The complete report, every panel

This is the real Aimosti report engine rendered over a made-up genome, deliberately loaded so every module has something to show. The panels below are rendered in full: 39 clinical conditions, 56 carrier genes, 27 traits, and 80 research entries, each one showing what a file could and could not say about it.

A whole-genome file also gets the wider ClinVar screen: the variants ClinVar lists as disease-causing in 1,363 more recessive and X-linked genes, and in 52 dominant genes whose conditions ClinGen rates actionable, shown here with the sample's own matches. The report carries 11 disease polygenic scores, and 5 emerging-research scores sit in a card of their own.

Deep Read reads a further 29 pharmacogenes. It is a paid add-on at $49 and it needs aligned reads (BAM or CRAM), so it is rendered here to show what it produces rather than what the report above includes.

What every panel below sits on: 1,054 panel regions of GRCh38, inside which ClinVar 2026-07 lists 483,295 classified variants. Those regions are the core of what is examined. The wider ClinVar screen adds 106,278 exact ClinVar positions, and each polygenic score adds its own markers. What we analyse lists what is reported, by name.

How to read it. A blank row means no variant was found there, which is not the same as the condition being ruled out. Callability is measured per gene, not per base, so treat it as strong evidence rather than a guarantee. Nothing here is a diagnosis.

This is the report you get. A re-analysis of the DNA file you already own, from $19 (chip) or $39 (whole-genome). Yours to keep, updated as the science moves.

Get your report Compare plans

Whole-genome (WGS) Chip upload (23andMe / Ancestry)

A chip array reads a fixed set of common spots; whole-genome sequencing reads (nearly) the entire genome. Both make a real, complete report, so flip between them to see what each file can tell you. A chip already powers pharmacogenomics, polygenic scores, traits, ancestry and the common single-gene findings; whole-genome data adds the rare-variant clinical & carrier panels and the wider ClinVar screen. Deep Read is a separate add-on and needs aligned reads (BAM or CRAM), which neither file above carries.

Your re-analysis report

Reference build GRCh38 · generated 2026-10-10

This year, in five lines

To discuss with a clinician2 findings in the established tier: F2, F5. Discuss with a clinician
For a prescriber4 medication flags: ABCG2, CYP2C19, SLCO1B1, VKORC1. The clinician hand-off carries the guidance for each. Discuss with a clinician
For family planningCarrier status needs a sequenced file; a genotyping chip does not read the panel.
Exploratory by designFrontier and the Fringe describe associations, not results; they carry no next step and are firewalled from everything above them.
What this file could not examineA genotyping chip does not read the clinical-findings or carrier panels.

Five lines derived from the sections below, never a verdict. Each links to the detail it summarises.

Counts only, never a verdict. Open each finding for the detail. An empty section means "not found in the regions examined", never "examined and clear".

Your genotyping chip, read in full. A complete report over everything an array genuinely calls: your pharmacogenomics, polygenic risk scores, traits, ancestry, and well-characterised single-gene findings (APOE, HFE, Lp(a), ABO). It's a real report, and it's yours to keep. A genotyping chip reads only a few hundred thousand pre-selected positions, about 0.02% of your genome, chosen for common variation. It does not sequence the rest, so it cannot find the rare or novel pathogenic variants the clinical and carrier modules look for: a “no finding” from chip data means “this chip never looked”, far more so than with whole-genome sequencing. Chip data suits common-variant traits, pharmacogenomic tag SNPs, and haplogroup ancestry, not clinical or carrier screening. The rare-variant clinical & carrier panels and the read-level Deep Read panel need a whole-genome (WGS) or aligned-reads file, a later upgrade whenever you want it, never a correction. Nothing here is missing for what a chip can do.

Tier 1 · Established

Clinical findings: single variants only

A genotyping chip can reliably read a small set of single-variant findings, but not the full clinical-findings panel or carrier status (those need whole-genome data). From your chip we read these specific variants: Factor V Leiden thrombophilia, Prothrombin G20210A thrombophilia. Any detected below are shown as single-variant findings: for that variant only, never a diagnosis. Upload a whole-genome (WGS) file for the full clinical and carrier panels.

Risk factors

Common, well-established variants that modestly shift risk: context to be aware of, never a diagnosis.

F2 Prothrombin G20210A thrombophilia ☆☆☆☆Low: a modest risk factor Discuss with a clinician

Prothrombin G20210A is a common, well-established variant that modestly raises the chance of abnormal blood clots. Like Factor V Leiden, it is a risk factor, not a disease or a diagnosis.

Evidence ☆☆☆☆ Effect Low: a modest risk factor

Variant: 11:46739505 G>A · rs1799963
Your genotype: heterozygous
ClinVar classification: Established common risk factor
ClinVar’s reported condition: Prothrombin G20210A thrombophilia
Confidence: ☆☆☆☆ curated single-variant finding
Population frequency: frequency unknown
Inheritance: Autosomal dominant (risk factor)

Prothrombin (clotting factor II, the F2 gene) is one of the proteins that lets blood clot. The G20210A variant (rs1799963) sits in a non-coding part of the gene and slightly increases how much prothrombin the body makes. A little more prothrombin means blood clots somewhat more readily than average.

This is a modest risk factor, not a diagnosis. It is one of the two most common inherited clotting variants in people of European ancestry (the other being Factor V Leiden), and the large majority of people who carry one copy never have a clotting problem. As with Factor V Leiden, the variant matters most in situations where the baseline clot risk is already raised: pregnancy, surgery, prolonged immobility, or oestrogen-containing contraception.

Carrying two copies, or carrying it alongside Factor V Leiden or other clotting risk factors, increases the effect further. The fair framing is "a common risk factor worth being aware of", not alarm.

This is a common risk factor, not a clinical diagnosis. If you have a personal or family history of blood clots, discuss it (and this finding) with a doctor before any decision (for example about contraception or clot prevention).

Sources: MedlinePlus Genetics: F2 gene · MedlinePlus Genetics: Prothrombin thrombophilia

F5 Factor V Leiden thrombophilia ☆☆☆☆Low: a modest risk factor Discuss with a clinician

Factor V Leiden is a common, well-established variant that modestly raises the chance of abnormal blood clots. It is a risk factor, not a disease or a diagnosis.

Evidence ☆☆☆☆ Effect Low: a modest risk factor

Variant: 1:169549811 C>T · rs6025
Your genotype: heterozygous
ClinVar classification: Established common risk factor
ClinVar’s reported condition: Factor V Leiden thrombophilia
Confidence: ☆☆☆☆ curated single-variant finding
Population frequency: frequency unknown
Inheritance: Autosomal dominant (risk factor)

Factor V Leiden (the F5 rs6025 variant) is one of the most common inherited variants affecting blood clotting in people of European ancestry. Normally a protein called activated protein C switches off clotting factor V at the right moment; the Leiden variant makes factor V slightly resistant to being switched off, so blood clots a little more readily than average.

This is a modest risk factor, not a diagnosis. Most people who carry one copy never have a clotting problem. The variant becomes more relevant in specific situations: during pregnancy, after surgery, with prolonged immobility, or alongside oestrogen-containing contraception, where the baseline clot risk is already raised.

Carrying two copies, or carrying it together with other clotting risk factors, increases the effect further. This is exactly the kind of common, well-replicated finding where the the fair framing is "a common risk factor worth being aware of", not alarm.

This is a common risk factor, not a clinical diagnosis. If you have a personal or family history of blood clots, discuss it (and this finding) with a doctor.

Sources: MedlinePlus Genetics: F5 gene · MedlinePlus Genetics: Factor V Leiden thrombophilia

APOE: Alzheimer’s-risk genetics

Sensitive · opt-in

Read this before you reveal it. APOE is the strongest common genetic influence on the risk of late-onset Alzheimer's disease. Your two ε alleles place you in a relative-risk band: higher or lower than the average ε3/ε3 genotype. It is one factor among many, it is not a diagnosis or a forecast, and nothing you can do changes the gene itself. APOE is a risk factor, not a diagnosis and not a prediction. Most people who carry an ε4 allele never develop Alzheimer's disease, and many people who develop it carry none. No treatment changes your APOE genotype, and clinical guidelines do not recommend APOE genotyping to predict dementia in people without symptoms. We show it because it is your information (which some people value for life, family and financial planning), not because it tells you what will happen.

Reveal my APOE result

APOE ε3/ε4 · incomplete Not established

Your APOE result could not be completed: rs7412 was never examined in your file, and the ε-diplotype is read from both markers together.

Your genotype: ε3/ε4 · rs7412 not examined
Relative risk vs ε3/ε3: Not established: the reading is incomplete

How to read this: ancestry, sex & the limits of the number

The size of the ε4 effect depends on genetic ancestry and sex. It is strongest in people of East Asian and European ancestry and substantially weaker in people of African ancestry; the relativities shown here come mostly from European-ancestry studies. Women who carry ε4 appear to be at somewhat higher risk than men, particularly with a single ε4 copy. Read every number as a broad population average, not your personal odds.

What this isn’t: APOE is a risk factor, not a diagnosis and not a prediction. Most people who carry an ε4 allele never develop Alzheimer's disease, and many people who develop it carry none. No treatment changes your APOE genotype, and clinical guidelines do not recommend APOE genotyping to predict dementia in people without symptoms. We show it because it is your information (which some people value for life, family and financial planning), not because it tells you what will happen.

rs7412 could not be read from your file, so this ε-diplotype is incomplete. The ε-diplotype is read from rs429358 and rs7412 together, so an unread position there means your ε2 status was not examined, not that it is absent. This is not a negative result. A whole-genome or gVCF file covering both markers can settle it.

2026-06-03 · MedlinePlus Genetics: APOE gene · NIA: Alzheimer's Disease Genetics Fact Sheet · Farrer et al. 1997, JAMA: APOE and Alzheimer risk meta-analysis (PubMed) · dbSNP: rs429358 · dbSNP: rs7412

Lipoprotein(a): inherited heart-risk marker

Lipoprotein(a), or Lp(a), is a largely inherited, independent and causal risk factor for heart disease and aortic valve narrowing. Your level is mostly fixed for life. We can read two common genetic markers that flag a high level, but they are a partial proxy, not a measurement.

LPALp(a)Likely elevated

A marker linked to elevated Lp(a) is present.

What this reads as: One tag allele: associated, on average, with a higher Lp(a) level and higher cardiovascular risk.
Markers found: rs10455872 (heterozygous)

You carry one copy of a tag allele associated with elevated lipoprotein(a). On average this points to a higher-than-typical Lp(a) and a modestly higher risk of heart disease and aortic stenosis, but the effect size varies widely between people, because the tag is only a partial stand-in for the actual level. Confirm with a blood test before drawing any conclusion.

How to read this: ancestry & why a blood test is the real answer

The two tag markers we can read were characterised mainly in people of European ancestry and capture only part of the genetic signal in other groups. A negative read is therefore even less informative outside European ancestry. Genetic ancestry never substitutes for the blood test.

Important: This is a genetic proxy, not a measurement. The main driver of Lp(a), the LPA KIV-2 repeat, cannot be read from a variant file, so these markers can flag a likely-high level but can NEVER rule one out. The only way to know your Lp(a) is a one-time blood test (reported in nmol/L); guidelines suggest measuring it once in adulthood. Discuss results and any treatment with a clinician. Lp(a) is not changed by diet, and targeted Lp(a)-lowering drugs are still in trials.

2026-06-03 · MedlinePlus Genetics: LPA gene · Clarke et al. 2009, NEJM: LPA variants (rs10455872, rs3798220) and coronary disease (PubMed) · dbSNP: rs10455872 · dbSNP: rs3798220

APOL1: inherited kidney-disease risk

APOL1 risk variants are the strongest common genetic influence on several kidney diseases. They are found almost entirely on West/Central African ancestral haplotypes, where they likely persist because they protect against African sleeping sickness. Two risk alleles raise kidney-disease risk. It is a risk factor, not a diagnosis, and an actionable one.

APOL1G0/G0 (partial)Not determined

Your APOL1 result could not be completed: G1 and G2 were never examined in your file, and it takes two risk alleles to reach the high-risk genotype.

Relative risk: Not established: the reading is incomplete
Risk alleles found: At least 0 of 2 · G1, G2 not examined

How to read this: ancestry & the limits of the number

The G1 and G2 risk alleles arose on, and are essentially confined to, West/Central African genetic ancestry. Ancestry is not race, and allele frequencies vary continuously across people. This finding applies to anyone who carries the alleles, however they identify. Outside that ancestry the alleles are vanishingly rare, so a 'no risk allele' result is expected there and tells you little.

Important: Two APOL1 risk alleles raise the risk of chronic kidney disease and FSGS, especially alongside a 'second hit' such as HIV or certain infections, but most people with the high-risk genotype never develop kidney failure. This is actionable in a good way: clinicians manage APOL1 high-risk status with blood-pressure control, avoidance of kidney-toxic drugs, and kidney-function monitoring, and APOL1-targeted treatments are emerging. Discuss with a clinician; this is not a diagnosis.

G1 (rs73885319) is not on your array; G2 (rs71785313) is a 6 bp deletion, not a single-base change, and a genotyping array does not type it. APOL1 risk is recessive-like — it takes TWO risk alleles to reach the high-risk genotype — so with G1 and G2 unexamined the count below is a minimum, not a total, and the high-risk genotype has not been ruled out. Treat G1 and G2 as "not examined", never as absent.

2026-06-07 · MedlinePlus Genetics: APOL1 gene

Hemochromatosis (HFE): inherited iron-overload risk

Hereditary haemochromatosis is a condition in which the body can absorb and store too much iron over many years. It is strongly linked to two common changes in the HFE gene. Carrying an at-risk combination raises the chance of iron overload, but many people who carry one never develop a problem, which is why this is a risk factor, not a diagnosis.

HFEno risk alleles (partial)No risk alleles · partial

Neither HFE risk change was found.

What this reads as: Neither HFE risk change was found: the common genotype for this gene.
Risk alleles found: None of the two HFE risk changes · C282Y, H63D not examined

Neither the C282Y nor the H63D change was reported in your file, which is the common result. This makes hereditary haemochromatosis from these two well-known HFE variants unlikely. It does not rule out every rarer cause of iron overload, and it is best read as 'not flagged' rather than a guarantee: a variant file cannot prove a position was covered, and H63D in particular cannot be resolved from genotyping-array data at all, because its two letters read the same on either strand. Where a marker could not be examined, the note on the card above names it.

How to read this: penetrance & why a blood test is the real answer

Penetrance is incomplete, and how incomplete depends on what is counted. Among people with two copies of C282Y, an Australian study found iron-overload disease in about 28% of men and 1% of women, while UK Biobank projected a haemochromatosis diagnosis by age 80 for about 56% of men and 40% of women. Many never develop disease from it, and it is less common and usually milder in women, who lose iron through menstruation and pregnancy. Genetics set the predisposition; whether iron actually accumulates depends on age, sex, diet, blood loss and other factors.

Important: This is a genetic risk factor, not a diagnosis and not a measurement of your iron levels. The only way to know your iron status is a simple blood test, ferritin and transferrin saturation, which your clinician can interpret. Hereditary haemochromatosis is very treatable when caught (usually by periodic blood removal), so an at-risk genotype is information to act on calmly with a clinician, never a cause for alarm.

C282Y (rs1800562) is not on your array; H63D (rs1799945) is not on your array. Treat C282Y and H63D as "not examined", never as absent: the genotype below reflects only the position(s) your file could actually be read for. A ferritin / transferrin-saturation blood test reads iron directly and settles the question.

2026-10-10 · MedlinePlus Genetics: HFE gene · GeneReviews: HFE Hemochromatosis (Porto et al.) · Allen et al. 2008, NEJM: Iron-overload-related disease in HFE hereditary hemochromatosis (PMID 18199861) · Lucas et al. 2024, BMJ Open: HFE genotypes, haemochromatosis diagnosis and clinical outcomes at age 80 years (PMID 38479735) · dbSNP: rs1800562 (C282Y) · dbSNP: rs1799945 (H63D)

What each HFE genotype has been measured to mean, and the chip strand problem →

Celiac HLA (DQ2.5 / DQ8): a rule-out, not a risk score

Celiac disease almost always occurs in people who carry one of two HLA types, DQ2.5 or DQ8, though most people who carry either type never develop it. We read two markers linked to these types. A positive result is common and only mildly informative. Not finding either type is the more useful result, the same reasoning a clinician uses to stop a celiac workup rather than start one, but that result only stands on a file that could have shown these markers in the first place.

HLA-DQnot called from this fileNot called

DQ2.5 and DQ8 not called from this file.

What this reads as: These two markers were not resolved from your file, so no celiac HLA read is available. This is not a negative result, only a gap in what this particular file can show.
Tag markers found: Neither marker could be read from this file, so this is not a negative

Neither DQ2.5 nor DQ8 could be read from your array with enough confidence to say anything either way. This is not a negative: your chip either did not include these two positions or reported genotypes there that could not be read. (On whole-genome sequencing the same status has a different cause: DQ2.5's marker is under-read by the sequencing method itself.) Either way, no celiac read is available from this file; a blood test for celiac antibodies remains the way to check. That test is done while gluten is still in the diet, because a gluten-free diet started beforehand can make it read falsely clear.

Neither marker could be resolved from your array, so this is not a negative. Either these two positions are not on your chip, or the genotypes it reported there could not be read.

2026-07-31 · MedlinePlus Genetics: Celiac disease · Monsuur et al. 2008, PLoS ONE: tag-SNP validation for DQ2.5/DQ8 (PMID 18509540) · Koskinen et al. 2009: DQ2.5/DQ8 tag-SNP validation in a Finnish population (PMID 19255754) · Karell et al. 2003, Human Immunology: HLA types in celiac patients lacking DQ2/DQ8 (PMID 12651074) · dbSNP: rs2187668 (DQ2.5 tag) · dbSNP: rs7454108 (DQ8 tag)

What DQ2 and DQ8 can and can't rule out, file type by file type →

Tier 2 · Well-supported

Pharmacogenomics

4 actionable

How your genotype may affect specific medicines, restating CPIC guideline guidance for gene–drug pairs callable from your file. Information to discuss with a prescriber, never an instruction from us. Some pairs are out of scope for a standard variant file (e.g. CYP2D6 and TPMT’s common *3A) and are deliberately omitted rather than guessed. CYP2D6 can be resolved from your aligned reads: see Deep Read.

ABCG2 rs2231142 variant (T)/rs2231142 variant (T) Poor function

A pump, also called BCRP, that pushes rosuvastatin back out of gut and liver cells. One common variant, rs2231142 (Q141K), weakens it, and the same dose then leaves more rosuvastatin in the blood.

Your alleles: rs2231142 variant (T) (Decreased function) · rs2231142 variant (T) (Decreased function)

Two copies of the rs2231142 variant: the lowest ABCG2 pump activity CPIC grades.

Rosuvastatin

Increased rosuvastatin exposure compared to normal and decreased function; unknown myopathy risk; increased lipid-lowering effects.

CPIC: with your SLCO1B1 result (decreased function), the guideline's line for this ABCG2 result is "Prescribe ≤10mg as a starting dose and adjust doses of rosuvastatin based on disease-specific and specific population guidelines. If dose >10mg needed for desired efficacy, consider an alternative statin or combination therapy (i.e., rosuvastatin plus non-statin guideline directed medical therapy) (PMID: 30423391)."

Every defining variant for this gene was resolved from your array: this call is read directly, not assumed.

Which defining variants were on your array (1)

rs2231142 (rs2231142 variant (T)): on your array (T/T)

CPIC (statins 2022) · CPIC: statins & SLCO1B1, ABCG2, CYP2C9

CYP2C19 *1/*2 Intermediate metabolizer

A liver enzyme that activates clopidogrel and clears several antidepressants and acid-blockers. Common variants raise or lower its activity.

Your alleles: *1 (Normal function) · *2 (No function)

Reduced activity from one no-function allele (and, in some genotypes, one increased-function allele that does not fully compensate).

Clopidogrel

Reduced active-metabolite formation and higher on-treatment platelet reactivity.

CPIC: for acute coronary syndrome / PCI, the guideline suggests an alternative antiplatelet (e.g. prasugrel or ticagrelor) where not contraindicated.

Escitalopram / citalopram

Reduced metabolism when compared to CYP2C19 normal metabolizers. Higher plasma concentrations may increase the probability of side effects.

CPIC: "Initiate therapy with recommended starting dose. Consider a slower titration schedule and lower maintenance dose than normal metabolizers."

Every defining variant for this gene was resolved from your array: this call is read directly, not assumed.

Carisoprodol, clobazam and diazepam: CPIC reviewed these with CYP2C19 (evidence level B/C) and publishes no CYP2C19-based dosing recommendation for these pairs, so there is no CPIC dosing guidance to restate here. Duloxetine, esomeprazole, flibanserin, fluoxetine and rabeprazole: CPIC reviewed these with CYP2C19 (evidence level C) and publishes no CYP2C19-based dosing recommendation for these pairs, so there is no CPIC dosing guidance to restate here. Fluvoxamine, paroxetine, venlafaxine and vortioxetine: CPIC's dosing guidance for these is through CYP2D6. Brivaracetam: CPIC rates the evidence linking it to CYP2C19 as level B, but the CPIC release we use publishes no dosing recommendation for the pair. That is not a sign your CYP2C19 result does not matter for it.

Which defining variants were on your array (3)

rs4244285 (*2): on your array (G/A)
rs4986893 (*3): on your array
rs12248560 (*17): on your array

CPIC (clopidogrel 2022; SSRIs 2023) · CPIC: clopidogrel & CYP2C19 · CPIC: SSRIs & CYP2C19/CYP2D6

SLCO1B1 *1/*5 Decreased function

A liver transporter that pulls statins out of the blood. Reduced-function variants raise statin exposure and the risk of muscle side-effects.

Your alleles: *1 (Normal function) · *5 (Decreased function)

One decreased-function (*5) allele raises blood statin levels.

Simvastatin

Higher exposure and a moderately increased myopathy risk, dose-dependent.

CPIC: the guideline's line for this result is "Prescribe an alternative statin depending on the desired potency (see Figure 1 of PMID: 35152405 for recommendations for alternative statins). If simvastatin therapy is warranted, limit dose to <20mg/day."

Every defining variant for this gene was resolved from your array: this call is read directly, not assumed.

Elagolix: CPIC reviewed it with SLCO1B1 (evidence level B/C) and publishes no SLCO1B1-based dosing recommendation for this pair, so there is no CPIC dosing guidance to restate here. Methotrexate: CPIC reviewed it with SLCO1B1 (evidence level C) and publishes no SLCO1B1-based dosing recommendation for this pair, so there is no CPIC dosing guidance to restate here.

Which defining variants were on your array (1)

rs4149056 (*5): on your array (T/C)

CPIC (statins 2022) · CPIC: statins & SLCO1B1 (and others)

VKORC1 A/A Highly increased sensitivity

The vitamin-K-epoxide-reductase gene warfarin acts on. A common promoter variant (-1639A) lowers its expression and raises warfarin sensitivity.

Your alleles: A (Reduced expression) · A (Reduced expression)

Two -1639A alleles markedly lower VKORC1 expression and raise warfarin sensitivity.

Warfarin

Markedly greater sensitivity; a notably lower dose is often required.

CPIC: warfarin dosing follows the genotype-guided algorithm; two -1639A alleles predict a markedly lower dose and closer INR monitoring.

Every defining variant for this gene was resolved from your array: this call is read directly, not assumed.

Which defining variants were on your array (1)

rs9923231 (A): on your array (T/T)

CPIC (warfarin 2017) · CPIC: warfarin, CYP2C9 & VKORC1

Genes with a typical (normal) result (3)

No genotype-driven prescribing change is expected for these genes.

CYP2C9 *1/*1 Normal metabolizer

A liver enzyme that clears warfarin, several NSAIDs and phenytoin. Reduced-function variants slow that clearance.

Your alleles: *1 (Normal function) · *1 (Normal function)

Typical CYP2C9 activity (activity score 2.0).

Warfarin

Normal warfarin clearance.

CPIC: warfarin dosing follows the genotype-guided algorithm together with VKORC1 and clinical factors; no CYP2C9-driven adjustment beyond it.

Every defining variant for this gene was resolved from your array: this call is read directly, not assumed.

Aceclofenac, diclofenac, flibanserin, indomethacin, lumiracoxib, nabumetone and naproxen: CPIC reviewed these with CYP2C9 (evidence level C) and publishes no CYP2C9-based dosing recommendation for these pairs, so there is no CPIC dosing guidance to restate here. Avatrombopag, dronabinol, erdafitinib and lesinurad: CPIC reviewed these with CYP2C9 (evidence level B/C) and publishes no CYP2C9-based dosing recommendation for these pairs, so there is no CPIC dosing guidance to restate here. Aspirin: CPIC's dosing guidance for it is through G6PD. Acenocoumarol: CPIC rates the evidence linking it to CYP2C9 as level B, but the CPIC release we use publishes no dosing recommendation for the pair. That is not a sign your CYP2C9 result does not matter for it. Siponimod: CPIC rates the evidence linking it to CYP2C9 as level A, but the CPIC release we use publishes no dosing recommendation for the pair. That is not a sign your CYP2C9 result does not matter for it. The FDA drug label lists CYP2C9 testing as required.

Which defining variants were on your array (2)

rs1799853 (*2): on your array
rs1057910 (*3): on your array

CPIC (warfarin 2017) · CPIC: warfarin, CYP2C9 & VKORC1

CYP4F2 *1/*1 Normal CYP4F2 function

An enzyme in vitamin-K recycling. The *3 variant lowers its activity, which slightly raises the warfarin dose people tend to need.

Your alleles: *1 (Normal function) · *1 (Normal function)

Two *1 alleles: typical CYP4F2 activity.

Warfarin

No CYP4F2-driven change to the predicted dose.

CPIC: warfarin dosing follows the genotype-guided algorithm (CYP2C9 + VKORC1 + clinical factors); no CYP4F2 adjustment for *1/*1.

Every defining variant for this gene was resolved from your array: this call is read directly, not assumed.

Acenocoumarol: CPIC rates the evidence linking it to CYP4F2 as level B, but the CPIC release we use publishes no dosing recommendation for the pair. That is not a sign your CYP4F2 result does not matter for it. Phenprocoumon: CPIC rates the evidence linking it to CYP4F2 as level B, but the CPIC release we use publishes no dosing recommendation for the pair. That is not a sign your CYP4F2 result does not matter for it.

Which defining variants were on your array (1)

rs2108622 (*3): on your array

CPIC (warfarin 2017) · CPIC: warfarin, CYP2C9, VKORC1 & CYP4F2

NUDT15 *1/*1 Normal metabolizer

An enzyme that detoxifies thiopurine drugs (azathioprine, mercaptopurine). Reduced-function variants raise the risk of severe bone-marrow suppression, and are most common in East-Asian and Hispanic ancestries.

Your alleles: *1 (Normal function) · *1 (Normal function)

Typical NUDT15 activity.

Azathioprine / mercaptopurine

Normal thiopurine handling expected.

CPIC: standard, label-recommended starting dose, titrated to clinical response as usual.

Every defining variant for this gene was resolved from your array: this call is read directly, not assumed.

Which defining variants were on your array (1)

rs116855232 (*3): on your array

CPIC (thiopurines 2018) · CPIC: thiopurines, TPMT & NUDT15

Could not be resolved from your file (2)

We don’t assign a result where the data can’t support one: see each gene’s note.

DPYD DPYD

The enzyme that breaks down the chemotherapy drugs 5-fluorouracil and capecitabine. Reduced-function variants can cause severe, occasionally fatal toxicity at standard doses.

One of this gene's defining variants could not be resolved from chip data (rs67376798 is on your array but is a strand-ambiguous SNP (A/T or C/G) that chip data can't orient, so it can't be resolved): this is "not examined", never a reference result.

Tegafur: CPIC reviewed it with DPYD (evidence level C) and publishes no DPYD-based dosing recommendation for this pair, so there is no CPIC dosing guidance to restate here.

Which defining variants were on your array (2)

rs3918290 (c.1905+1G>A): on your array
rs67376798 (c.2846A>T): on your array, but not resolvable from chip data (see the note above for why)

CPIC (fluoropyrimidines 2017) · CPIC: fluoropyrimidines & DPYD

HLA-B HLA-B

An immune-system gene whose variants sharply raise the risk of two distinct severe drug reactions. HLA-B*15:02 raises the risk of Stevens-Johnson syndrome / toxic epidermal necrolysis with the seizure and nerve-pain drug carbamazepine and its relatives, and is most common in people of South-East and East-Asian genetic ancestry. HLA-B*57:01 raises the risk of a severe hypersensitivity reaction to the HIV drug abacavir. The tag allele has a frequency of about 3% in Finns and 4% in Europeans, and because it is not confined to one ancestry group, clinical guidelines call for screening every patient before the drug is started, regardless of ancestry. Both are read via strongly-linked tag markers; ancestry is not race, and a negative result is not a global exclusion.

rs144012689 is not on your array. *15:02 could not be assessed. Your array typed rs2395029 and found no *57:01.

Which defining variants were on your array (2)

rs144012689 (*15:02): not on your array
rs2395029 (*57:01): on your array

CPIC (carbamazepine/oxcarbazepine & HLA-B 2017; abacavir & HLA-B 2014) · CPIC: carbamazepine/oxcarbazepine & HLA-B · CPIC: abacavir & HLA-B

Traits

Well-replicated, benign wellness traits that clear our evidence gate: catalogued in the GWAS Catalog, genome-wide significant, and replicated across independent studies. Each is a tendency, not a verdict.

Read from your file: 5 of 5. The other 0 had no variant reported in your file, so they’re read as the common (reference) genotype: an inference, not a measurement (see each card’s note).

ALDH2 Alcohol flush response Flush likely

Whether you tend to flush (going red, warm and sometimes queasy) soon after drinking alcohol.

Your genotype: G/A (rs671)
Most-associated outcome: Flush likely

One inactive (A) allele: ALDH2 works at reduced capacity, so acetaldehyde builds up and flushing, warmth or nausea after alcohol are likely. This matters for health, not just comfort: acetaldehyde is a recognised carcinogen, and with this genotype drinking raises the risk of alcohol-associated oesophageal and head-and-neck cancer above average. (The same enzyme activates the heart drug nitroglycerin, which can work less well in ALDH2-deficient people.) Information, not advice.

Why we include this: the evidence

Source: Functional variant: Crabb et al. 1989 (J Clin Invest) · PMID:2562960
Significance: Below genome-wide significance
Replication: Crabb et al. 1989 demonstrated biochemically that the ALDH2*2 allele (Glu504Lys) is dominant-negative: heterozygotes have markedly reduced enzyme activity. The variant is extensively replicated across East-Asian cohorts and has been shown to confer risk of alcohol-associated cancers; A allele frequency in gnomADg:eas = 0.2249.
Where established: The A allele is common in East-Asian ancestry (gnomADg:eas ~22%) and nearly absent in European and African ancestry (gnomADg:ALL ~0.8%)
Effect: The A allele dominantly reduces ALDH2 activity, causing the flush response

ALDH2 clears acetaldehyde, alcohol's toxic first breakdown product. The Lys504 (A) enzyme is nearly inactive and acts as a dominant-negative inhibitor of the wild-type subunit, so even one A allele substantially reduces tetrameric enzyme activity, allowing acetaldehyde to accumulate and trigger flushing.

MedlinePlus Genetics: ALDH2 gene · NIH/PMC: acetaldehyde, ALDH2 deficiency and alcohol-associated cancer

CYP1A2 Caffeine metabolism Slower metaboliser

How quickly your liver tends to clear caffeine, set largely by the CYP1A2 enzyme.

Your genotype: C/A (rs762551)
Most-associated outcome: Slower metaboliser

You carry at least one C allele, associated with slower caffeine clearance: caffeine tends to linger longer, so the same cup can feel stronger or disturb sleep more. A modest tendency, not a rule; habit and dose matter more day to day.

Why we include this: the evidence

Source: PharmGKB · PA27093
Significance: Below genome-wide significance
Replication: CYP1A2 *1F (rs762551) is the canonical PharmGKB marker for CYP1A2 inducibility (gene PA27093); slower vs faster caffeine clearance by genotype is documented across multiple pharmacokinetic studies.
Where established: Studied across several ancestries; effect sizes are modest
Effect: The A allele is associated with faster (more inducible) CYP1A2 activity; C-allele carriers tend to be slower metabolisers

rs762551 marks the CYP1A2 *1F haplotype; CYP1A2 performs the bulk of caffeine breakdown, and its inducibility differs by genotype (and by smoking).

MedlinePlus Genetics: CYP1A2 gene

ABCC11 Earwax type & body odour Dry earwax

A single ABCC11 variant that sets whether your earwax is wet or dry, and tracks with how much underarm body odour you tend to produce.

Your genotype: T/T (rs17822931)
Most-associated outcome: Dry earwax

Two T alleles: dry, flaky earwax and a tendency toward noticeably less underarm body odour, because the apocrine glands make fewer of the compounds skin bacteria turn into smell. Entirely benign; just a difference.

Why we include this: the evidence

Source: Functional variant: Yoshiura et al. 2006 (Nat Genet) · PMID:16444273
Significance: Below genome-wide significance
Replication: Confirmed in multiple populations and by functional biochemical studies; the T/T genotype reliably predicts dry earwax and reduced apocrine secretion across ethnic groups. In East-Asian populations (gnomADg:eas ~84%) the T allele is the major allele.
Where established: The dry (T) allele is common in East-Asian ancestry (gnomADg:eas ~84%) and rarer in European and African ancestry (gnomADg:ALL ~14%)
Effect: The T allele is recessive: T/T gives dry earwax and reduced odour; a C allele gives wet earwax

ABCC11 encodes an ATP-binding cassette transporter expressed in earwax and apocrine sweat glands. The 538G>A substitution (Gly180Arg; T allele) causes protein misfolding and degradation, disabling the transporter and giving dry flaky earwax and less underarm odour.

MedlinePlus Genetics: ABCC11 gene

MCM6 Lactase persistence Likely lactase persistent

Whether your body tends to keep producing lactase, the enzyme that digests the milk sugar lactose, into adulthood.

Your genotype: T/T (rs4988235)
Most-associated outcome: Likely lactase persistent

You carry at least one persistence (T) allele, so you most likely keep digesting lactose comfortably as an adult. This is a tendency, not a guarantee: tolerance also depends on gut bacteria and how much dairy you eat.

Why we include this: the evidence

Source: Functional variant: Enattah et al. 2002 (Nat Genet) · PMID:11788828
Significance: Below genome-wide significance
Replication: One of the most replicated human trait associations; the −13910 C>T enhancer variant was first identified in Finnish and other European cohorts and has since been confirmed in many independent cohorts worldwide. Functional studies show the T allele creates a binding site for Oct-1 transcription factor, maintaining LCT expression into adulthood. MAF is the minor (C / non-persistence) allele frequency in gnomADg:NFE (0.3647); the T / persistence allele is the majority allele in Northern Europeans.
Where established: Best established in Northern European-ancestry populations; the persistence (T) allele reaches ~64% in NFE but is much rarer in East Asian, Middle-Eastern, and most African populations; other persistence variants exist in African and Middle-Eastern populations and are not covered here
Effect: The T allele is dominantly associated with continued lactase production into adulthood

rs4988235 sits in an MCM6 intron that acts as an enhancer of the neighbouring LCT (lactase) gene; the T allele keeps LCT transcribed after weaning via an Oct-1 binding site.

MedlinePlus Genetics: Lactose intolerance

ACTN3 Muscle fibre type (ACTN3) Endurance-associated

A common variant in the ACTN3 "speed gene" that nudges muscle toward power/sprint vs endurance tendencies.

Your genotype: T/T (rs1815739)
Most-associated outcome: Endurance-associated

Two stop (T) alleles: you make no α-actinin-3, which is completely benign and common. This genotype is slightly more frequent in endurance athletes than power athletes. A tendency only; trainability dominates.

Why we include this: the evidence

Source: Functional variant: Yang et al. 2003 (Am J Hum Genet) · PMID:12879365
Significance: Below genome-wide significance
Replication: One of the most studied muscle-genetics variants. Yang et al. 2003 (Am J Hum Genet, N=429 elite athletes + controls) first showed the R577X null allele (T) is under-represented in sprint/power athletes. Confirmed in many independent cohorts worldwide. The null allele is common: ~18% of people are T/T worldwide (gnomADg:ALL T = 0.3751, so T/T ≈ 14%). Functional and population evidence is strong; this is not a statistical GWAS association but a protein-null variant with a well-characterised phenotypic shift.
Where established: Studied across multiple ancestries; the X (T) allele is common worldwide, with gnomADg:ALL frequency 0.3751
Effect: The functional C allele is over-represented in elite power athletes; T/T removes α-actinin-3 from fast-twitch fibres

rs1815739 introduces a premature stop codon (R577X) in ACTN3, eliminating α-actinin-3 from fast-twitch (type II) muscle fibres. T/T individuals rely entirely on α-actinin-2, a structural shift associated with endurance-leaning muscle characteristics, and entirely benign.

MedlinePlus Genetics: ACTN3 gene

Traits we couldn’t resolve from your file (22)

The marker wasn’t callable or the genotype isn’t in the curated table: see each note.

ADH1B Alcohol metabolism speed (ADH1B)

How quickly your liver converts alcohol (ethanol) into acetaldehyde, the first step in breaking down a drink, shaped largely by the ADH1B variant you carry.

Your genotype: Not read (rs1229984)

Why we include this: the evidence

Source: Functional variant: Edenberg & McClintick 2018 (Alcohol Clin Exp Res) review; original ADH1B*2 characterisation Bosron & Li 1986, Hurley et al. 1990 · PMID:30320893
Significance: Below genome-wide significance
Replication: The ADH1B*2 allele (His48Arg, C allele at rs1229984) produces an alcohol dehydrogenase beta2 subunit with 40–100x higher Vmax than the beta1 (T allele) form. This was biochemically characterised by Jörnvall et al. 1984 and Hurley et al. 1990 and has been replicated in hundreds of pharmacokinetic and epidemiological studies. The C allele reaches ~30% in East-Asian populations (gnomADg:EAS = 0.2961) and ~3.6% in European (gnomADg:NFE = 0.0357). GWAS confirm strong protective associations with alcohol dependence and aerodigestive cancer risk.
Where established: The C (fast/ADH1B*2) allele is common in East-Asian ancestry (~30%) and much rarer in European and African ancestry (~4% and ~1% respectively); maf reported as gnomADg:EAS (the population where it is most common)
Effect: The C allele (Arg48, ADH1B*2) is dominant for fast ethanol oxidation: even one copy substantially accelerates acetaldehyde production after drinking

ADH1B encodes the beta subunit of the class-I liver alcohol dehydrogenase that performs the first oxidative step (ethanol → acetaldehyde) in alcohol metabolism. The His48Arg (T>C) substitution stiffens the active site in a way that dramatically increases catalytic rate. Faster ADH1B means acetaldehyde accumulates more rapidly after drinking, producing nausea, flushing and discomfort, which acts as a natural deterrent to heavy drinking. The C allele is associated with protection against alcohol use disorder and alcohol-related cancers, but acetaldehyde is itself a carcinogen, so rapid production without equally fast ALDH2 clearance carries its own risks.

This trait's marker is not on this array, so it was never read. We leave it unresolved rather than show the common (reference) version, which would look like a result. A sequenced file, or an array that carries this marker, can settle it.

Edenberg & McClintick 2018: ADH1B and alcohol use disorders review (Alcohol Clin Exp Res) · MedlinePlus Genetics: ADH1B gene

OR2M7 region Asparagus urine smell detection

Whether you can smell the distinctive odour some people produce in urine after eating asparagus.

Your genotype: Not read (rs4481887)

Why we include this: the evidence

Source: Functional variant: Markt et al. 2016 (BMJ) · PMID:27965198
Significance: Below genome-wide significance
Replication: Markt et al. 2016 (BMJ, N=6,909) performed a GWAS of asparagus anosmia and identified rs4481887 in an olfactory-receptor gene cluster as the top hit; Eriksson et al. 2010 (PLoS Genet, PMID:20585627, N=9,126) first reported this association from a 23andMe web-based study. The trait is olfactory-cluster-localised and the A allele (smell-detected) reaches 20-31% in European populations.
Where established: Established mainly in European-ancestry cohorts
Effect: The A allele is associated with a greater ability to detect the asparagus-urine odour

rs4481887 lies within a cluster of olfactory-receptor genes (including OR2M7) on chromosome 1; these receptors are thought to detect the sulphur-containing methanethiol compounds asparagus metabolism produces.

GWAS Catalog: asparagus anosmia

TAS2R38 Bitter taste sensitivity (TAS2R38)

Whether bitter compounds, including in coffee, broccoli, and certain medicines, tend to taste strongly bitter to you, shaped mainly by the TAS2R38 haplotype you carry.

Your genotype: Not read (rs713598)

Why we include this: the evidence

Source: Functional variant: Kim et al. 2003 (Science) · PMID:12595690
Significance: Below genome-wide significance
Replication: Kim et al. 2003 positionally cloned TAS2R38 as the receptor underlying PTC/PROP bitter taste; the three-SNP PAV/AVI haplotype has been replicated in dozens of independent cohorts. The taster (PAV) and non-taster (AVI) haplotypes account for the majority of the normal population variance in bitter sensitivity. rs713598 and rs1726866 MAFs in gnomADg:ALL are ~0.447 and ~0.472 respectively, confirming both as common variants.
Where established: PAV and AVI haplotypes are found across all major ancestry groups; relative frequency varies (AVI is rarer in some East-Asian populations)
Effect: The PAV haplotype (G at rs713598 / G at rs1726866) confers bitter taste sensitivity; AVI (C/A) is non-tasting; heterozygotes are intermediate

TAS2R38 encodes a bitter taste receptor. The PAV form (Pro49/Ala262/Val296) is correctly folded and signals in response to phenylthiocarbamide, PROP and other bitter glucosinolates. The AVI form is non-functional for these ligands. Bitter sensitivity influences diet preferences and perception of off-notes in vegetables and beverages.

Kim et al. 2003: positional cloning of TAS2R38 in Science · MedlinePlus Genetics: Taste

CASC16 locus Chronotype (morning vs evening tendency)

A genetic tendency toward being a morning person or an evening person, shaped partly by a variant in the CASC16 region near chromosomal region 16q12.

Your genotype: Not read (rs12927162)

Why we include this: the evidence

Source: GWAS Catalog · GCST003429, GCST007576
Significance: Genome-wide significant (p ≤ 5×10⁻⁸)
Replication: Hu et al. 2016 (Nat Commun, GCST003429, N=89,283 Europeans) identified rs12927162-A as significantly associated with morning chronotype (OR=1.099, p=2e-12). Jones et al. 2019 (Nat Commun, GCST007576, N=697,832 Europeans) independently confirmed the locus with a larger beta (p=2e-32). The A allele consistently marks morning preference across both studies. gnomADg:ALL minor allele G frequency = 0.1914.
Where established: Characterised primarily in European-ancestry cohorts; replication in other ancestries is limited for this specific locus
Effect: The A allele is associated with morning preference; the G allele is associated with a tendency toward eveningness

rs12927162 sits in an intergenic region of the 16q12 gene desert (CASC16 locus), approximately 100 kb downstream of TOX3. The mechanism is not fully resolved; the region may harbour regulatory elements influencing circadian-clock gene networks. Chronotype is a polygenic trait shaped by many loci. This single variant explains only a fraction of individual variation. External factors such as light exposure, caffeine, age and social schedules dominate day-to-day timing.

Hu et al. 2016: chronotype GWAS (GCST003429, PMID:26835600) · Jones et al. 2019: chronotype GWAS in 697,828 individuals (GCST007576, PMID:30696823)

OR6A2 Cilantro (coriander) soapy taste

Whether fresh cilantro tends to taste soapy to you, linked to a smell-receptor gene cluster.

Your genotype: Not read (rs72921001)

Why we include this: the evidence

Source: Functional variant: Eriksson et al. 2012 (Flavour) · DOI:10.1186/2044-7248-1-22
Significance: Below genome-wide significance
Replication: Eriksson et al. 2012 (Flavour; preprint arXiv:1209.2096) identified rs72921001 near OR6A2 as the lead variant for cilantro soapy taste in a discovery cohort of 14,604 European-ancestry participants and replicated it in a distinct set of 11,851 (lead p=6.4e-9, OR=0.81 per A allele, i.e. the A allele is protective, so the C allele is the soapy-associated allele). Effect is modest; the locus sits in a cluster of olfactory receptor genes implicated in aldehyde perception.
Where established: Established mainly in European and South Asian ancestry cohorts; effect is modest and taste perception is also shaped by culture and exposure
Effect: The C allele is associated with a higher chance of perceiving cilantro as soapy

rs72921001 lies in a cluster of olfactory-receptor genes (including OR6A2) on chromosome 11 that detect the aldehydes giving cilantro its aroma; variant likely affects receptor sensitivity to these compounds.

Eriksson et al. 2012, Flavour: A genetic variant near olfactory receptor genes influences cilantro preference (preprint) · MedlinePlus Genetics: Genes and smell

AHR Coffee consumption tendency (AHR)

A genetic tendency to drink more or fewer cups of coffee per day on average, shaped partly by the aryl hydrocarbon receptor gene near this variant.

Your genotype: Not read (rs4410790)

Why we include this: the evidence

Source: GWAS Catalog · GCST001032, GCST002650
Significance: Genome-wide significant (p ≤ 5×10⁻⁸)
Replication: Cornelis et al. 2011 (PLoS Genet, GCST001032, N=47,431 Europeans) identified rs4410790 near AHR as a top locus for habitual caffeine intake (p=2e-19); Cornelis et al. 2015 (Mol Psychiatry, GCST002650, N=91,462+30,062 Europeans) confirmed the AHR locus for coffee cups per day. gnomADg:ALL MAF of T allele = 0.4586.
Where established: Characterised primarily in European-ancestry cohorts; the AHR-region signal for coffee intake has been observed in additional populations
Effect: The C allele is associated with higher habitual coffee intake (more cups per day); the T allele is associated with lower intake

rs4410790 lies near AHR (aryl hydrocarbon receptor), a transcription factor that induces CYP1A2 and other metabolic genes. The AHR pathway modulates caffeine clearance and the brain's response to coffee's aromatic compounds. Individuals with higher AHR-mediated metabolism clear caffeine faster, feel less stimulated per cup and tend to drink more. This is a behavioural tendency from a brain–metabolism interaction; habit, culture and stress all override the genetic signal.

Cornelis et al. 2011: GWAS of caffeine intake (GCST001032) · MedlinePlus Genetics: AHR gene

ACKR1 Duffy blood-group antigen (malaria resistance & neutrophil baseline)

Whether your red cells carry the Duffy antigen. The 'null' result is common in people of African genetic ancestry and brings two unrelated, well-established effects: resistance to Plasmodium vivax malaria, and a naturally lower normal neutrophil count.

Your genotype: Not read (rs2814778)

Why we include this: the evidence

Source: GWAS Catalog · GCST001302, GCST004620
Significance: Genome-wide significant (p ≤ 5×10⁻⁸)
Replication: A textbook, strongly replicated functional regulatory variant; confirmed in GCST001302 (Crosslin 2011, eMERGE leukocyte count, N=13,533, p=7e-55) and GCST004620 (Astle 2016, UK Biobank basophil+neutrophil count, N=170,143, p=2e-12)
Where established: The C (null) allele is common in people of West/Central African genetic ancestry and uncommon elsewhere. Ancestry is not race, and allele frequencies vary continuously across people. This result applies to anyone who carries the variant, however they identify.
Effect: C/C silences ACKR1 on red cells (Duffy-null), giving P. vivax resistance and a lower neutrophil baseline

The C allele disrupts a GATA-1 binding site, so the Duffy antigen/chemokine receptor is not made on red cells. Plasmodium vivax uses that receptor to invade red cells, so Duffy-null cells resist it. The same biology is linked to a constitutionally lower neutrophil count.

MedlinePlus Genetics: ACKR1 gene

ADGRG6 Attached or free earlobes (ADGRG6)

School biology often presents earlobe attachment as a one-gene, dominant-or-recessive trait. It is not: a 2017 study of nearly 75,000 people found 49 regions of the genome involved. This card reads one of the strongest, near the ADGRG6 gene.

Your genotype: Not read (rs58122955)

Why we include this: the evidence

Source: GWAS Catalog (EBI): GCST005193, GCST005192, Shaffer et al. 2017 (AJHG) · GCST005193, GCST005192
Significance: Genome-wide significant (p ≤ 5×10⁻⁸)
Replication: Found in a meta-analysis of three cohorts whose earlobes were scored by expert raters from photographs (1,791 European American, 5,062 Latin American and 2,857 Chinese participants; GCST005193, p = 2e-14), then replicated in an independent cohort of 64,950 European-ancestry 23andMe research participants who reported their own earlobes (odds ratio about 1.3 per A allele, p = 3e-87). GCST005192 is the combined analysis of all 74,660 people. The effect pointed the same way in all four cohorts.
Where established: Found across European American, Latin American and Chinese cohorts and replicated in European-ancestry participants; the effect was largest in the European American cohort and smaller, in the same direction, in the other two. The A allele is common in every reference population: about 26% of European chromosomes and 20% of Finnish ones. No African-ancestry cohort was studied.
Effect: Each copy of the A allele nudged earlobes toward being attached, with the same direction in all three expert-rated cohorts and an odds ratio of about 1.3 per copy in the independent replication cohort. It is one of 49 regions involved, so a single marker leaves most of the outcome to the others.

ADGRG6 (also called GPR126) encodes a cell-surface receptor involved in development; disrupting it in zebrafish causes a swollen inner ear, and the same region was earlier linked to earlobe size. The associated variant is intronic and no study has yet shown which variant in the region does the work. If it is this one, it most likely acts by changing how the gene is regulated while the outer ear forms, not by changing the protein.

Shaffer et al. 2017, American Journal of Human Genetics: Multiethnic GWAS reveals polygenic architecture of earlobe attachment (PMID 29198719) · Shaffer et al. 2017, free full text (PubMed Central PMC5812923) · GWAS Catalog: GCST005193 — lobe attachment (rater scored) · GWAS Catalog: GCST005192 — lobe attachment (rater-scored or self-reported) · dbSNP: rs58122955

HERC2 Eye colour (blue vs brown)

The main common genetic switch behind blue versus brown eyes, though eye colour is polygenic, so this is a strong hint, not a verdict.

Your genotype: Not read (rs12913832)

Why we include this: the evidence

Source: GWAS Catalog · GCST000685, GCST000710
Significance: Genome-wide significant (p ≤ 5×10⁻⁸)
Replication: The single most replicated eye-colour locus; explains a large share of blue/brown variation; independently confirmed in GCST000685 (Liu 2010, N=9,494) and GCST000710 (Eriksson 2010, N=9,126)
Where established: Best established in European-ancestry populations, where blue/brown variation is common
Effect: The G allele is recessive for blue; G/G is strongly associated with blue eyes, A/_ with brown

rs12913832 lies in an intron of HERC2 that regulates the neighbouring OCA2 pigment gene; the G allele lowers OCA2 expression in the iris, reducing melanin and favouring blue eyes.

MedlinePlus Genetics: Eye color

IRF4 Freckling tendency (IRF4)

A genetic tendency toward freckles and fair, sun-sensitive skin, shaped by a well-studied variant in an enhancer of the IRF4 gene in pigment cells.

Your genotype: Not read (rs12203592)

Why we include this: the evidence

Source: Functional variant: Praetorius et al. 2013 (Cell) · PMID:24267888
Significance: Below genome-wide significance
Replication: Praetorius et al. 2013 (Cell, PMID:24267888) showed in melanocytes, zebrafish and mice that rs12203592 lies within an enhancer of IRF4: the T allele impairs binding of TFAP2A, which with the melanocyte master regulator MITF drives the enhancer, and IRF4 cooperates with MITF to activate tyrosinase (TYR). The variant's association with freckles, sun sensitivity, blue eyes and brown hair was first reported in large pigmentation GWAS (Han 2008; Eriksson 2010) and the mechanism was then established functionally in vivo.
Where established: Established mainly in European-ancestry cohorts; the T allele and freckling are most common in lighter-skinned populations. Freckling also depends heavily on sun exposure and age.
Effect: The T allele is associated with more freckling and fairer, more sun-sensitive skin; the C allele with less.

rs12203592 sits in a melanocyte enhancer in intron 4 of IRF4. The T allele weakens TFAP2A binding, reducing IRF4 expression; IRF4 normally cooperates with MITF to switch on tyrosinase (TYR), the rate-limiting melanin-synthesis enzyme. Less IRF4 means less even melanin production, which presents as freckles and lighter, sun-sensitive skin (Praetorius et al. 2013).

Praetorius et al. 2013, Cell: A polymorphism in IRF4 affects human pigmentation · MedlinePlus Genetics: IRF4 gene

FUT2 FUT2 secretor status (norovirus resistance, gut microbiome, B12)

Whether you secrete blood-group sugars into saliva, the gut and other body fluids. Non-secretors resist the dominant strains of norovirus and tend to have a different gut microbiome and higher serum vitamin B12.

Your genotype: Not read (rs601338)

Why we include this: the evidence

Source: Functional variant: Carlsson et al. 2009 (PLoS One) · PMID:19440360
Significance: Below genome-wide significance
Replication: The FUT2 G428A (c.461G>A, p.Trp154Ter) nonsense variant has been studied extensively; homozygous carriers lack FUT2 enzyme activity and are non-secretors. Carlsson 2009 confirmed protection against symptomatic norovirus GII.4 in a clinical cohort. The A allele frequency in gnomADg:nfe = 0.4776, confirming it as a common polymorphism in Europeans.
Where established: rs601338 is the common non-secretor allele in people of European genetic ancestry. Other populations carry different non-secretor alleles this single marker does not capture (for example se357 in East Asia), so a 'secretor' call from this marker alone is not a global determination. Ancestry is not race; this applies to anyone carrying the variant.
Effect: A/A removes FUT2 enzyme activity → non-secretor

FUT2 adds blood-group sugars to mucosal surfaces. The A allele (Trp154Ter stop codon) inactivates the enzyme, so A/A people are non-secretors. Many gut pathogens, notably norovirus GII.4, need those sugars to attach, so non-secretors resist them.

MedlinePlus Genetics: FUT2 gene

KITLG Hair colour shade (KITLG)

A genetic tendency toward lighter (blond) versus darker (brown or black) hair, shaped partly by a regulatory variant near the KITLG gene on chromosome 12.

Your genotype: Not read (rs12821256)

Why we include this: the evidence

Source: GWAS Catalog · GCST006988, GCST006989
Significance: Genome-wide significant (p ≤ 5×10⁻⁸)
Replication: Morgan et al. 2018 (Nat Commun, PMID:30531825) genome-wide study of hair colour in 323,317 UK Biobank participants: rs12821256-C associated with blond vs brown/black (GCST006988, p=4×10^-30); same SNP replicated in brown vs black comparison (GCST006989, N=283,920). The KITLG locus is one of several replicated hair-colour loci alongside MC1R, HERC2/OCA2, and IRF4. gnomADg:ALL MAF for the C (blond) allele = 0.0685.
Where established: Best characterised in European-ancestry cohorts; blond hair is most common in Northern European populations
Effect: The C allele (minor allele, ~7% global frequency) is associated with lighter, blond hair; the T allele (reference, ~93% frequency) is associated with darker brown or black hair

rs12821256 lies in a regulatory region upstream of KITLG (also known as SCF, stem cell factor), which encodes a ligand that supports melanocyte survival and migration. The C allele is thought to reduce KITLG expression in hair follicles, lowering melanin production and favouring blond colouring. Hair colour is highly polygenic; MC1R, HERC2/OCA2, IRF4 and dozens of other loci jointly determine the specific shade.

Morgan et al. 2018: Hair colour GWAS in UK Biobank (Nat Commun) · MedlinePlus Genetics: Hair color

EDAR Hair thickness & shape

A tendency toward thicker, straighter hair strands, influenced by a well-studied variant in EDAR.

Your genotype: Not read (rs3827760)

Why we include this: the evidence

Source: Functional variant: Kamberov et al. 2013 (Cell) · PMID:23415220
Significance: Below genome-wide significance
Replication: EDAR 370A (rs3827760) is the classic functional hair-morphology variant: identified for East-Asian hair thickness (Fujimoto et al. 2008) and shown causal in a mouse model expressing the selected EDAR allele (Kamberov et al. 2013, Cell). The derived G allele is at high frequency in East-Asian and Native-American ancestry.
Where established: The G (derived) allele is common in East-Asian and Native-American ancestry and rare in European/African ancestry. Read this trait with ancestry in mind
Effect: The G (370Ala) allele is associated with thicker, straighter hair fibres and shovel-shaped incisors

rs3827760 changes an amino acid in EDAR, a receptor guiding the development of hair, teeth and sweat glands; the derived allele increases EDAR signalling.

MedlinePlus Genetics: EDAR gene

TCHH Straight vs curly hair (TCHH)

Whether your hair tends to grow straight or to wave and curl, read at the trichohyalin gene — the locus that explains more of the straight-hair difference among Europeans than any other common variant.

Your genotype: Not read (rs11803731)

Why we include this: the evidence

Source: GWAS Catalog (EBI): GCST005191, Liu et al. meta-analysis (Human Molecular Genetics) · GCST005191
Significance: Genome-wide significant (p ≤ 5×10⁻⁸)
Replication: GCST005191 is a meta-analysis with a discovery sample of 16,763 European-ancestry individuals and a separate replication sample of 12,201 people spanning four independent cohorts — 2,340 European, 2,899 Han Chinese, 709 Uyghur and 6,238 Latin American. The locus was first reported by Medland et al. 2009 (AJHG, PMID:19896111) in Australian and Dutch samples, so the association is replicated both within GCST005191 and across separate publications.
Where established: Discovered and best characterised in European-ancestry samples, then replicated in Han Chinese, Uyghur and Latin American cohorts. The T allele is common in Europeans (0.23) and very rare in African and East Asian reference populations, so this variant explains hair-shape differences within Europe rather than between continents — East Asian straight hair is largely a different locus (EDAR).
Effect: The T allele is associated with straighter hair, the A allele with more wave and curl. It is the largest single common-variant effect on hair shape in Europeans, and still only part of the picture.

TCHH encodes trichohyalin, a structural protein of the inner root sheath of the hair follicle. Trichohyalin is cross-linked by transglutaminases and helps set the mechanical shape of the hair shaft as it hardens; variation at this locus changes the follicle's cross-sectional asymmetry, and an asymmetric follicle extrudes a curved — curlier — shaft.

GWAS Catalog: GCST005191 — strand of hair shape · Liu et al., Human Molecular Genetics: meta-analysis identifying 8 novel loci for head hair shape (PMID:29220522) · Medland et al. 2009, AJHG: Common variants in the trichohyalin gene are associated with straight hair in Europeans (PMID:19896111)

TWIST2 Androgenetic hair thinning tendency (TWIST2)

A genetic tendency toward androgenetic hair thinning (the most common form of hair loss), shaped partly by a regulatory variant near the TWIST2 gene. This reflects a cosmetic tendency, not a disease prediction.

Your genotype: Not read (rs11684254)

Why we include this: the evidence

Source: GWAS Catalog · GCST007020, GCST006661
Significance: Genome-wide significant (p ≤ 5×10⁻⁸)
Replication: Yap et al. 2018 (Nat Commun, PMID:30573740, GCST007020, N=205,327 European males) identified rs11684254-G as one of the strongest genome-wide signals for male-pattern baldness (p≈2×10^-308, beta=increase); independently confirmed in Hagenaars et al. 2017 (PLoS Genet, PMID:28196072, GCST006661, N=52,874 British males, p≈1×10^-40, beta=increase). Also replicated in Hillmer 2018 (GCST007815, N=22,518, p=1×10^-39, β=0.29). The TWIST2 locus is one of the most consistently replicated autosomal loci for androgenetic alopecia. gnomADg:ALL MAF for the G allele = 0.4396.
Where established: Best characterised in European-ancestry males; androgenetic alopecia is universal but prevalence and genetic architecture vary by ancestry
Effect: The G allele (rs11684254-G, ~44% global frequency) is associated with greater tendency to develop androgenetic hair thinning; the C allele is associated with lower tendency at this locus

rs11684254 lies upstream of TWIST2 (Twist Family BHLH Transcription Factor 2), a regulatory transcription factor involved in follicular development and stem cell differentiation. The G allele is thought to alter TWIST2 expression in hair follicles, influencing the miniaturisation process characteristic of androgenetic alopecia. The androgen-receptor pathway (driven by dihydrotestosterone) is the main driver; the TWIST2 locus likely modulates follicle sensitivity. Hair loss is highly polygenic (dozens of loci contribute alongside this one), and lifestyle, age, and androgens strongly mediate the trait.

Yap et al. 2018: Dissection of male pattern baldness (Nat Commun) · Hagenaars et al. 2017: Genetic prediction of male pattern baldness (PLoS Genet)

FADS1 Omega fatty acid desaturation (FADS1)

How efficiently your body converts short-chain omega-6 and omega-3 fats from food into the longer-chain forms (arachidonic acid, EPA and DHA) that your cells use directly.

Your genotype: Not read (rs174537)

Why we include this: the evidence

Source: GWAS Catalog · GCST002721, GCST90060989
Significance: Genome-wide significant (p ≤ 5×10⁻⁸)
Replication: Mozaffarian et al. 2013 (PLoS Genet, GCST002721, N=10,421 multi-ancestry) found rs174537-T associated with trans-18:2 (linoleic acid isomer) measurement; a second study (GCST90060989, N=5,662 Pakistani) confirmed FADS1-region association with phosphatidylserine levels (p=4e-20). The FADS1 locus is the most replicated genetic locus for polyunsaturated fatty acid metabolism across dozens of independent cohorts. gnomADg:ALL MAF of T allele = 0.2888.
Where established: Effect replicated across European, African, Asian and mixed-ancestry cohorts; T allele is common in all major populations
Effect: The T allele is associated with altered FADS1 desaturase activity and higher proportions of 18-carbon fatty acid precursors relative to long-chain products (lower elongation/desaturation efficiency); G allele is associated with more efficient conversion

rs174537 is the lead tagging SNP for the FADS1–FADS2 gene cluster on chromosome 11. FADS1 (delta-5 desaturase) and FADS2 (delta-6 desaturase) catalyse the rate-limiting steps converting linoleic acid (18:2n-6) to arachidonic acid (20:4n-6), and alpha-linolenic acid (18:3n-3) to EPA (20:5n-3) and DHA (22:6n-3). The T-allele haplotype is associated with reduced desaturase activity, so T/T individuals tend to have lower arachidonic acid and long-chain n-3 PUFA and higher precursor fatty acids from diet. This affects inflammatory signalling, membrane composition and cardiovascular markers, all tendencies, not diagnoses.

MedlinePlus Genetics: FADS1 gene · Mozaffarian et al. 2013: GWAS of circulating fatty acids (GCST002721)

ZEB2 region Photic sneeze reflex (ACHOO)

Whether bright light tends to make you sneeze: the so-called photic sneeze reflex.

Your genotype: Not read (rs10427255)

Why we include this: the evidence

Source: GWAS Catalog · GCST000706, GCST007687
Significance: Genome-wide significant (p ≤ 5×10⁻⁸)
Replication: Identified and independently replicated in GCST000706 (Eriksson 2010, 23andMe, N=9,126, European) and GCST007687 (Chinese ancestry GWAS, N=3,417, p=6e-20)
Where established: Established mainly in European-ancestry cohorts; independently replicated in Chinese cohort
Effect: The C allele is associated with a higher chance of sneezing in response to bright light

rs10427255 sits near ZEB2; the reflex is thought to involve cross-wiring of visual and sneeze pathways, though the exact biology is still being studied.

GWAS Catalog: photic sneeze reflex

MC1R Red hair and fair skin tendency (MC1R)

A genetic tendency toward red or auburn hair colouring and fair skin, shaped by two classic MC1R loss-of-function variants known as 'R' alleles. Carrying one or two R alleles shifts the hair follicle's pigment balance toward warm, reddish tones and increases freckling tendency.

Your genotype: Not read (rs1805007)

Why we include this: the evidence

Source: Functional variant: Valverde et al. 1995 (Nature Genetics) · PMID:7581459
Significance: Below genome-wide significance
Replication: Valverde et al. 1995 (Nat Genet, PMID:7581459) identified MC1R R allele variants in >80% of individuals with red hair or fair non-tanning skin but <20% of dark-haired individuals, establishing these as the primary common genetic determinants of red pigmentation. Hundreds of subsequent studies across diverse cohorts confirm R151C (rs1805007) and R160W (rs1805008) as the two most common and most penetrant red-hair R alleles. The GWAS Catalog documents rs1805007-T with OR=12.47 for hair colour and OR=4.37 for freckles (p values ≤ 2×10^-55). gnomADg:ALL MAF for rs1805007-T ≈ 4.6%; rs1805008-T ≈ 4.6%.
Where established: R alleles are common in populations of Northern European ancestry, especially Irish, Scottish, and British; rare in East Asian and sub-Saharan African populations
Effect: Each T allele (R151C or R160W) reduces MC1R signalling, shifting pigment production from dark eumelanin toward reddish pheomelanin; carrying one or two R alleles raises the probability of red or auburn hair and fair skin with freckling tendency

MC1R encodes the melanocortin-1 receptor, which controls the eumelanin/pheomelanin switch in melanocytes. Binding of α-MSH activates the receptor, favouring dark eumelanin. The R151C and R160W substitutions impair receptor function, tipping the balance toward reddish pheomelanin. The result is a lighter hair and skin phenotype with increased sensitivity to UV and a tendency for warm hair tones. Hair colour is polygenic: HERC2/OCA2, IRF4, KITLG and other loci also contribute, so these two variants are a strong signal, not a complete predictor.

Valverde et al. 1995: MC1R variants and red hair (Nature Genetics) · MedlinePlus Genetics: Hair color

PAX8 Sleep duration (PAX8)

How long your body tends to sleep, read at PAX8 — the most consistently replicated common-variant signal for sleep duration. Set expectations before you read it: the whole distance between the two extreme genotypes is about five minutes a night.

Your genotype: Not read (rs62158211)

Why we include this: the evidence

Source: GWAS Catalog (EBI): GCST003839, Jones et al., sleep duration (PLoS Genetics) · GCST003839
Significance: Genome-wide significant (p ≤ 5×10⁻⁸)
Replication: GCST003839 pairs a discovery sample of 127,573 British individuals with a separate replication sample of 47,180 — the two independent cohorts criterion 3 asks for. The locus then reappears in Lane et al. 2016 (GCST003980, T +0.039 h, p=5e-14), Hammerschlag et al. 2017 (GCST004694, p=1e-12) and Doherty et al. 2018 (GCST006914, p=6e-17), with the direction identical every time. Those are corroborations rather than independent samples — all of them are U.K. cohorts drawing on substantially the same UK Biobank participants, so they should not be counted as extra cohorts. The one that genuinely adds something is Doherty 2018, which measured sleep with a wrist accelerometer instead of a questionnaire: the same signal in the same people through a different instrument, which is what makes self-report bias an unlikely explanation for it.
Where established: Discovered and replicated almost entirely in British and other European-ancestry samples, so the effect size outside Europe is not well established. The T allele is common in every reference population but varies: 0.2107 in the 1000 Genomes European panel and 0.1768 in the Finnish one, against 0.1230 in East Asian and 0.0764 in African panels. Because the whole effect is a few minutes either way, this is a variant that describes a population average rather than something that would explain why one person needs more sleep than another.
Effect: Each copy of the G allele is associated with about 0.039 hours — roughly 2.3 minutes — less sleep per night (SE 0.005), so the gap between the two homozygous genotypes is on the order of five minutes. The independently measured estimates agree closely: Lane et al. report +0.039 hours per T allele, and Doherty et al. find the same direction using accelerometer-derived sleep. Sleep duration is highly polygenic and dominated by behaviour, light exposure, age and health, so this locus is a real signal that explains a very small share of why people differ.

Honestly unresolved, and worth saying so. PAX8 encodes a paired-box transcription factor best known for driving thyroid and kidney development, and rs62158211 is an intronic variant rather than a coding change — so there is no demonstrated causal chain from this base to a night's sleep. The most cited plausible route is thyroid: PAX8 is essential for thyroid follicular cells, and thyroid hormone measurably shapes sleep architecture and sleep need. That remains a hypothesis. What earns this locus its place under §4 criterion 6, which asks for a plausible mechanism or a canonical well-characterised locus, is the second branch: PAX8 is the single most reproducible common-variant sleep-duration signal in the literature, not an isolated statistical tag.

GWAS Catalog: GCST003839 — sleep duration · Jones et al., PLoS Genetics: GWAS analyses identify new morningness and sleep duration loci (PMID:27494321) · Doherty et al., Nature Communications: GWAS of accelerometer-derived sleep measures (PMID:30531941) · dbSNP: rs62158211

SLC45A2 Tan or burn: tanning response (SLC45A2)

Whether your skin answers strong sun with a tan or with a burn, read at SLC45A2 — the pigment gene that carries one of the strongest signals of natural selection in the European genome. Most European genomes give the same answer at this variant, and it is the answer where sun buys the most damage for the least tan.

Your genotype: Not read (rs16891982)

Why we include this: the evidence

Source: GWAS Catalog (EBI): GCST005897, Visconti et al., tanning response to sun exposure (Nature Communications) · GCST005897
Significance: Genome-wide significant (p ≤ 5×10⁻⁸)
Replication: GCST005897 has a U.K. discovery sample of 121,296 people (46,768 low-tanning cases against 74,528 moderate- and high-tanning controls) and a separate replication sample of 55,382 (15,547 cases, 39,835 controls) recruited across the United States, Australia, the Netherlands and the United Kingdom — two independent cohorts, 176,678 people in total. SLC45A2 Phe374Leu is additionally one of the most heavily replicated pigmentation variants in the literature, reported across European, East Asian, South Asian and South American samples since the mid-2000s, and it is a textbook signal of positive selection in West Eurasia.
Where established: Discovered and characterised in European-ancestry samples, and the frequency picture is the important caveat: the light G allele is close to fixed in Europe (0.9384 in the 1000 Genomes European panel, 0.9596 in the Finnish one) and close to absent in the East Asian (0.0060) and African (0.0356) panels. This variant therefore separates Europeans from other continental groups far more sharply than it separates Europeans from each other. For a Finnish reader the practical consequence is that roughly 92% of Finnish genomes land on the same genotype here, so this trait describes a population-wide setting more often than a personal difference.
Effect: The C allele (Leu374) is associated with a better tanning response; the G allele (Phe374) with skin that reddens and burns instead of browning. In Visconti et al. the C allele lowers the odds of the study's low-tan-response phenotype (beta 0.918, standard error 0.173, p = 2e-176). Tanning ability is polygenic — MC1R, IRF4, HERC2/OCA2 and TYR all contribute — so this is the single largest common-variant dial rather than the whole mechanism.

SLC45A2 encodes a transporter in the melanosome membrane that helps hold melanosomal pH in the narrow range where tyrosinase works. Tyrosinase is the rate-limiting enzyme of melanin synthesis, and it is sharply pH-sensitive. The 374Phe form regulates that pH less effectively, so tyrosinase runs slower, less dark eumelanin accumulates in the melanosome, and the melanocyte hands the surrounding keratinocytes lighter, less UV-absorbent pigment. Less eumelanin means less of the shielding that converts a UV dose into a tan rather than into an inflammatory sunburn.

GWAS Catalog: GCST005897 — low tan response · Visconti et al., Nature Communications: GWAS in 176,678 Europeans reveals genetic loci for tanning response to sun exposure (PMID:29739929) · dbSNP: rs16891982 (SLC45A2 Phe374Leu) · Evolution of skin-pigmentation variation in West Eurasia (PNAS)

FUT2 Vitamin B12 tendency (FUT2)

A genetic influence on circulating vitamin B12, linked to the FUT2 gene that also governs secretor status: non-secretors tend to have higher serum B12.

Your genotype: Not read (rs602662)

Why we include this: the evidence

Source: GWAS Catalog · GCST90277442, GCST000358
Significance: Genome-wide significant (p ≤ 5×10⁻⁸)
Replication: Grarup et al. 2013 (PLoS Genet, GCST90277442, N=38,229 Danish/Icelandic) confirmed rs602662-A as the top FUT2-region hit for vitamin B12 (p=2e-139, increase); Hazra et al. 2009 (Nat Genet, GCST000358, N=3,620 Europeans) first reported this association (p=3e-20, beta=+49.77 pg/ml). The FUT2 locus is among the most robustly replicated B12 loci. gnomADg:ALL MAF of A allele = 0.4598.
Where established: Well-characterised in European and Scandinavian ancestry cohorts; the A allele frequency varies by population
Effect: The A allele is associated with higher serum vitamin B12; G allele (non-secretor haplotype region) is associated with lower circulating B12

rs602662 tags the FUT2 non-secretor haplotype. FUT2 adds fucose sugars to gut mucosal surfaces; non-secretor status (A allele) alters the intestinal microbiome and gut physiology in ways that increase vitamin B12 absorption or reduce bacterial competition for the vitamin. The FUT2 locus is distinct from intrinsic-factor deficiency and pernicious anaemia. This is a common-variant modulation of B12 status within the normal range.

Grarup et al. 2013: Vitamin B12 and folate GWAS (GCST90277442) · MedlinePlus Genetics: FUT2 gene

GC Vitamin D tendency (GC / VDBP)

A genetic nudge toward lower or higher circulating vitamin D, set partly by how efficiently your body carries the vitamin through your bloodstream.

Your genotype: Not read (rs2282679)

Why we include this: the evidence

Source: GWAS Catalog · GCST000664, GCST005367
Significance: Genome-wide significant (p ≤ 5×10⁻⁸)
Replication: Wang et al. 2010 (Lancet, GCST000664, N=6,722 Europeans) identified rs2282679 in GC as a top locus for 25-hydroxyvitamin D; Manousaki et al. 2017 (PLoS Genet, GCST005367, N=79,366 initial + 42,757 replication) confirmed this as one of the strongest common vitamin D loci. gnomADg:ALL MAF of G allele = 0.2176.
Where established: Effect well characterised in European-ancestry cohorts; replicated in other ancestries but effect size varies
Effect: The G (minor) allele at rs2282679 is associated with lower 25-hydroxyvitamin D levels; T/T carriers tend toward higher circulating vitamin D

rs2282679 lies in the GC gene, which encodes the vitamin D-binding protein (VDBP / DBP) that transports 25-hydroxyvitamin D and 1,25-dihydroxyvitamin D through the circulation. The G allele is associated with a structural variant that reduces the binding efficiency or circulating concentration of VDBP, lowering transported vitamin D. This is a tendency: sun exposure, diet and supplementation remain the dominant determinants of actual status.

MedlinePlus Genetics: GC gene (vitamin D binding protein) · Wang et al. 2010: Common genetic determinants of vitamin D insufficiency (GCST000664)

Traits we looked at and decided not to report (5)

Other services report these. We assessed each against the same evidence bar the traits above clear, and it didn’t hold up, so the reason is here rather than the trait.

Sweet preference · Hwang LD et al., Am J Clin Nutr 2019 (PMID:31005972)
The largest study of sweet taste found only one genome-wide-significant signal, and it is not a taste gene. The hit is rs11642841 in FTO (P = 3.8e-8), a variant best known for its association with body weight, and it was found for how much sugar people eat rather than how sweet they find it. The sweet-taste receptor genes themselves (TAS1R2, TAS1R3, GNAT3) had, in the authors' own words, limited support across all three European samples, and every sweet-perception result was below genome-wide significance. Reporting this would mean relabelling a body-weight variant as a taste result.
Motion sickness · Hromatka BS et al., Hum Mol Genet 2015 (PMID:25628336)
The evidence is statistically strong but has never been independently repeated. The one study surveyed 80,494 people about car sickness and found 35 variants past the genome-wide threshold, the strongest at P = 4e-44 — but it describes itself as the first such study, it draws on a single customer database, and the trait is self-reported. Our gate asks for at least two independent groups of people before we report a result, and that is the criterion we will not bend: a finding from one cohort that no one else has reproduced is a promising lead, not a result about you.
Hangover response
There is no study to assess. The GWAS Catalog holds no genome-wide association study for hangover under any name, so there is no published, replicated evidence linking specific variants to how badly a person suffers after drinking. What we can say about alcohol from your file we already say: the alcohol-flush response (ALDH2) and alcohol metabolism (ADH1B) are both reported, and they are the parts with real evidence behind them.
HIV resistance (CCR5-delta-32) · Samson M et al., Nature 1996 (PMID:8898206); Liu R et al., Cell 1996 (PMID:8756719)
The biology here is real and unusually well understood: people who carry two copies of a 32-base-pair deletion in CCR5 lack the surface receptor that the common strain of HIV uses to enter cells, and are strongly protected against infection by it. We still do not report it as a trait, and the reason is that we cannot read it directly. The deletion is not a single-letter change, and a standard variant file does not call it. What we would actually be reading is a nearby single-letter marker that usually travels with the deletion, and then inferring the deletion from that. The inference is right most of the time and wrong some of the time, and nothing in your file tells you which case you are in. Our gate asks that a trait be callable directly from your data, and for a result this consequential we would rather report nothing than report a guess. The marker remains visible in the Frontier tier, where the inference is labelled as an inference.
Unibrow (eyebrow convergence)
There is no catalogued study to assess. The marker consumer sites cite for this trait, rs2395845, is not in the GWAS Catalog at all — the catalog has no record of it, so there are no published association statistics for us to check it against. The underlying claim traces to a single study of facial features in a Latin American cohort, which is interesting work, but it is one cohort, and our gate asks for at least two independent ones before we describe a result as being about you. With no catalog entry and no replication, there is nothing here we can stand behind.

ABO blood type

Likely · not a blood test

Your likely ABO blood group, inferred from three positions in the ABO gene. This is genetics, not a blood test. It is a strong indication for most people but not a clinical result.

ABOgroup not determinedTier 2 · Well-supported

Your blood group couldn't be confidently inferred: the defining markers were missing, inconsistent, or carried a rare variant this simple three-SNP method doesn't resolve. A blood test is the reliable answer.

Important: This is a likely blood group inferred from DNA, not a measured one. Rare ABO subgroups, the cis-AB and weak-A/B variants, and the messy O deletion mean genetics can occasionally disagree with a real blood test, so confirm with an actual blood test, and never use this for transfusion or any medical decision.

Your file's blood-group markers were never examined: the ABO group is read from three positions, one of which is a small insertion and another a C/G change whose two strands can't be told apart — neither can be typed from a genotyping array. This is "not examined", not a result, and it is the reason no group is shown. A blood test is the reliable answer, and a whole-genome file would let us infer it.

2026-06-07 · NCBI dbSNP: rs8176719 (ABO 261delG) · Yamamoto: Molecular genetics of the ABO blood group system (Annals of Blood)

Polygenic scores

Tendency · not a diagnosis

Not reportable Abdominal aortic aneurysm 29-variant score (Klarin et al., Circulation 2020) Tier 2 · Well-supported

Examined, not reportable: only 1 of 29 scoring variants (3%) were found in your file, below the 50% we require to place you on this score’s distribution. This is a coverage gap, not a result.

A polygenic score adds up many small-effect variants into a single statistical tendency. It is not a measurement, not a prediction of whether you will develop the condition, and not a diagnosis. Most of the risk for a complex disease comes from the rest of your genome, your environment, and chance — none of which this number captures.

2026-06-14 · PGS Catalog PGS000753 (restated; analytic European reference distribution under allele-frequency assumptions)

Not reportable Atrial fibrillation 142-variant score (Frederiksen et al., Heart 2023) Tier 2 · Well-supported

Examined, not reportable: only 0 of 142 scoring variants (0%) were found in your file, below the 50% we require to place you on this score’s distribution. This is a coverage gap, not a result.

2026-06-14 · PGS Catalog PGS005159 (restated; analytic European reference distribution under allele-frequency assumptions)

Not reportable Breast cancer 65-variant score (Zhang et al., PLoS Med 2018) Tier 2 · Well-supported

Examined, not reportable: only 0 of 65 scoring variants (0%) were found in your file, below the 50% we require to place you on this score’s distribution. This is a coverage gap, not a result.

2026-06-14 · PGS Catalog PGS000051 (restated; analytic European reference distribution under allele-frequency assumptions)

Not reportable Coronary artery disease 241-variant primary-prevention score (Marston et al., JAMA Cardiology 2023) Tier 2 · Well-supported

Examined, not reportable: only 2 of 241 scoring variants (1%) were found in your file, below the 50% we require to place you on this score’s distribution. This is a coverage gap, not a result.

2026-06-14 · PGS Catalog PGS003438 (restated; analytic European reference distribution under allele-frequency assumptions)

Not reportable Colorectal cancer 95-variant score (Huyghe et al., Nature Genetics 2018) Tier 2 · Well-supported

Examined, not reportable: only 0 of 95 scoring variants (0%) were found in your file, below the 50% we require to place you on this score’s distribution. This is a coverage gap, not a result.

2026-06-14 · PGS Catalog PGS000765 (restated; analytic European reference distribution under allele-frequency assumptions)

Not reportable Gout (serum-urate score) 114-variant serum-urate score (Tin et al., Nature Genetics 2019) Tier 2 · Well-supported

Examined, not reportable: only 0 of 114 scoring variants (0%) were found in your file, below the 50% we require to place you on this score’s distribution. This is a coverage gap, not a result.

2026-06-14 · PGS Catalog PGS000126 (restated; analytic European reference distribution under allele-frequency assumptions)

Not reportable Kidney cancer 105-variant score (Purdue et al., Nature Genetics 2024; its 2 X-chromosome variants left out) Tier 2 · Well-supported

Examined, not reportable: only 0 of 105 scoring variants (0%) were found in your file, below the 50% we require to place you on this score’s distribution. This is a coverage gap, not a result.

2026-06-14 · PGS Catalog PGS004908 (restated; analytic European reference distribution under allele-frequency assumptions)

Not reportable LDL cholesterol (polygenic score) 223-variant LDL-cholesterol score (Trinder et al., JAMA Cardiology 2020) Tier 2 · Well-supported

Examined, not reportable: only 0 of 223 scoring variants (0%) were found in your file, below the 50% we require to place you on this score’s distribution. This is a coverage gap, not a result.

2026-06-14 · PGS Catalog PGS000115 (restated; analytic European reference distribution under allele-frequency assumptions)

Not reportable Prostate cancer 128-variant score (Jia et al., JNCI Cancer Spectrum 2020) Tier 2 · Well-supported

Examined, not reportable: only 0 of 128 scoring variants (0%) were found in your file, below the 50% we require to place you on this score’s distribution. This is a coverage gap, not a result.

2026-06-14 · PGS Catalog PGS000719 (restated; analytic European reference distribution under allele-frequency assumptions)

Not reportable Type 2 diabetes 47-variant European GWAS-significant SNP score (Liu et al., 2023) Tier 2 · Well-supported

Examined, not reportable: only 0 of 47 scoring variants (0%) were found in your file, below the 50% we require to place you on this score’s distribution. This is a coverage gap, not a result.

2026-06-14 · PGS Catalog PGS004226 (restated; analytic European reference distribution under allele-frequency assumptions)

Not reportable Venous thromboembolism 297-variant score (Klarin et al., Nature Genetics 2019) Tier 2 · Well-supported

Examined, not reportable: only 0 of 297 scoring variants (0%) were found in your file, below the 50% we require to place you on this score’s distribution. This is a coverage gap, not a result.

2026-06-14 · PGS Catalog PGS000043 (restated; analytic European reference distribution under allele-frequency assumptions)

credibility firewall · exploratory below

Tier 3 · Emerging / Speculative

Frontier

Emerging research

Read from your file: 10 of 10. The other 0 had no variant reported in your file, so they’re read as the common (reference) genotype: an inference, not a measurement (see each card’s note).

ACTN3 ACTN3: fast-twitch 'sports gene' (sprint/power vs endurance) Endurance-leaning (X/X) Replicated · small effect

Whether your fast-twitch muscle fibres make alpha-actinin-3. The R577X variant is the most-studied 'athlete gene': real on average across populations, but a tiny influence on any one person next to training.

Your genotype: T/T (Fitness & performance)
Most-associated reading: Endurance-leaning (X/X)

No functional alpha-actinin-3 (~18% of people): fast-twitch fibres lean slightly toward endurance-type properties. Associated on average with a touch less raw power and is more common in some endurance athletes, but plenty of strong, fast X/X people exist.

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: PubMed: ACTN3 elite-athlete cohorts (Yang et al. 2003 and replications)
Effect: The X (T) allele is a stop variant: X/X people make no functional alpha-actinin-3 in fast-twitch fibres. R/R is over-represented among elite sprint/power athletes and X/X among some endurance athletes, but the average group difference is small and easily swamped by training, body type and other genes.

Alpha-actinin-3 is a structural protein in fast-twitch (type II) muscle fibres. R577X (rs1815739) truncates it, so the ~18% of people who are X/X produce none and compensate with alpha-actinin-2: a subtle shift toward endurance-type fibre properties, not a deficit.

This is the most replicated result in sports genetics, yet it explains only a sliver of performance variance: training, recovery and dozens of other variants dominate. It does not pick your sport.

What this isn’t: Not a talent test, not a training prescription, and not a ceiling: X/X world-class sprinters and R/R endurance champions both exist.

MedlinePlus Genetics: ACTN3 gene · dbSNP: rs1815739

ADORA2A ADORA2A: caffeine sensitivity & sleep Intermediate Mixed evidence

The adenosine A2A receptor that caffeine blocks. A variant here is linked to how wired or anxious caffeine makes you, and how much it fragments your sleep.

Your genotype: T/C (Sleep & chronotype)
Most-associated reading: Intermediate

One of each allele: an intermediate, and the most common, profile.

How seriously to take this: the evidence

Strength: Mixed evidence
Source of the claim: PubMed: caffeine-challenge & sleep studies
Effect: One genotype at rs5751876 is associated on average with more caffeine-induced anxiety and lighter, more disrupted sleep after caffeine. Habitual intake and tolerance shift this strongly.

Caffeine works largely by blocking adenosine A2A receptors (encoded by ADORA2A). Variation in the receptor gene alters individual sensitivity to that blockade, including its anxiogenic and sleep-fragmenting effects.

Caffeine response is shaped at least as much by CYP1A2 metabolism (see your Traits report), tolerance, dose and timing as by this single receptor variant.

What this isn’t: Not medical guidance on caffeine, and not a diagnosis of anxiety or insomnia.

MedlinePlus Genetics: ADORA2A gene · dbSNP: rs5751876

ALDH2 Alcohol flush: acetaldehyde clearance (ALDH2) Reduced clearance, flush-prone Established biology · inferred call

Whether alcohol makes you flush, feel queasy and get a racing heart from even a small drink. One of the best-characterised functional variants in the human genome tilts how fast your body clears acetaldehyde, the toxic first breakdown product of alcohol.

Your genotype: G/A (Substance response)
Most-associated reading: Reduced clearance, flush-prone

One variant copy already cripples most enzyme activity; carriers commonly flush and tolerate alcohol poorly.

How seriously to take this: the evidence

Strength: Established biology · inferred call
Source of the claim: Decades of biochemical work + large East-Asian-ancestry cohorts (PubMed): one of the best-characterised functional human variants
Effect: The A allele (ALDH2*2, a lysine substitution) cripples the enzyme that clears acetaldehyde, so it builds up after drinking: the cause of facial flushing, nausea and a racing heart. A single A copy already produces a large effect because the variant subunit poisons the whole enzyme complex; two copies is more pronounced. It is strongly protective against heavy drinking, and raises upper-aerodigestive cancer risk in people who drink despite the reaction.

ALDH2 is the mitochondrial enzyme that turns acetaldehyde (the immediate product of ethanol breakdown) into harmless acetate. The Glu504Lys substitution sits at the subunit interface and acts dominant-negatively, so even heterozygotes lose most enzyme activity and accumulate acetaldehyde: a known irritant and carcinogen.

The flush is a real biological signal, but its intensity varies a lot between people and with how much you drink: the genotype tells you the tendency, not exactly how you will feel.

What this isn’t: Not a diagnosis of any disease, and not on its own a prediction that you will or won't develop cancer or an alcohol problem, and never a green light to 'drink through' the reaction.

dbSNP: rs671 · Glu504Lys of ALDH2 and risk of human diseases (review, PMC4600480)

BDNF BDNF: Val66Met, learning & neuroplasticity One Met allele Mixed evidence

A growth factor that supports learning, memory and exercise-driven brain plasticity. The Val66Met variant subtly changes how much BDNF is released on demand. Its memory effects are small and inconsistent; its clearest genome-wide links are with body weight and with taking up smoking.

Your genotype: C/T (Cognition & stress)
Most-associated reading: One Met allele

One Met allele: slightly altered activity-dependent BDNF release; population studies show only small, inconsistent average differences on memory or plasticity measures. In genome-wide studies it goes with a slightly lower body mass index and a slightly lower chance of taking up smoking.

How seriously to take this: the evidence

Strength: Mixed evidence
Source of the claim: Neuroimaging and cognition cohorts (inconsistent); genome-wide studies of body mass index (Yengo et al. 2018) and smoking initiation (Saunders et al. 2022)
Effect: The Met (T) allele is associated with somewhat reduced activity-dependent BDNF secretion and, on average, small differences in episodic memory and exercise-induced plasticity. Findings are inconsistent across populations. What has replicated at genome-wide scale lies elsewhere: each Met allele goes with a slightly lower body mass index and a slightly lower chance of ever taking up smoking, in samples of hundreds of thousands to millions of people.

Val66Met (rs6265) sits in the pro-BDNF region and alters intracellular trafficking and activity-dependent release of BDNF at the synapse: the molecule central to long-term potentiation and learning.

Cognitive effects are small, study-dependent and partly ancestry-specific; the Met allele is common and is not a deficit. Sleep, aerobic exercise and learning load move BDNF far more than genotype does.

What this isn’t: Not a measure of memory ability or intelligence, and not a diagnosis of any kind.

Yengo et al. 2018, Human Molecular Genetics: height and BMI in about 700,000 Europeans (PMID 30239722) · Saunders et al. 2022, Nature: tobacco and alcohol use in 3.4 million people (PMID 36477530) · GWAS Catalog: associations for rs6265 · MedlinePlus Genetics: BDNF gene · dbSNP: rs6265

CYP1A2 Caffeine metabolism: fast or slow (CYP1A2) Intermediate inducibility Mixed evidence

Whether the enzyme that clears most of your caffeine ramps up readily (so caffeine clears fast) or stays lower (so it lingers). A single regulatory variant nudges the dial, but how much coffee you actually drink, plus smoking and medicines, matter far more.

Your genotype: C/A (Substance response)
Most-associated reading: Intermediate inducibility

One high-inducibility allele: an in-between metaboliser profile.

How seriously to take this: the evidence

Strength: Mixed evidence
Source of the claim: PubMed: many small-to-medium caffeine and pharmacokinetic studies; widely used by consumer-DNA products, but replication of hard outcomes is patchy
Effect: The A allele is generally tied to a more inducible CYP1A2 enzyme, so AA individuals are often labelled 'fast' caffeine metabolisers and C-carriers 'slower', with caffeine lingering longer. In practice the effect shows up mostly as inducibility (e.g. in smokers or heavy coffee drinkers) and is modest and context-dependent. Reported links between this variant and outcomes like blood pressure or heart-attack risk after coffee are genuinely mixed and should be read cautiously.

CYP1A2 is a liver cytochrome-P450 enzyme that clears about 95% of caffeine and many drugs. rs762551 sits in intron 1 near regulatory sequence and is thought to influence how strongly the gene is induced rather than the enzyme's intrinsic shape, which is why environmental inducers like smoking and diet interact with the genotype.

A single regulatory SNP whose real-world impact is small next to how much coffee you drink, smoking, medicines and other genes: don't treat 'fast/slow' as a fixed personal verdict.

What this isn’t: Not a medical test, not a heart-disease predictor, and not a basis for changing caffeine intake on health grounds.

dbSNP: rs762551 · Functional SNP rs762551 in CYP1A2 and coffee intake (ScienceDirect)

CLOCK CLOCK: chronotype (morningness / eveningness) One evening-associated allele Contested

A core circadian-clock gene. The 3111T/C variant has been linked, inconsistently, to a tendency toward eveningness and later sleep timing, and genome-wide studies of chronotype in hundreds of thousands of people have not found it.

Your genotype: A/G (Sleep & chronotype)
Most-associated reading: One evening-associated allele

One copy of the evening-associated allele: a small, inconsistent nudge toward later timing at most, and one that large genome-wide studies did not detect.

How seriously to take this: the evidence

Strength: Contested
Source of the claim: PubMed: circadian-genetics studies (mixed replication); no genome-wide association in the GWAS Catalog, including the 351 chronotype loci of Jones et al., Nat Commun 2019
Effect: The evening-associated allele has been linked in some cohorts to greater eveningness and delayed sleep timing, but replication is patchy and effect sizes are small. The GWAS Catalog lists no genome-wide association for this variant with any trait, and it is not among the 351 chronotype loci found in 697,828 people (Jones et al. 2019).

CLOCK encodes a core transcription factor of the circadian oscillator. rs1801260 lies in the 3' UTR and may subtly affect transcript regulation and clock period.

Chronotype is highly polygenic (dozens of loci) and strongly shaped by light exposure, age and behaviour; this single variant captures only a sliver and is among the weaker reported associations.

What this isn’t: Not a diagnosis of a sleep disorder and not a fixed verdict on when you must sleep.

Jones et al. 2019, Nature Communications: chronotype in 697,828 individuals (PMID 30696823) · GWAS Catalog: rs1801260 (no associations listed) · GeneCards: CLOCK gene · dbSNP: rs1801260

COMT COMT: dopamine clearance and the 'warrior / worrier' story Met/Met (slower clearance) Mixed evidence

How quickly your prefrontal cortex clears dopamine. The Val158Met variant changes the enzyme's activity three- to fourfold, which is solid chemistry. The popular story that it sorts people into stress-proof 'warriors' and anxious 'worriers' has no support from large studies.

Your genotype: A/A (Cognition & stress)
Most-associated reading: Met/Met (slower clearance)

Two Met alleles: the slower form of the enzyme, so dopamine lingers a little longer in the prefrontal cortex. Some studies found slightly better working memory in calm conditions; pooled results are small, and large studies give the 'worrier' label no support.

How seriously to take this: the evidence

Strength: Mixed evidence
Source of the claim: Cognitive-genetics cohorts and a meta-analysis of executive function (Barnett et al., Mol Psychiatry 2007); genome-wide studies (GWAS Catalog) for what the variant measurably changes
Effect: The Met (A) allele lowers COMT enzyme activity, so prefrontal dopamine lingers longer; the Val (G) allele clears it faster. Genome-wide studies confirm the chemistry, with very strong associations for blood levels of COMT and of molecules it acts on, and show nothing comparable for personality, anxiety or intelligence. A pooled analysis of 12 studies of one executive-function test found a small advantage for Met/Met over Val/Val in healthy people (d = 0.29), and the effect was larger in the earliest studies.

COMT breaks down synaptic dopamine. Val158Met (rs4680) changes the enzyme's thermostability, so Met/Met carriers have roughly 3–4× lower activity and higher tonic prefrontal dopamine than Val/Val carriers.

The 'warrior/worrier' story is a popular simplification of a U-shaped, task- and stress-dependent effect: neither genotype is 'better', and a single SNP explains only a sliver of the variance.

What this isn’t: Not a measure of intelligence or resilience, and not a prediction of your response to any medication: discuss stimulants or psychiatric medicines only with a prescriber.

Barnett et al. 2007, Molecular Psychiatry: meta-analysis of COMT Val158Met and executive function (PMID 17325717) · GWAS Catalog: associations for rs4680 · MedlinePlus Genetics: COMT gene · dbSNP: rs4680

ANKK1 DRD2 / ANKK1: Taq1A, dopamine reward signalling A1 carrier Mixed evidence

The Taq1A variant beside the dopamine D2-receptor gene, long studied for differences in reward learning, motivation and addiction vulnerability.

Your genotype: G/A (Cognition & stress)
Most-associated reading: A1 carrier

One A1 allele: associated on average with somewhat lower D2-receptor availability; the behavioural differences reported are small.

How seriously to take this: the evidence

Strength: Mixed evidence
Source of the claim: PubMed: neuroimaging & behavioural-genetics cohorts
Effect: The A1 (A) allele is associated on average with lower striatal D2-receptor density and subtle differences in reward sensitivity and reinforcement learning. The link to addiction is real but weak and heavily modified by environment.

rs1800497 (Taq1A) is a missense variant in ANKK1, immediately adjacent to DRD2; A1 carriers show reduced striatal D2-receptor availability on PET imaging, shifting dopaminergic reward signalling.

Early 'reward deficiency' claims were overstated; effect sizes are small and addiction risk is overwhelmingly driven by environment, not this SNP.

What this isn’t: Not a predictor of addiction, willpower or any disorder: it is one small dopaminergic modifier among many.

MedlinePlus Genetics: DRD2 gene · dbSNP: rs1800497

MTHFR MTHFR: C677T, the most over-hyped variant in consumer genetics C/T (mildly reduced) Popularly over-hyped

You have almost certainly been told MTHFR matters. Here is the plain version: for the vast majority of people it does not, which is exactly why our clinical module does NOT report it.

Your genotype: G/A (Metabolism & appetite)
Most-associated reading: C/T (mildly reduced)

One 677T allele: about 65% of typical enzyme activity, with no health implication at normal folate intake. Extremely common.

How seriously to take this: the evidence

Strength: Popularly over-hyped
Source of the claim: ACMG practice guidance; MedlinePlus Genetics
Effect: The 677T (A) allele modestly lowers MTHFR enzyme activity. T/T can raise homocysteine slightly when folate intake is low: an effect that essentially disappears with normal folate status.

MTHFR helps convert folate into its active form for homocysteine metabolism. C677T (rs1801133) reduces the enzyme's thermostability, lowering activity to about 65% (C/T) or 30% (T/T) of normal, but adequate dietary folate compensates.

Major bodies, including the ACMG, advise AGAINST testing MTHFR for thrombophilia or recurrent pregnancy loss: the evidence does not support the elaborate claims sold around it. The depression and ADHD claims, often sold with methylfolate as the remedy, fare no better: the largest genome-wide studies of depression (807,553 people) and of ADHD (225,534 people) list no genome-wide significant signal near MTHFR. We surface it here only to set the record straight.

What this isn’t: NOT a cause of your symptoms, not a reason for special supplements without medical advice, and not a clinical finding: it is a normal, common variant.

MedlinePlus Genetics: MTHFR gene · ACMG practice guideline: lack of evidence for MTHFR testing (PubMed) · Howard et al. 2019, Nature Neuroscience: genome-wide meta-analysis of depression in 807,553 people (PMID 30718901) · Demontis et al. 2023, Nature Genetics: genome-wide analyses of ADHD, 27 loci (PMID 36702997)

PPARGC1A PPARGC1A: Gly482Ser, endurance & mitochondria One Ser allele Mixed evidence

PGC-1alpha is the master regulator of mitochondrial biogenesis: the adaptation behind aerobic training. The Gly482Ser variant is studied for differences in endurance trainability.

Your genotype: C/T (Fitness & performance)
Most-associated reading: One Ser allele

One Ser allele: small, inconsistent average differences in endurance measures at most.

How seriously to take this: the evidence

Strength: Mixed evidence
Source of the claim: PubMed: exercise-genomics cohorts
Effect: The Ser (T) allele has been associated in some studies with slightly lower endurance performance and trainability and with some metabolic traits, but results vary across cohorts.

PPARGC1A (PGC-1alpha) drives mitochondrial biogenesis and oxidative-fibre adaptation to aerobic exercise. Gly482Ser (rs8192678) may modestly affect the function of this transcriptional co-activator.

Athletic performance is massively polygenic and training-dominated; single-gene 'sports genetics' is mostly entertainment, and the effect sizes here are small and inconsistent.

What this isn’t: Not a ceiling on your fitness and not a basis for any training or talent decision.

GeneCards: PPARGC1A gene · dbSNP: rs8192678

Entries we couldn’t resolve from your file (43)

The marker wasn’t callable or the genotype isn’t in the curated table: see each note.

ACE Endurance vs power: the ACE 'sport gene' Contested

The original 'sport gene': the ACE insertion/deletion has been linked, loosely, to an endurance lean (I) or a power lean (D). We can only INFER it from a nearby tag SNP, and the performance story is contested, a curiosity, never a talent verdict.

Your genotype: Not read (Fitness & performance)

How seriously to take this: the evidence

Strength: Contested
Source of the claim: Tag-SNP proxy (rs4343, r²≈0.88 with the I/D in Europeans) for the classic ACE I/D performance literature, which is mixed-to-contested
Effect: Some studies link the ACE I (insertion) allele to a modest endurance lean and the D (deletion) allele to a power/strength lean, but results conflict and effect sizes are tiny. Because rs4343 only tags the actual I/D variant through linkage, any read-out here is a double inference: not a direct measurement. Treat it as a curiosity, not a verdict on athletic type.

ACE encodes angiotensin-converting enzyme, part of the renin-angiotensin system that regulates blood pressure and tissue perfusion. The D-associated allele is loosely tied to higher circulating ACE (hypothesised power/fast-twitch physiology), the I-associated allele to lower ACE (proposed endurance efficiency). The actual causal variant is a 287 bp Alu insertion/deletion that a single-base engine can't type: rs4343 only stands in for it.

Training, recovery, nutrition and sport-specific practice dominate athletic outcomes far more than this variant; single-gene 'sport gene' claims are widely oversold in consumer reports, and this is an inferred, contested one.

What this isn’t: Not a talent test and not a predictor of whether you can become an athlete.

This entry's marker is not on this array, so it was never read. We leave it unresolved rather than show the common (reference) version, which would look like a result. A sequenced file, or an array that carries this marker, can settle it.

dbSNP: rs4343 · rs4343 as a homogeneous-assay proxy for ACE I/D (PubMed 18057531)

ADH1B Alcohol metabolism speed (ADH1B) Established biology · inferred call

How fast your liver performs the very first step of breaking down alcohol. A high-activity variant speeds that step dramatically, producing an unpleasant acetaldehyde spike that, on average, nudges carriers toward drinking less.

Your genotype: Not read (Substance response)

How seriously to take this: the evidence

Strength: Established biology · inferred call
Source of the claim: Classic enzyme-kinetics data + multiple large multi-ancestry cohorts (PubMed): one of the most robust findings in alcohol genetics
Effect: The C (His48) allele encodes an enzyme that converts ethanol to acetaldehyde dramatically faster, transiently spiking acetaldehyde and producing an unpleasant reaction that discourages heavy drinking. Carriers show lower rates of alcohol dependence (reported odds ratios around 0.3-0.4 for the protective allele) and tend to drink less. The size of the effect depends on ancestry-specific allele frequency and on the ALDH2 background.

ADH1B is a liver alcohol dehydrogenase that performs the first step of ethanol metabolism. The His48 substitution raises the enzyme's turnover roughly 70-80 fold, so ethanol is oxidised to acetaldehyde faster; the resulting acetaldehyde bump is aversive: especially when ALDH2 clearance is also slow.

This shifts drinking behaviour on average: it is not a switch. Many other genetic and environmental factors shape alcohol use, so individual outcomes vary widely.

What this isn’t: Not a test for, or a diagnosis of, alcohol use disorder, and not a prediction of any individual's drinking on its own.

dbSNP: rs1229984 · ADH1B and alcohol dependence across European and African ancestry (PMC3252425)

BCO1 Beta-carotene converter status (BCO1) Replicated · small effect

Whether your genes make you a more or less efficient converter of plant beta-carotene (the orange pigment in carrots and sweet potato) into active vitamin A. A common variant shifts the balance; total diet and fat intake matter far more for your vitamin A status.

Your genotype: Not read (Nutrition & metabolism)

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: GWAS (NHGRI-EBI GWAS Catalog), Ferrucci et al. 2009, plus replication of BCO1 carotenoid associations
Effect: The G allele tracks lower BCO1 enzyme efficiency, so beta-carotene is converted to vitamin A more slowly and circulating carotenoid levels run higher. The reference T allele tracks more efficient conversion and lower circulating carotenoids. The site explains only a couple of percent of the variation.

BCO1 is the enzyme that splits dietary beta-carotene into retinal on the way to vitamin A; the variant subtly lowers the enzyme's activity, leaving more unconverted carotenoid in the blood.

A small, replicated effect. Vitamin A status is dominated by what you eat, how much fat is in the meal and overall health, not this single marker. Higher circulating carotenoid is not the same as deficiency.

What this isn’t: Not a diagnosis, not a supplement recommendation, and not a vitamin A deficiency test.

NHGRI-EBI GWAS Catalog, rs6564851 · dbSNP, rs6564851

CADM2 Risk-taking & processing speed (CADM2) Replicated · small effect

A gene that turns up again and again in genome-wide studies of risk-taking, sociability and how fast people process information. The effects are real across huge samples but vanishingly small per person: population statistics, not a personality readout.

Your genotype: Not read (Social & personality)

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: Large GWAS meta-analyses (CHARGE processing-speed, UK Biobank risk-taking) + broader CADM2 behavioural-genetics literature: genome-wide significant, tiny per-variant effect
Effect: The T allele has been tied to slightly faster information-processing speed and, intriguingly, to slightly lower self-reported risk-taking at the same spot. These are statistical tendencies across very large samples, not individual-level predictions: each copy shifts the trait by a tiny fraction of a standard deviation.

CADM2 (cell adhesion molecule 2) helps neurons form and maintain synaptic connections, especially in reward, impulse-control and processing-speed circuits. The variant is intronic and likely nudges CADM2 expression rather than changing the protein; the exact regulatory mechanism is not pinned down.

Each copy shifts the trait by a tiny fraction of a standard deviation; CADM2 associations are polygenic-context effects, and behaviour is overwhelmingly shaped by environment and many other genes.

What this isn’t: Not a 'risk-taking gene', an intelligence test or a personality diagnosis: just one of thousands of common variants nudging population averages.

dbSNP: rs17518584 · GWAS for processing speed implicates CADM2 (Mol Psychiatry 2016)

GDF5 Joint cartilage and osteoarthritis: a GDF5 promoter variant Replicated · small effect

GDF5 is a growth factor that helps build joints and cartilage. A common variant in the switch that controls it turns the gene down slightly and is one of the best-replicated genetic links to osteoarthritis of the knee and hip, though each copy shifts the odds only a little.

Your genotype: Not read (Fitness & performance)

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: Association plus promoter assays, Miyamoto et al. 2007 (Nature Genetics), two Japanese hip-osteoarthritis cohorts with replication in Japanese and Han Chinese knee osteoarthritis; genome-wide significant for knee osteoarthritis in Icelandic and UK data, Styrkarsdottir et al. 2018 (Nature Genetics)
Effect: The A allele (the '+104T' susceptibility allele on the gene's own strand) was associated with higher odds of hip and knee osteoarthritis in East Asian cohorts and, at genome-wide significance, of knee osteoarthritis in a large European study. The per-copy effect is modest, and the risk allele is the more common one in Europeans.

The variant sits in the GDF5 core promoter, the stretch of DNA that controls how much of the gene is made. In cartilage-forming cells the susceptibility allele drove less gene activity, and lower GDF5 during joint development and repair is thought to leave cartilage slightly less resilient.

Osteoarthritis depends far more on age, body weight, past injuries and joint loading than on this variant, and many carriers of the higher-risk allele never develop symptomatic arthritis. Because that allele is the majority one in Europeans, carrying it is ordinary. It is also very rare in African ancestry, where these findings say little.

What this isn’t: Not an osteoarthritis diagnosis, not a measure of your joints' current state, and not a prediction about any sport or activity.

Miyamoto et al. 2007, Nature Genetics: A functional polymorphism in the 5' UTR of GDF5 is associated with susceptibility to osteoarthritis (PMID 17384641) · Styrkarsdottir et al. 2018, Nature Genetics: Meta-analysis of Icelandic and UK data sets identifies new osteoarthritis loci (PMID 30374069) · NHGRI-EBI GWAS Catalog: rs143383

CCR5 CCR5Δ32: the HIV-resistance deletion (read via a proxy SNP) Established biology · inferred call

The famous 32-base deletion that, in two copies, makes most HIV-1 unable to enter your cells: the genotype behind the 'Berlin' and 'London' patients. We can't type the deletion itself, so we read a nearby tag SNP that travels with it in people of European descent. That makes this an inference, not a direct readout.

Your genotype: Not read (Immunity & infection)

How seriously to take this: the evidence

Strength: Established biology · inferred call
Source of the claim: PubMed / GWAS: CCR5Δ32 (rs333) as HIV-1 co-receptor knockout; rs113341849 used as its r²≈0.97 tag SNP where the deletion can't be genotyped directly.
Effect: CCR5 is the main co-receptor R5-tropic HIV-1 uses to enter cells. Δ32 truncates the protein so it never reaches the surface. Δ32/Δ32 people are strongly resistant to R5-tropic HIV-1; Δ32/+ people who do get infected tend to progress more slowly. The biology here is solid: what's soft is that we infer your Δ32 status from a proxy SNP rather than reading the deletion.

The CCR5 protein sits on immune-cell surfaces and is the doorway most HIV-1 strains use. The Δ32 allele deletes 32 bases, frame-shifting the gene so no working receptor is made. With two copies there is essentially no doorway, so R5-tropic HIV-1 can't get in. rs113341849 sits nearby and is inherited together with Δ32 in European-descent populations, so its A allele stands in for the deletion we can't directly see.

This is a PROXY, not the deletion: the tag travels with Δ32 about 97% of the time in Europeans, and that link is weak-to-meaningless in non-European ancestry. There, a 'no tag' result tells you little. Resistance is also not immunity: X4-tropic HIV-1 strains can still infect Δ32/Δ32 people, and one Δ32 copy is NOT protection from getting HIV. Nothing here is medical advice. Do not change how you protect yourself from HIV based on this card.

What this isn’t: Not a direct test for the deletion, not immunity, not a clinical result, and not a reason to alter prevention. Δ32/Δ32 also carries a downside: links to worse symptomatic West Nile virus and possibly more severe influenza, so 'resistance allele' doesn't mean 'better immune system'. (This is the gene the 2018 CRISPR-baby experiment targeted; editing it in healthy embryos was condemned worldwide.)

MedlinePlus Genetics: CCR5 gene · SNPedia: rs333 (CCR5-Δ32) · dbSNP: rs113341849 (tag SNP)

COL5A1 Tendon stiffness & flexibility (COL5A1) Mixed evidence

A collagen variant linked, modestly, to how stiff or supple your tendons and ligaments tend to be, and a small tilt in soft-tissue injury risk and range of motion. The signal is real but population-dependent and easily swamped by training.

Your genotype: Not read (Fitness & performance)

How seriously to take this: the evidence

Strength: Mixed evidence
Source of the claim: Collagen soft-tissue / tendon-injury candidate-gene literature, meta-analysed: modest signal in Caucasian cohorts, doesn't replicate in Asian ones
Effect: Some studies associate the TT genotype with a modestly higher risk of musculoskeletal soft-tissue injuries (tendon/ligament) and reduced joint flexibility, with the C allele appearing protective and linked to greater range of motion. The effect is small, population-dependent, and does not replicate everywhere: a wellness-curiosity signal, not a clinical injury predictor.

COL5A1 encodes the alpha-1 chain of type V collagen, which regulates the assembly and diameter of type I collagen fibrils in tendons and ligaments. rs12722 sits in the 3' untranslated region and is thought to influence mRNA stability/expression; the T variant is proposed to shift the collagen ratio toward thinner, less tensile-strong fibrils.

Load management, warm-up, biomechanics and training history drive injury risk far more than this single variant, and the association is inconsistent across populations.

What this isn’t: Not a talent test and not a diagnosis of injury-proneness.

dbSNP: rs12722 · COL5A1 rs12722 and soft-tissue injuries: meta-analysis (PMC5880610)

CRY1 Night-owl family variant: CRY1 Mixed evidence

A change in a core clock gene that made people in several families sleep and wake about an hour later than their relatives. In a large population study the same variant shifted sleep by only a few minutes. It is uncommon in most Europeans and very rare in Finland.

Your genotype: Not read (Sleep & chronotype)

How seriously to take this: the evidence

Strength: Mixed evidence
Source of the claim: Family study (Patke et al., Cell 2017) and a population test in 1,444 UK Biobank carriers (Weedon et al., PLoS Genet 2022)
Effect: In the families where it was found, carriers had delayed sleep phase: sleep and wake times shifted roughly an hour later, sometimes with broken sleep. In UK Biobank, 10 % of carriers called themselves 'definitely an evening person' against 8 % of everyone else, and activity monitors put their sleep about 5 minutes later on average, a difference at the edge of what the study could detect. Carriers were no more likely than anyone else to have a recorded sleep disorder. Outside the families that brought it to light, its effect is small.

CRY1 is one of the brakes of the circadian clock. The variant changes a splice site, so the protein is made without the stretch coded by exon 11. In cell experiments the shortened protein binds the clock's activators CLOCK and BMAL1 more tightly, which lowers the output of the genes they switch on and lengthens the clock's cycle.

Sleep timing is shaped by age, light, work and school hours, and by hundreds of common variants with tiny effects each. One family-discovered variant does not decide when you can sleep.

What this isn’t: Not a diagnosis of a sleep disorder, and not a test for ADHD: one research group's claim of a link, drawn from families selected for having both ADHD and insomnia, has not been confirmed in the general population. A carrier result does not mean your sleep pattern is fixed, and a non-carrier result does not rule out delayed sleep phase, which usually has no single genetic cause.

Patke et al. 2017, Cell: CRY1 mutation in familial delayed sleep phase disorder (PMID 28388406) · Weedon et al. 2022, PLoS Genetics: Mendelian sleep and circadian variants in a population setting · Onat et al. 2020, J Clin Invest: the claimed ADHD link this card does not support (PMID 32538895) · dbSNP: rs184039278

DRD4 Novelty-seeking, the contested version (DRD4) Contested

A dopamine-receptor promoter variant often tied to novelty-seeking or 'the wanderlust gene': included as a contested entry. The personality fame really rests on a repeat in the gene that we cannot type, and the SNP associations are small and inconsistent. It tells you very little.

Your genotype: Not read (Social & personality)

How seriously to take this: the evidence

Strength: Contested
Source of the claim: DRD4 promoter SNP; meta-analyses find at most a small, inconsistently replicated novelty-seeking/ADHD association, and the famous signal is the exon-3 48bp VNTR (7R), not this SNP
Effect: This SNP tells you very little. The novelty-seeking, extraversion and ADHD associations attached to rs1800955 are small, disputed and frequently fail to replicate; functional studies have even questioned whether it meaningfully changes DRD4 transcription. Any single-SNP 'personality readout' framing is oversold.

rs1800955 sits in the DRD4 promoter (~521 bp upstream of the start), where it could in principle nudge gene expression, though direct transcriptional effects have been hard to confirm. The famous DRD4 'novelty-seeking' signal comes from the exon-3 48bp VNTR (the 7R repeat allele): a separate length polymorphism, NOT this SNP, and not inferable from it.

Personality is not genetic destiny: traits like curiosity or impulsivity arise from environment, life experience and the combined tiny effects of very many genes, not one promoter SNP. Don't use this to label yourself or anyone, diagnose ADHD, or predict temperament.

What this isn’t: Explicitly NOT a novelty-seeking, 'adventure gene', impulsivity or ADHD diagnostic test, and not a personality assessment.

dbSNP: rs1800955 · No direct effect of the -521 C/T polymorphism on DRD4 transcription (PMC1481588)

FAAH Anandamide tone (FAAH) Mixed evidence

A variant in the enzyme that breaks down anandamide: one of the body's own cannabis-like signalling molecules. Carriers of the low-activity allele have higher anandamide tone, loosely tied to differences in anxiety, reward and pain. The molecular effect is solid; the behavioural read-out is noisy.

Your genotype: Not read (Cognition & stress)

How seriously to take this: the evidence

Strength: Mixed evidence
Source of the claim: Functional biochemistry (enzyme stability) is solid; the behavioural associations come from many small-to-moderate human studies with conflicting directions
Effect: The minor (A) allele lowers the enzyme's activity, so A-carriers tend to have higher anandamide levels, and studies have loosely linked this to differences in anxiety, fear-extinction, reward and pain. The psychological signal is real-but-noisy: present in some cohorts, absent or reversed in others.

FAAH degrades anandamide and related fatty-acid amides that act on cannabinoid receptors. The Pro129Thr change makes the enzyme more prone to degradation, shortening its half-life and raising steady-state anandamide. Higher anandamide tone is the proposed route to altered threat processing and reward signalling.

Effect sizes for anxiety/well-being/pain are small and frequently fail to replicate. This is NOT the dramatic 'no fear, no pain' phenotype of the rare FAAH-OUT case (Jo Cameron), which involved additional mutations.

What this isn’t: Not a pain-immunity, fearlessness or 'happiness' gene, and not a clinical or psychiatric diagnosis.

dbSNP: rs324420 · FAAH rs324420: biological pathways review (PMC10606937)

FADS1 Omega-3/6 conversion efficiency (FADS1) Replicated · small effect

How readily your body upgrades plant-form omega-3 and omega-6 fats into the long-chain forms (EPA, arachidonic acid) the body actually uses. One of the most reproducible nutrition-genetics signals, though what you eat matters far more than your genotype.

Your genotype: Not read (Nutrition & metabolism)

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: GWAS Catalog + multiple PUFA GWAS/meta-analyses (InCHIANTI and others): one of the most reproducible nutrigenetic loci, modest per-person effect
Effect: People carrying the G allele tend to convert dietary plant omega-3 (ALA) and omega-6 (LA) into the long-chain forms (EPA, arachidonic acid) more readily, so they show higher blood levels from the same diet. T-allele carriers convert less efficiently, with TT individuals showing roughly 30-40% lower arachidonic acid/EPA in some studies. This is a tendency in conversion efficiency, not a guarantee of any particular blood level.

rs174537 tags variation that affects FADS1 expression (the T allele is linked to increased promoter methylation and lower FADS1 protein). FADS1 encodes the Δ5-desaturase, a rate-limiting step turning precursor PUFAs into long-chain EPA and arachidonic acid. Lower activity means more reliance on preformed EPA/DHA from diet, such as oily fish.

Actual omega-3/6 status is driven mostly by what you eat: oily fish and direct EPA/DHA largely bypass this conversion step. The genotype shifts a tendency, it does not set your fatty-acid levels.

What this isn’t: Not a blood fatty-acid measurement and not a diagnosis of omega-3 deficiency.

dbSNP: rs174537 · GWAS of plasma PUFAs (InCHIANTI), PLoS Genetics

FKBP5 Stress-axis regulation, contested (FKBP5) Contested

A variant in a gene that tunes the body's stress-hormone feedback loop, studied for an interaction between childhood adversity and later PTSD or depression. The mechanism is plausible but the gene-by-environment evidence is contested: this is a research hypothesis, never a test.

Your genotype: Not read (Cognition & stress)

How seriously to take this: the evidence

Strength: Contested
Source of the claim: Candidate-gene / gene-by-environment stress-biology literature (Klengel/Binder group and follow-ups); GxE psychiatry has a poor replication record overall
Effect: Some studies report that, among people who experienced significant childhood adversity, T-allele carriers show modestly altered stress-hormone (cortisol/glucocorticoid) regulation and a somewhat higher reported rate of PTSD or depressive symptoms, but only in interaction with environment, never as a standalone effect. The associations are small, the literature is mixed, and several independent replications have been weak or null.

FKBP5 encodes a co-chaperone that regulates glucocorticoid-receptor sensitivity, part of the negative-feedback loop on the stress (HPA-axis) response. rs1360780 is an intronic variant proposed to act allele-specifically: in T-carriers, early-life trauma is reported to drive demethylation of an FKBP5 regulatory region. This is a proposed epigenetic-interaction mechanism, not a proven deterministic switch.

NOT a PTSD or depression test, and it cannot tell you whether you will experience either. Mental health is shaped by thousands of variants plus environment, relationships and chance; the gene-environment claim here is weak, debated and unreliable for any individual prediction.

What this isn’t: Not a psychiatric diagnosis and not a prediction that you will develop any mental-health condition.

Klengel et al. 2013, Nat Neurosci: FKBP5 demethylation mediates gene–childhood-trauma interaction · Border et al. 2019, Am J Psychiatry: no support for historical candidate-gene/GxE hypotheses for depression

FTO FTO: appetite & body-weight set-point Replicated · small effect

The most-replicated common obesity-associated gene. The effect is genuine but small, and works mostly through appetite and satiety rather than metabolism.

Your genotype: Not read (Metabolism & appetite)

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: GWAS Catalog: large obesity meta-analyses
Effect: Each risk (A) allele is associated on average with roughly 1–1.5 kg higher body weight and somewhat reduced satiety. It is robustly replicated but explains only a tiny fraction of body-weight variance.

FTO risk variants act largely in the brain on appetite-regulating circuits (and on nearby genes such as IRX3/IRX5), nudging satiety and energy intake rather than basal metabolic rate.

The per-allele effect is small and fully modifiable: physical activity measurably blunts the FTO association, and diet and environment dominate the outcome.

What this isn’t: Not a diagnosis of obesity and not destiny: it shifts a probability slightly, nothing more.

MedlinePlus Genetics: FTO gene · GWAS Catalog: rs9939609

IGF2BP2 IGF2BP2: a lead type 2 diabetes marker Replicated · small effect

One of the first and most reproducible common variants linked to type 2 diabetes risk. It nudges the odds by a little; type 2 diabetes is highly polygenic and dominated by weight, diet and activity.

Your genotype: Not read (Metabolism & appetite)

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: GWAS (NHGRI-EBI GWAS Catalog): among the earliest and most-replicated type 2 diabetes loci (2007 onward)
Effect: Each copy of the T allele at rs4402960 is associated with modestly higher odds of type 2 diabetes: a per-allele odds ratio of roughly 1.1 across large studies. It is one of dozens of common variants that each shift the odds only slightly.

rs4402960 lies in an intron of IGF2BP2, a gene in the insulin-like growth factor pathway that is active in pancreatic beta-cell development. The variant is thought to subtly affect insulin secretion rather than insulin resistance.

A real but small per-allele effect. Type 2 diabetes is highly polygenic and strongly shaped by weight, diet, activity and age: this single marker barely moves any one person's risk, and it is already one of the markers inside our polygenic type-2-diabetes score.

What this isn’t: Not a diagnosis of diabetes or pre-diabetes, and not a prediction that you will or won't develop it.

NHGRI-EBI GWAS Catalog: rs4402960 (type 2 diabetes) · dbSNP: rs4402960

IL6R The IL-6 receptor trade-off: heart versus allergy (IL6R) Replicated · small effect

A common change in the receptor for interleukin-6, a major inflammation signal, mimics a mild version of an arthritis drug that blocks the same receptor. It is linked to lower inflammation markers and slightly lower heart-disease odds, but also to a higher chance of allergic conditions such as asthma and eczema.

Your genotype: Not read (Immunity & infection)

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: Mendelian-randomisation and meta-analysis studies of up to 133,449 people, IL6R MR Consortium and IL6R Genetics Consortium 2012 (The Lancet); asthma GWAS, Ferreira et al. 2011 (The Lancet); large protein, allergy and arthritis GWAS in the NHGRI-EBI GWAS Catalog
Effect: Each copy of the C allele (alanine at position 358) was associated with higher circulating IL-6 and soluble IL-6 receptor, lower C-reactive protein, and about 5 percent lower odds of coronary heart disease. The same allele is linked to lower odds of rheumatoid arthritis and higher odds of asthma, hay fever and eczema (asthma odds ratio about 1.09 per copy).

The alanine version of the receptor is thought to be cut off the cell surface more readily, which fits the much higher levels of free-floating receptor measured in carriers. Less signalling through the cell-bound receptor dampens inflammation in a way that resembles a weak dose of tocilizumab, a drug that blocks this receptor. Why the same change raises allergy odds is not settled.

Small effects in both directions, which is the point of this card: the same allele that nudges heart and arthritis odds down nudges allergy odds up. None of these effects is large enough to matter for any one person on its own, and heart disease, arthritis and allergy are each shaped by many other factors.

What this isn’t: Not a diagnosis of asthma, allergy, arthritis or heart disease, not a measure of your inflammation, and not a statement about whether any medicine suits you.

IL6R MR Consortium 2012, The Lancet: The interleukin-6 receptor as a target for prevention of coronary heart disease (PMID 22421340) · IL6R Genetics Consortium Emerging Risk Factors Collaboration 2012, The Lancet: IL-6 receptor pathways in coronary heart disease (PMID 22421339) · Ferreira et al. 2011, The Lancet: Identification of IL6R and chromosome 11q13.5 as risk loci for asthma (PMID 21907864) · NHGRI-EBI GWAS Catalog: rs2228145

TMPRSS6 Iron levels and the hepcidin brake (TMPRSS6) Replicated · small effect

TMPRSS6 helps set how much iron the body absorbs by controlling hepcidin, the hormone that acts as a brake on iron uptake. A common variant in it is one of the strongest genetic influences on everyday iron and haemoglobin levels.

Your genotype: Not read (Nutrition & metabolism)

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: Two independent genome-wide association studies published together: Benyamin et al. 2009 (iron status, Nature Genetics) and Chambers et al. 2009 (haemoglobin in 16,001 people of European and Indian Asian ancestry, Nature Genetics); confirmed in very large blood-cell GWAS (NHGRI-EBI GWAS Catalog)
Effect: The A allele (valine at position 736) was associated with lower serum iron, lower transferrin saturation and smaller red blood cells, and with haemoglobin about 0.13 g/dL lower per copy. The G allele goes with the opposite, slightly higher readings. These are small shifts within the normal range.

TMPRSS6 is an enzyme on liver cells that keeps hepcidin production in check. Hepcidin closes the gate that lets iron out of the gut and out of storage. The valine version appears to restrain hepcidin less well, so hepcidin runs a little higher and slightly less iron reaches the blood.

A real, replicated effect, but a small one next to diet, menstrual and other blood losses, pregnancy, blood donation and inflammation, which drive iron levels far more. Iron status is measured with blood tests that a clinician interprets; this card does not stand in for them.

What this isn’t: Not a diagnosis of iron deficiency, anaemia or iron overload, says nothing about haemochromatosis (that is the separate HFE section), and is not a reason to change iron intake.

Benyamin et al. 2009, Nature Genetics: Common variants in TMPRSS6 are associated with iron status and erythrocyte volume (PMID 19820699) · Chambers et al. 2009, Nature Genetics: Genome-wide association study identifies variants in TMPRSS6 associated with hemoglobin levels (PMID 19820698) · NHGRI-EBI GWAS Catalog: rs855791

PCSK9 Lifelong lower LDL: the R46L variant (PCSK9) Replicated · small effect

A variant that partly switches off PCSK9, the protein a class of cholesterol drugs is designed to block. Carriers tend to have lower LDL cholesterol from birth. It is uncommon in most of Europe and about two to three times as frequent in Finland.

Your genotype: Not read (Metabolism & appetite)

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: Prospective cohort study, Cohen et al. 2006 (New England Journal of Medicine), 9,524 white participants followed for 15 years; confirmed as one of the strongest LDL signals in large lipid GWAS including Finnish cohorts (NHGRI-EBI GWAS Catalog)
Effect: In the 2006 study, carriers of the T allele (leucine at position 46) had about 15 percent lower LDL cholesterol and about half the rate of coronary heart disease over 15 years compared with non-carriers. A modest LDL reduction carried from birth seems to add up over decades, which is the main lesson the finding taught.

PCSK9 tags the liver's LDL receptors for destruction. Fewer receptors means less LDL cholesterol is cleared from the blood. The R46L change reduces PCSK9's activity, so more receptors survive and blood LDL runs lower. Drugs that block PCSK9 work on the same principle.

Lower average LDL is not protection on its own: diet, smoking, blood pressure and many other variants still drive heart risk, and carriers do develop heart disease. The heart-risk figure comes from one cohort and is imprecise: the study's range runs from a 21 to a 68 percent reduction. Only a lipid test shows your actual cholesterol, and a clinician interprets it.

What this isn’t: Not a cholesterol reading, not a heart-risk score, and not a reason to treat any lipid result differently.

Cohen et al. 2006, NEJM: Sequence variations in PCSK9, low LDL, and protection against coronary heart disease (PMID 16554528) · NHGRI-EBI GWAS Catalog: rs11591147 · dbSNP: rs11591147

PNPLA3 Liver fat: the I148M variant (PNPLA3) Established biology · inferred call

The single strongest common genetic influence on how much fat the liver stores. The I148M version of PNPLA3 is carried by about two in five people of European ancestry and is linked to more liver fat and higher liver-enzyme readings, with weight and alcohol shaping how much it matters.

Your genotype: Not read (Metabolism & appetite)

How seriously to take this: the evidence

Strength: Established biology · inferred call
Source of the claim: Genome-wide association, Romeo et al. 2008 (Nature Genetics), since replicated in dozens of studies across ancestries; among the strongest signals in the NHGRI-EBI GWAS Catalog for liver enzymes, fatty liver and cirrhosis
Effect: The G allele (methionine at position 148) was associated with more liver fat and with liver inflammation in the discovery study; liver fat was more than twice as high in G/G people as in non-carriers. Large later studies link it to higher liver-enzyme levels and higher average odds of fatty-liver disease and cirrhosis. The effect is amplified by excess body fat (Stender et al. 2017), and the same allele is linked to alcohol-related cirrhosis.

PNPLA3 is an enzyme on the surface of fat droplets inside liver cells. The methionine version escapes the cell's normal clean-up tagging and piles up on those droplets; experiments in mice show that this build-up, more than any loss of enzyme activity, is what makes fat accumulate in the liver (BasuRay et al. 2017, 2019).

A strong, well-replicated effect on liver fat, but most carriers never develop serious liver disease: weight, alcohol, diabetes and other variants shape the outcome far more than this one marker. Liver health is assessed with blood tests and imaging that a clinician interprets, and this card does not replace them.

What this isn’t: Not a diagnosis of fatty liver or any liver disease, not a measure of your current liver fat, and not a prediction that you will develop cirrhosis.

Romeo et al. 2008, Nature Genetics: Genetic variation in PNPLA3 confers susceptibility to nonalcoholic fatty liver disease (PMID 18820647) · Stender et al. 2017, Nature Genetics: Adiposity amplifies the genetic risk of fatty liver disease conferred by multiple loci (PMID 28436986) · BasuRay et al. 2019, PNAS: Accumulation of PNPLA3 on lipid droplets is the basis of associated hepatic steatosis (PMID 31019090) · NHGRI-EBI GWAS Catalog: rs738409 · dbSNP: rs738409

HSD17B13 HSD17B13: a protective liver variant (read via a proxy SNP) Established biology · inferred call

A common variant that breaks the liver enzyme HSD17B13 is linked to lower liver-enzyme levels and lower average odds of chronic liver disease. We cannot read the variant itself, a one-letter insertion, so we read a nearby marker that travels with it in people of European ancestry. That makes this an inference, not a direct readout.

Your genotype: Not read (Metabolism & appetite)

How seriously to take this: the evidence

Strength: Established biology · inferred call
Source of the claim: Exome sequencing plus health records, Abul-Husn et al. 2018 (New England Journal of Medicine), 46,544 people with replication in independent cohorts; rs6834314 is the variant's GWAS tag (r2 0.94 in 1000 Genomes Europeans, 0.89 in Finns)
Effect: The protective insertion was associated with lower ALT and AST liver-enzyme levels and with lower average odds of alcohol-related liver disease, non-alcoholic liver disease and cirrhosis: for example about 42 percent lower odds of alcoholic liver disease in one-copy carriers and 53 percent lower in two-copy carriers. The biology is solid; what is soft here is that we infer the insertion from the G allele of a linked marker.

HSD17B13 is an enzyme on fat droplets in liver cells. The insertion disrupts how its message is spliced, producing an unstable, truncated protein with reduced activity. Losing the enzyme does not stop fat building up, but it was associated with less progression from simple fatty liver to inflamed liver (steatohepatitis), and it softened the liver injury linked to the PNPLA3 I148M variant.

This is a PROXY: the marker matches the insertion in roughly nine out of ten European-ancestry chromosomes, and the link is weak in African ancestry (r2 about 0.17), where a result here says little. 'Lower odds' is not immunity: carriers still develop liver disease, especially with heavy drinking or excess weight. Liver health is judged from tests a clinician interprets.

What this isn’t: Not a direct test for the HSD17B13 variant, not a liver diagnosis, and not a licence to discount alcohol or weight as risks.

Abul-Husn et al. 2018, NEJM: A protein-truncating HSD17B13 variant and protection from chronic liver disease (PMID 29562163) · NHGRI-EBI GWAS Catalog: rs6834314 (tag SNP) · dbSNP: rs72613567 (the insertion this card infers)

TM6SF2 Liver fat versus blood fat: the E167K variant (TM6SF2) Replicated · small effect

A less common variant, carried by about one in eight people of European ancestry, that keeps more fat inside the liver and sends less of it into the blood. Carriers tend to have more liver fat but lower blood cholesterol and triglycerides.

Your genotype: Not read (Metabolism & appetite)

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: Exome-wide association, Kozlitina et al. 2014 (Nature Genetics), and an independent Norwegian exome study of lipids, Holmen et al. 2014 (Nature Genetics); confirmed in very large lipid and liver GWAS (NHGRI-EBI GWAS Catalog)
Effect: The T allele (lysine at position 167) was associated with higher liver fat and higher levels of the liver enzyme ALT, and at the same time with lower LDL cholesterol, total cholesterol and triglycerides in the blood. Most carriers have one copy; two copies are uncommon.

TM6SF2 helps the liver package fat into particles (VLDL) for export into the bloodstream. The lysine version produces about half as much working protein, so less fat leaves the liver: more is stored in liver cells and less circulates in the blood. Knocking the gene down in mice reproduced both effects.

A real but moderate effect, and a two-sided one: what it means for heart health over a lifetime is still being worked out, and this card makes no heart-risk claim. Weight, alcohol and diet shape liver fat far more than this variant, and liver health is judged from tests a clinician interprets.

What this isn’t: Not a diagnosis of fatty liver or of any lipid disorder, and not a reading of your actual cholesterol.

Kozlitina et al. 2014, Nature Genetics: Exome-wide association study identifies a TM6SF2 variant that confers susceptibility to NAFLD (PMID 24531328) · Holmen et al. 2014, Nature Genetics: Coding variation identifies a candidate causal variant in TM6SF2 influencing total cholesterol (PMID 24633158) · NHGRI-EBI GWAS Catalog: rs58542926

CETP HDL & longevity curio (CETP I405V) Mixed evidence

A cholesterol-transfer variant linked, in some long-lived families, to higher 'good' HDL cholesterol, larger lipoprotein particles and slower memory decline. The longevity story is real but inconsistent across populations: interesting, not a verdict on how long you'll live.

Your genotype: Not read (Aging & longevity)

How seriously to take this: the evidence

Strength: Mixed evidence
Source of the claim: Directly genotyped coding SNP; Ashkenazi centenarian/offspring cohorts (Barzilai lineage) plus other studies: real but population-heterogeneous
Effect: In several cohorts the Val/Val (G/G) genotype is associated with higher HDL cholesterol, larger HDL/LDL particles, and, in Ashkenazi Jewish centenarian families, exceptional longevity and slower memory decline versus Ile/Ile. These are population-specific, modest-effect associations that have not held up uniformly elsewhere (the direction even flips in some East-Asian samples), so any single result is easily over-sold.

CETP shuttles cholesteryl esters from HDL to other lipoproteins; the Val405 form is linked to lower CETP activity, which lets HDL accumulate as larger, cholesterol-rich particles. Larger lipoprotein particle size is the proposed link to favourable cardiovascular and cognitive ageing, but the amino-acid change is predicted benign, so any effect is subtle and indirect.

Longevity is overwhelmingly driven by lifestyle, environment and chance; this one SNP explains only a sliver of the variance, and its direction even flips between ancestries.

What this isn’t: Not a lifespan prediction and not a verdict on how long someone will live.

dbSNP: rs5882 · CETP I405V and memory decline / dementia (Sanders et al., PMC3047443)

FOXO3 Longevity-associated FOXO3 variant Replicated · small effect

FOXO3 is one of only a couple of genes repeatedly linked to reaching very old age across many populations. A common variant is associated with slightly better odds of exceptional longevity, but lifestyle and luck dominate by far.

Your genotype: Not read (Aging & longevity)

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: GWAS and candidate-gene studies, Willcox et al. 2008, PNAS (Honolulu cohort); replicated in German, Italian, Chinese and other long-lived cohorts
Effect: The G allele at this site, which is the GRCh38 reference base, is the longevity-associated allele. People with two copies (the G/G genotype) showed modestly higher odds of reaching very old age than carriers of the T allele. Per-allele effects on lifespan are small and emerge only at the population level.

FOXO3 is a transcription factor in insulin and stress-response pathways linked to cellular maintenance; the variant is thought to subtly increase protective FOXO3 activity, though the exact functional change is still studied.

One of the better-replicated longevity signals, yet exceptional age is overwhelmingly shaped by environment, behaviour, chance and many other genes. This single marker barely moves any one person's odds.

What this isn’t: Not a prediction of your lifespan, not a health diagnosis, and not a reason to change anything you do.

Willcox et al. 2008, PNAS, FOXO3 and human longevity (PMID 18765803) · dbSNP, rs2802292

KL Klotho KL-VS: heterozygote-advantage curio Contested

An unusual longevity/cognition variant where carrying exactly ONE copy (not two) is the form linked to better outcomes in some studies. The signal is genuinely contested: it fails to replicate in several large cohorts, so read it as a research curiosity, not a benefit.

Your genotype: Not read (Aging & longevity)

How seriously to take this: the evidence

Strength: Contested
Source of the claim: Directly genotyped coding SNP used as a KL-VS haplotype tag; positive cohorts + meta-analyses, but large non-replications (Newcastle 85+, UK Biobank)
Effect: KL-VS heterozygotes (one copy) have, in several studies, shown better cognition, higher circulating klotho, and a longevity edge at older ages, while non-carriers and, importantly, homozygous-variant individuals do worse. It is a balanced, heterozygote-advantage pattern, NOT a dose response, and it does not replicate in all cohorts, so it is easy to over-sell.

The Val352 substitution alters klotho protein trafficking/secretion: heterozygotes show increased secreted klotho whereas homozygous KL-VS carriers paradoxically show reduced klotho: a plausible basis for one copy being optimal. Klotho influences FGF23/phosphate, insulin/IGF-1 and oxidative-stress pathways, the proposed routes to its effects.

Lifespan and cognitive ageing are dominated by environment, lifestyle and chance; this single haplotype explains very little and its longevity association fails to replicate in major cohorts. rs9536314 alone only infers KL-VS status.

What this isn’t: Not a lifespan prediction and not a longevity verdict.

dbSNP: rs9536314 · No KLOTHO–longevity association in Newcastle 85+ and UK Biobank (PMC8893196)

MAOA The 'warrior gene' myth (MAOA) Popularly over-hyped

The notorious 'warrior gene': included to set the record straight, not to label anyone. The headline story rests on a repeat in the gene's promoter that we cannot read, and the aggression narrative is badly oversold and has a real history of misuse. This SNP says essentially nothing about you.

Your genotype: Not read (Social & personality)

How seriously to take this: the evidence

Strength: Popularly over-hyped
Source of the claim: A synonymous MAOA tag SNP repeatedly press-styled as a 'warrior gene' readout; the fame actually attaches to the MAOA promoter uVNTR (3R/4R), a tandem repeat we cannot type
Effect: This SNP says very little about you. Reported links between rs6323 and aggression, ADHD or psychiatric traits are weak, frequently non-replicated, and confounded by population structure and study design; the popular 'warrior gene' story is badly oversold relative to the evidence. At most this is a faint, indirect proxy that should never be read as a behavioural prediction.

rs6323 is a synonymous change at codon 297 (arginine either way), so it does not alter the MAOA protein; any influence would be regulatory or via weak linkage. Critically, the headline MAOA promoter VNTR (the famous 3-repeat vs 4-repeat "uVNTR") is a separate, length-based polymorphism: it is NOT this SNP and is not captured by it. MAOA is on the X chromosome, so males carry a single copy.

Behaviour is not genetic destiny: aggression and personality are overwhelmingly shaped by environment, upbringing and the combined tiny effects of many genes: there is no single 'violence gene'. This locus has a documented history of misuse, including ethnic stereotyping and courtroom misapplication, and must never be used to label, predict or judge anyone.

What this isn’t: Explicitly NOT a 'warrior gene', aggression, violence or criminality test, and not a personality or psychiatric diagnosis.

dbSNP: rs6323 · Chasing the 'warrior gene': why it looks like a dud (Genetic Literacy Project)

MEIS1 Restless legs and broken sleep: MEIS1 Replicated · small effect

A common variant in MEIS1 that turns up in study after study of insomnia, and much more strongly in studies of restless legs syndrome, the urge to move the legs at night. Each copy nudges the odds a little. It is more common in Finland than elsewhere in Europe.

Your genotype: Not read (Sleep & chronotype)

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: Genome-wide studies of insomnia (Hammerschlag et al. 2017; Jansen et al. 2019; Lane et al. 2019, all Nature Genetics) and of restless legs syndrome (GWAS Catalog GCST005042, GCST011995)
Effect: Each copy of the T allele raised the odds of reporting insomnia by about 14 to 27 % in samples of up to 1.3 million people, one of the strongest single signals in insomnia genetics. The same allele roughly doubles the odds of restless legs syndrome per copy, which suggests that part of its insomnia signal comes from people kept awake by restless legs. About 1 in 9 Europeans and 1 in 6 Finns carry at least one copy.

MEIS1 is a transcription factor that guides development of the brain and spinal cord, and it is one of the best-established genes behind restless legs syndrome. This variant sits inside an intron rather than changing the protein, so it most likely acts on how the gene is regulated. The route from there to restless legs or broken sleep has not been worked out.

Insomnia has many causes, from stress, pain and shift work to other sleep disorders, and hundreds of genetic variants each add a little. One allele with a modest effect does not explain anyone's sleep. Restless legs syndrome is recognised from symptoms, and most carriers of this allele never develop it.

What this isn’t: Not a diagnosis of insomnia or restless legs syndrome, and not a sign that poor sleep is fixed by your genes. Not carrying the allele does not rule either out.

Jansen et al. 2019, Nature Genetics: genome-wide analysis of insomnia in 1,331,010 individuals (PMID 30804565) · Hammerschlag et al. 2017, Nature Genetics: genome-wide association analysis of insomnia complaints (PMID 28604731) · GWAS Catalog: associations for rs113851554 (insomnia, restless legs syndrome) · dbSNP: rs113851554

WWC1 KIBRA memory-performance variant (WWC1) Mixed evidence

An intronic variant in the WWC1 (KIBRA) gene was an early, much-discussed hit for episodic memory: carriers of one allele recalled slightly more in some studies. Later work has been inconsistent, so this is an emerging, contested association rather than a settled one.

Your genotype: Not read (Cognition & memory)

How seriously to take this: the evidence

Strength: Mixed evidence
Source of the claim: GWAS and candidate-gene studies, Papassotiropoulos et al. 2006, Science; later meta-analysis (Milnik et al. 2012) with mixed results
Effect: T-allele carriers (C/T or T/T) were reported in the original and several follow-up studies to outperform C/C individuals on delayed episodic-recall tasks. A 2012 meta-analysis found the effect smaller and less consistent than first reported.

WWC1 encodes KIBRA, a protein involved in synaptic plasticity and memory formation in the hippocampus; the variant is intronic and any functional effect on memory is indirect and not fully mapped.

A genuinely famous early finding that has not replicated cleanly. Memory is highly polygenic and shaped by sleep, age, education and practice; this single marker explains very little.

What this isn’t: Not a diagnosis, not a dementia test, and not a measure of your intelligence or memory ceiling.

Papassotiropoulos et al. 2006, Science, KIBRA and human memory (PMID 17053149) · dbSNP, rs17070145

MTNR1B Glucose & melatonin crossover (MTNR1B) Replicated · small effect

A common variant in a melatonin receptor that nudges fasting blood sugar slightly upward, and sits at an unusual crossover where melatonin and late-night eating affect glucose handling more in carriers. A small statistical tilt, not a verdict.

Your genotype: Not read (Metabolism & appetite)

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: GWAS Catalog + large European meta-analyses and Asian replication cohorts: robustly replicated, small per-allele effect
Effect: Carrying the G allele is associated, on average, with very slightly higher fasting blood-glucose readings and a small nudge in long-term type-2-diabetes risk-factor profiles: a tiny statistical tilt, not a switch. MTNR1B also sits at a melatonin/circadian crossover: in carriers, melatonin (including evening doses, or melatonin-disrupting late eating and shift schedules) appears to blunt glucose handling more than in non-carriers, which is why sleep and meal timing are part of this variant's story.

The melatonin receptor MTNR1B is expressed in pancreatic beta cells; the G allele is linked to higher receptor expression and reduced glucose-stimulated insulin release, especially overnight when melatonin is high. The net effect is slightly lower early insulin output and modestly higher fasting glucose, and a genotype that interacts with melatonin exposure.

A common, low-effect-size variant: the per-copy change in glucose is small and population-averaged. Diet, weight, activity, sleep, age and many other genes dominate the actual outcome.

What this isn’t: Not a diabetes diagnosis, not a prediction you will develop diabetes, and not a reason to start or stop melatonin without a clinician.

dbSNP: rs10830963 · MTNR1B G-allele and impaired fasting glycemia / T2D (PubMed 19324940)

CHRNA5 Nicotine dependence signal (CHRNA5) Replicated · small effect

A nicotinic-receptor variant that, among people who smoke, nudges the odds of smoking more heavily and finding it harder to quit. It is one of the most reliably replicated common variants in smoking genetics, and still only a small, probabilistic nudge per person.

Your genotype: Not read (Substance response)

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: Candidate-gene + genome-wide association studies of nicotine dependence / cigarettes-per-day, with many independent replications and functional cell-line work
Effect: Carrying the A (Asn398) allele is associated, on average, with somewhat heavier smoking and a higher likelihood of nicotine dependence among people who smoke; the effect is graded (one copy less than two). It is also linked to lung-cancer risk, but largely indirectly: through heavier and longer smoking, not as an independent cancer switch. The effect sizes are population-level averages and small for any one individual.

The variant changes amino acid 398 (Asp to Asn) in the alpha-5 subunit of neuronal nicotinic acetylcholine receptors. The Asn398 form shows altered receptor function: reduced response to agonist and changes in calcium signalling and desensitisation: plausibly blunting the early aversive effects of nicotine and shifting reward and habit dynamics.

A probabilistic risk modifier, not a verdict: most of whether and how much someone smokes is driven by environment, social context and other genes, and many A-allele carriers never smoke or smoke lightly.

What this isn’t: It does not mean a person is destined to smoke, become dependent, or get lung cancer, and it is not a diagnosis.

dbSNP: rs16969968 · The CHRNA5-A3-B4 gene cluster and smoking (review, PMC5152594)

OPRM1 Mu-opioid receptor variant (OPRM1 A118G) Replicated · small effect

A much-studied change in the brain's main opioid receptor that was once thought to shape pain, reward and how people respond to opioids, naltrexone and alcohol. Those claims have not held up. What has held up, in very large genome-wide studies, is a small association with opioid use disorder.

Your genotype: Not read (Substance response)

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: Genome-wide studies of opioid use disorder (Zhou et al., JAMA Psychiatry 2020; Kember et al., Nat Neurosci 2022); for the drug-response claims, candidate-gene and clinical-trial analyses whose meta-analyses largely temper the early findings
Effect: In genome-wide studies of opioid use disorder (Zhou et al. 2020; Kember et al. 2022), the G (Asp40) allele is slightly less common among people with the disorder: an odds ratio of about 0.89 per copy in a cross-ancestry meta-analysis, passing the genome-wide threshold in several analyses. The older claims, that carriers respond differently to naltrexone, to opioid pain relief or to alcohol, come from smaller studies whose meta-analyses find the effects small, inconsistent and often non-significant after correcting for multiple testing. Either way the effect is weak at the individual level. Allele frequency varies widely by ancestry (much higher in East-Asian-ancestry populations).

The variant changes amino acid 40 (Asn to Asp) in the mu-opioid receptor and removes a putative N-glycosylation site. Proposed effects on receptor expression and binding exist, but the functional consequences in humans remain debated and not cleanly established.

One of the most famous 'promising then deflated' pharmacogenetic variants: it must not be used to choose, dose or predict response to any medication. Those decisions belong with a prescriber.

What this isn’t: It does not reliably predict how a person will respond to opioids, naltrexone or alcohol, it is not a test of addiction risk, and it is not a basis for any treatment decision.

dbSNP: rs1799971 · Kember et al. 2022, Nature Neuroscience: cross-ancestry meta-analysis of opioid use disorder (PMID 36171425) · Zhou et al. 2020, JAMA Psychiatry: OPRM1 functional coding variant and opioid use disorder (PMID 32492095) · Meta-analysis: OPRM1 rs1799971 moderating naltrexone response in AUD (PMC7340566)

OXTR OXTR: rs53576, the oxytocin receptor & social behaviour Contested

A famous (and famously contested) variant in the oxytocin-receptor gene, popularly tied to empathy and sociality. We include it precisely to show how shaky 'social genetics' can be.

Your genotype: Not read (Social & personality)

How seriously to take this: the evidence

Strength: Contested
Source of the claim: PubMed: social-genetics studies (replication disputed)
Effect: Early studies linked the G allele to higher self-reported empathy, optimism and stress resilience. Large replication attempts have often failed, and the consensus is now sceptical.

OXTR encodes the oxytocin receptor, central to social bonding. rs53576 is intronic, so any functional effect is indirect and unproven: part of why the behavioural claims do not replicate well.

This is a cautionary tale: a single intronic SNP does not meaningfully determine personality, and most early 'candidate-gene' social findings have not held up. Treat any reading as entertainment.

What this isn’t: Not a measure of your empathy, kindness or social ability: behaviour here is overwhelmingly non-genetic.

dbSNP: rs53576 · Meta-analysis: a sociability gene? (OXTR rs53576, PubMed)

PER3 Early-bird family variant: PER3 Replicated · small effect

Two linked changes in the clock gene PER3, found in a family whose members fell asleep and woke hours earlier than most people. In UK Biobank, carriers were somewhat more likely to call themselves morning people and slept about 7 minutes earlier. The pair is five times more common in Finland than elsewhere in Europe.

Your genotype: Not read (Sleep & chronotype)

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: Family study (Zhang et al., PNAS 2016) and a population test in about 1,500 UK Biobank carriers and 149 Finnish carriers (Weedon et al., PLoS Genet 2022)
Effect: In the original family, carriers had familial advanced sleep phase: a body clock running hours ahead. In UK Biobank, 30 % of carriers described themselves as 'definitely a morning person' against 24 % of non-carriers, a difference that passed the genome-wide significance threshold, and activity monitors put their sleep about 7 minutes earlier. In the Finnish cohorts the same study examined, carriers and non-carriers called themselves morning people equally often (23 % and 22 %), though only 149 carriers were counted.

PER3 is one of the period proteins of the circadian clock. The two changes, proline 415 to alanine and histidine 417 to arginine, sit next to each other and are inherited together. In cell experiments they made PER3 less stable and less able to stabilise its partners PER1 and PER2, which help set the clock's timing. Mice carrying the human version shifted their sleep-wake cycle when days were kept short.

A few minutes of average shift is far smaller than the hours seen in the family. Family members also scored higher on questionnaires about low and seasonal mood; that has not been shown in the general population, and this card says nothing about mood. The card reads one of the two changes, which population data show are inherited together.

What this isn’t: Not a diagnosis of a sleep-timing disorder and not a marker for depression or seasonal mood. Being a carrier does not make an early schedule right for you, and not carrying it says nothing about whether you are a morning person.

Zhang et al. 2016, PNAS: a PERIOD3 variant, a circadian phenotype and a seasonal mood trait (PMID 26903630) · Weedon et al. 2022, PLoS Genetics: Mendelian sleep and circadian variants in a population setting · dbSNP: rs139315125 (H417R, the change this card reads) · dbSNP: rs150812083 (P415A, inherited with it)

FUT2 Secretor status: norovirus and B12 (FUT2) Replicated · small effect

Whether you are a 'secretor', meaning you display blood-group sugars in saliva, gut and other secretions. The common nonsecretor variant is associated with strong resistance to the most widespread strains of norovirus, and with somewhat higher vitamin B12 blood levels, because the same sugars that norovirus uses to attach are absent.

Your genotype: Not read (Immunity & infection)

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: Functional and association studies on secretor status, plus GWAS (NHGRI-EBI GWAS Catalog) for vitamin B12 and norovirus susceptibility
Effect: The A allele is a stop variant (Trp154Ter); two copies (A/A) make a nonsecretor, robustly associated with resistance to the dominant GII.4 norovirus strains and with higher blood vitamin B12. G/G and G/A are secretors and susceptible to those strains.

FUT2 builds the H-antigen sugars on gut and secretion surfaces. Most noroviruses dock onto those sugars to infect; nonsecretors lack them, so the common strains cannot attach. The same biology shifts how B12 is handled in the gut.

Resistance is strain-specific, not absolute: some rarer norovirus strains still infect nonsecretors. The B12 effect is a small population-level shift. Secretor status also relates to other gut and infection traits not described here.

What this isn’t: Not a guarantee against stomach bugs and not a diagnosis or a B12-deficiency test.

NHGRI-EBI GWAS Catalog, rs601338 · dbSNP, rs601338

ADRB1 Natural short sleeper: ADRB1 Contested

A second, equally rare 'short sleeper' variant: carriers in one studied family felt rested on roughly four to six hours. When 69 carriers were found in a large population study, they slept as long as everyone else. As with the DEC2 version, nearly everyone is reference and has a normal sleep need; this is a curiosity, not a target.

Your genotype: Not read (Sleep & chronotype)

How seriously to take this: the evidence

Strength: Contested
Source of the claim: Shi et al., Neuron 2019: one large family (FNSS2); confirmed by a CRISPR knock-in mouse. Not seen in the general population: Weedon et al., PLoS Genet 2022 (69 UK Biobank carriers)
Effect: This extremely rare allele is associated with feeling rested on roughly four to six hours of sleep, less than the typical adult need. It was found in one multigenerational family. A population study later found 69 carriers among about 166,000 UK Biobank participants with sequenced exomes: they reported 7.1 hours of sleep a night on average against 7.2 for everyone else, and were no more likely to report six hours or less. So the family's short sleep did not show up in carriers from the general population, and almost everyone is reference (C/C) here anyway.

ADRB1 encodes the beta-1 adrenergic receptor; the Ala187Val change sits in a conserved region and yields a less stable receptor with reduced signalling. In mice the variant makes neurons in a wake-promoting region of the dorsal pons more active, which is thought to shorten sleep while preserving daytime function.

A rare familial short-sleep trait carried by one studied family: NOT advice or permission to sleep less. For the non-carrier majority, cutting sleep produces real sleep deprivation and health costs.

What this isn’t: Not a common 'you can run on four hours' marker, and not a population-wide sleep-need dial.

dbSNP: rs776439595 · OMIM: Short sleep, familial natural, 2 (FNSS2, 618591) · Weedon et al. 2022, PLoS Genetics: Mendelian sleep and circadian variants in a population setting

BHLHE41 Natural short sleeper: DEC2 (BHLHE41) Single study

The famous (and vanishingly rare) variant behind people who feel fully rested on about two hours less sleep than average. In the one population check, its few carriers slept as long as everyone else. Almost everyone carries the reference version and has an ordinary sleep need: this is included as a curio, not a goal.

Your genotype: Not read (Sleep & chronotype)

How seriously to take this: the evidence

Strength: Single study
Source of the claim: He et al., Science 2009: the first human short-sleep mutation; reproduced in knock-in mice and flies
Effect: This extremely rare allele is linked to a lifelong tendency to feel fully rested on roughly two hours less sleep than average, without daytime impairment. It has been reported in only a handful of families worldwide, so essentially everyone is reference (G/G) here and has an ordinary sleep need. The association rests on very small human samples, even though animal models recreate the trait. In UK Biobank, the 10 carriers found among about 166,000 sequenced participants reported 7.3 hours of sleep on average against 7.2 for non-carriers (Weedon et al. 2022): too few to settle the question, and no sign of the short-sleep trait.

DEC2/BHLHE41 is a transcriptional repressor in the circadian clock; the Pro384Arg change reduces its repressive activity. In model animals this raises arousal-promoting orexin (hypocretin) signalling in the hypothalamus, which is thought to compress sleep duration while preserving sleep quality.

Short sleep here is a rare inherited trait carried by a few families: it is NOT a license or an instruction to sleep less. For the reference (non-carrier) majority, deliberately curtailing sleep causes genuine sleep deprivation.

What this isn’t: Not a common 'sleep less and feel fine' switch you can flip, and not advice to shorten your sleep.

dbSNP: rs121912617 · ClinVar: BHLHE41 p.Pro384Arg, familial natural short sleep 1 · Weedon et al. 2022, PLoS Genetics: Mendelian sleep and circadian variants in a population setting

SLC6A4 Serotonin transporter (5-HTTLPR): the famous null Contested

The serotonin-transporter promoter: once the most famous 'depression-and-stress' result in all of behavioural genetics, and now the textbook example of a finding that didn't replicate. Included for exactly that reason: it has essentially no predictive value for any individual.

Your genotype: Not read (Cognition & stress)

How seriously to take this: the evidence

Strength: Contested
Source of the claim: Legacy candidate-gene literature (5-HTTLPR), now largely discredited for depression prediction; the 2003 stress-interaction failed large replication (Border et al. 2019)
Effect: Early studies proposed that serotonin-transporter promoter variation interacted with stressful life events to raise depression risk, and rs25531 was used to refine the classic long/short promoter calls. This famous interaction did NOT hold up: large well-powered studies and meta-analyses, culminating in Border et al. 2019 across hundreds of thousands of people, found no reliable association and no gene-by-environment effect. The plain reading is that this locus has essentially no demonstrated predictive value for depression in individuals.

SLC6A4 encodes the serotonin transporter (the SSRI target). rs25531 does not act alone: it tags and modifies the 5-HTTLPR, a roughly 43 bp insertion/deletion in the promoter (the 'long' vs 'short' allele), NOT a simple SNP. Because the real functional element is an indel that this SNP only partially indexes, a single-SNP read-out is inherently incomplete.

NOT a depression test, and it carries essentially no reliable predictive power for any individual; the large-replication evidence is that the historical claims were false positives. Mood and mental health are multi-factorial (polygenic, environmental and circumstantial) and this single position should not change how anyone thinks about their risk.

What this isn’t: Not a psychiatric diagnosis and not a prediction of depression or any mental-health outcome.

Border et al. 2019, Am J Psychiatry: the large-scale non-replication of 5-HTTLPR×stress · dbSNP: rs25531

TCF7L2 Glucose handling: the strongest common T2D variant (TCF7L2) Replicated · small effect

The single most reproducible common-variant link to type-2-diabetes risk and how the body releases insulin. Even so, the per-copy effect is small and most carriers never develop diabetes: a risk-factor nudge weighed alongside lifestyle and family history.

Your genotype: Not read (Metabolism & appetite)

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: GWAS Catalog flagship signal + many large case-control studies and meta-analyses across ancestries: the most reproducible common T2D variant, modest per-allele effect
Effect: Carrying the T allele is associated, on average, with a somewhat higher long-term chance of impaired glucose handling and type-2 diabetes: the strongest common-variant tilt known for this trait, but per copy the effect is still small in absolute terms (odds ratio roughly 1.3-1.5). It nudges a risk-factor profile, not a fate; most T-allele carriers never develop diabetes.

TCF7L2 is a transcription factor in the Wnt signalling pathway; the T allele alters regulatory activity in pancreatic islets and is linked to impaired glucose-stimulated insulin secretion and a reduced incretin (GLP-1) effect. The result is a poorer first-phase insulin response: it mainly affects insulin output rather than insulin resistance.

Even as the strongest common T2D variant, the per-allele effect is modest and population-averaged; absolute risk for any individual stays low-to-moderate. Body weight, diet, activity, age, ancestry and many other variants dominate the real-world outcome.

What this isn’t: Not a diabetes diagnosis and not a prediction that you will become diabetic: only a probabilistic risk-factor marker.

dbSNP: rs7903146 · The Role of TCF7L2 in Type 2 Diabetes (ADA, Diabetes 2021)

TERC Telomere length: the TERC region Replicated · small effect

Telomeres are the protective caps on the ends of chromosomes; they shorten as cells divide and are often called a biological clock. A common variant near TERC, part of the enzyme that rebuilds them, is one of the best-replicated genetic influences on telomere length. Longer is not simply better.

Your genotype: Not read (Aging & longevity)

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: Genome-wide association, Codd et al. 2010 (2,917 people, replication in 9,492) and Codd et al. 2013 (37,684 people, replication in 10,739), both Nature Genetics; confirmed in a TOPMed whole-genome study (Taub et al. 2022; NHGRI-EBI GWAS Catalog GCST90103979)
Effect: The T allele is associated with shorter average telomeres in white blood cells. For a closely linked marker at the same locus, each copy of the shorter-telomere allele corresponded to roughly 75 base pairs, about three and a half years' worth of normal age-related shortening.

TERC is the RNA template that telomerase, the telomere-rebuilding enzyme, copies onto chromosome ends. Variants in this region may shift how much TERC is made, and with it how well telomeres are topped up over a lifetime.

Telomere length is a double-edged measure, not a score where longer wins. Shorter-telomere alleles across several genes were linked to higher coronary-disease odds, while at this very marker the longer-telomere allele is linked to higher melanoma risk. One marker shifts average length only slightly, and age, smoking and other factors matter more.

What this isn’t: Not a measurement of your telomeres, not your biological age, and not a prediction of lifespan or of any disease.

Codd et al. 2010, Nature Genetics: Common variants near TERC are associated with mean telomere length (PMID 20139977) · Codd et al. 2013, Nature Genetics: Identification of seven loci affecting mean telomere length and their association with disease (PMID 23535734) · NHGRI-EBI GWAS Catalog: rs10936599

GCKR Triglycerides up, blood sugar down: the GCKR seesaw Replicated · small effect

A common variant in the liver's glucose-sensing machinery that pushes two blood markers in opposite directions: it tends to raise triglycerides while lowering fasting blood sugar. A tidy example of why a single variant is rarely simply 'good' or 'bad'.

Your genotype: Not read (Metabolism & appetite)

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: Fine-mapping across 12 cohorts and more than 45,000 people of several ancestries, Orho-Melander et al. 2008 (Diabetes); one of the strongest triglyceride signals in the NHGRI-EBI GWAS Catalog
Effect: The T allele (leucine at position 446) is associated with higher fasting triglycerides and, at the same time, with lower fasting glucose; it is also linked to slightly higher C-reactive protein, an inflammation marker. Each copy moves these levels by a small amount.

GCKR makes glucokinase regulatory protein, which holds the liver's glucose-processing enzyme (glucokinase) in reserve. The leucine version is less effective at restraining glucokinase, so the liver takes up and processes more glucose. That lowers blood sugar but feeds more raw material into fat production, which shows up as higher triglycerides.

Both effects are small per copy and population-averaged; diet, alcohol, weight and many other variants move triglycerides and glucose far more. Only blood tests show your actual levels, and a clinician interprets them.

What this isn’t: Not a diabetes or lipid diagnosis, and not a prediction of either. The lower-glucose and higher-triglyceride tilts do not cancel out into a verdict.

Orho-Melander et al. 2008, Diabetes: Common missense variant in GCKR associated with increased triglyceride and CRP but lower fasting glucose (PMID 18678614) · Beer et al. 2009, Human Molecular Genetics: GCKR P446L acts through increased glucokinase activity in liver (PMID 19643913) · NHGRI-EBI GWAS Catalog: rs1260326 · dbSNP: rs1260326

SLC2A9 Uric acid levels: the kidney's urate transporter (SLC2A9) Replicated · small effect

Uric acid is a normal waste product that, at high levels, can crystallise in joints as gout. SLC2A9 makes a transporter that decides how much urate the kidneys hand back to the blood, and a common variant in it is the strongest known genetic influence on urate levels.

Your genotype: Not read (Nutrition & metabolism)

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: Genome-wide association, Dehghan et al. 2008 (The Lancet; Framingham and Rotterdam with replication in ARIC), and Vitart et al. 2008 (Nature Genetics; Croatian scan with UK, Croatian and German replication); NHGRI-EBI GWAS Catalog GCST000242
Effect: The T allele was associated with lower blood uric acid in white and in black participants, and with about 40 percent lower odds of gout per copy (odds ratio 0.59) in white participants. Across studies the gene explains a few percent of the differences in urate between people, a large share for a single gene.

SLC2A9, also called GLUT9, carries urate across kidney-cell membranes, and it moved urate when expressed in frog eggs. How much urate the kidney reabsorbs instead of passing into urine sets much of the blood level, so variants that change this transporter shift urate up or down.

The strongest single gene for urate is still one input among many: diet, alcohol, body weight, kidney function, certain medicines and other genes all move urate, and most people with higher-urate genotypes never get gout. Urate is measured with a blood test that a clinician interprets.

What this isn’t: Not a gout diagnosis, not a measure of your current uric acid, and not a dietary instruction.

Dehghan et al. 2008, The Lancet: Association of three genetic loci with uric acid concentration and risk of gout (PMID 18834626) · Vitart et al. 2008, Nature Genetics: SLC2A9 is a newly identified urate transporter influencing serum urate concentration, urate excretion and gout (PMID 18327257) · NHGRI-EBI GWAS Catalog: rs16890979

ALPL Vitamin B6 levels (ALPL) Replicated · small effect

How much active vitamin B6 circulates in your blood depends partly on alkaline phosphatase, an enzyme that breaks it down. A common variant near the ALPL gene shifts that enzyme's level and, with it, your typical B6 reading.

Your genotype: Not read (Nutrition & metabolism)

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: Genome-wide association, Tanaka et al. 2009 (American Journal of Human Genetics; Italian and US cohorts with independent replication), and Hazra et al. 2009 (Human Molecular Genetics; separate US cohorts); NHGRI-EBI GWAS Catalog GCST000358 and GCST000483
Effect: Each copy of the C allele was associated with about 1.45 ng/mL lower blood vitamin B6 in the discovery study, and the region reached genome-wide significance again in an independent study. The T allele goes with lower alkaline phosphatase and slightly higher B6.

Alkaline phosphatase, made by the ALPL gene, removes the phosphate from pyridoxal 5'-phosphate, the active form of vitamin B6, as part of its normal breakdown and transport. The variant appears to influence how much of the enzyme is made, so more enzyme means a little less active B6 in the blood.

A modest shift: diet, supplements, kidney function and inflammation influence B6 readings far more. A blood test is what shows your actual level, and a clinician interprets it.

What this isn’t: Not a diagnosis of B6 deficiency or of any bone or enzyme disorder, and not a reason to change your diet or supplements.

Tanaka et al. 2009, American Journal of Human Genetics: GWAS of vitamin B6, vitamin B12, folate and homocysteine (PMID 19303062) · Hazra et al. 2009, Human Molecular Genetics: Genome-wide significant predictors of metabolites in the one-carbon metabolism pathway (PMID 19744961) · NHGRI-EBI GWAS Catalog: rs4654748

CYP2R1 Vitamin-D tendency: activation enzyme (CYP2R1) Replicated · small effect

A second vitamin-D tendency variant: this one in the liver enzyme that activates vitamin D into the measured form. A small, reproducible nudge that pairs with the GC binding-protein variant; sunlight and supplements still dominate.

Your genotype: Not read (Nutrition & metabolism)

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: GWAS Catalog / SUNLIGHT consortium large meta-analyses + replications: genome-wide significant and consistently replicated, small per-allele effect
Effect: Here the reference allele A is the favourable, higher-vitamin-D one and the common G allele is the risk allele. Each copy of the G allele is associated, on average, with modestly lower measured 25-hydroxyvitamin D and somewhat higher odds of insufficiency (roughly 1.2x per G allele in the discovery cohort). A small, reproducible nudge, not a determinant.

CYP2R1 encodes the main 25-hydroxylase that converts vitamin D into 25-hydroxyvitamin D in the liver: the form that blood tests measure. The G allele is linked to lower CYP2R1 activity/expression, so less precursor is hydroxylated into 25(OH)D, lowering the measured level.

Like the GC variant, real-world 25(OH)D is governed mostly by sun exposure, season, latitude, skin tone and supplementation; this genotype only shifts a baseline tendency.

What this isn’t: Not a vitamin-D blood test and not a diagnosis of vitamin-D deficiency.

dbSNP: rs10741657 · Common variants in CYP2R1 and GC predict vitamin D (PMC3937412)

GC Vitamin-D tendency: binding protein (GC) Replicated · small effect

A nudge in how much vitamin D your blood tends to carry, via the main protein that ferries it around. A top, reproducible genetic signal for vitamin-D level, but sunlight, season and supplements move the needle far more.

Your genotype: Not read (Nutrition & metabolism)

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: GWAS Catalog / SUNLIGHT consortium large meta-analyses (tens of thousands of Europeans) + replications: highly consistent direction, small per-allele effect
Effect: Each copy of the G allele is associated, on average, with modestly lower measured 25-hydroxyvitamin D in the blood, and a higher likelihood of falling into the insufficient range in some cohorts. T-allele carriers tend to run a little higher. The effect is real and reproducible but small at the individual level.

GC encodes vitamin-D binding protein, which carries about 85-90% of circulating 25(OH)D in the blood. rs2282679 tags variants altering this binding protein (it is tightly linked to the rs4588/rs7041 protein isoforms), changing how much 25(OH)D is held in circulation and thus the total level that is measured.

Sunlight (UVB) exposure, latitude, season, skin tone and supplementation dominate actual vitamin-D status far more than this genotype. The variant nudges a baseline tendency.

What this isn’t: Not a vitamin-D blood test and not a diagnosis of vitamin-D deficiency.

dbSNP: rs2282679 · Common variants of GC and 25(OH)D (PMC3613945)

SLC30A8 Zinc in insulin cells: the ZnT8 variant (SLC30A8) Replicated · small effect

SLC30A8 makes a zinc transporter found almost only in the pancreas's insulin-making cells. A common variant in it was one of the first type-2-diabetes signals found by genome-wide scans, and rare broken versions of the gene turned out to be protective, a twist that made it a drug-target story.

Your genotype: Not read (Metabolism & appetite)

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: Genome-wide association, Sladek et al. 2007 (Nature), replicated in large multi-ancestry type-2-diabetes GWAS (NHGRI-EBI GWAS Catalog); rare loss-of-function variants studied in about 150,000 people, Flannick et al. 2014 (Nature Genetics)
Effect: The C allele (arginine at position 325) is the common higher-risk version: each copy is associated with modestly higher odds of type 2 diabetes. The T allele (tryptophan) goes with slightly lower fasting glucose and HbA1c. Separately, people carrying rare variants that break the gene had about 65 percent lower odds of type 2 diabetes.

ZnT8 loads zinc into the granules where insulin is stored, and zinc helps pack insulin for storage. How the common amino-acid change affects the transporter is still debated, and so is why rare broken copies protect; the rare-variant result was read by its authors as a sign that turning ZnT8 down could help prevent type 2 diabetes.

A small per-copy effect on a disease driven mostly by weight, activity, age and many other genes. The higher-risk allele is the majority version in Europeans, so carrying it is ordinary. This marker is already one of the inputs to the polygenic type-2-diabetes score.

What this isn’t: Not a diabetes diagnosis, not a measure of your blood sugar, and not a prediction that you will or will not develop diabetes.

Sladek et al. 2007, Nature: A genome-wide association study identifies novel risk loci for type 2 diabetes (PMID 17293876) · Flannick et al. 2014, Nature Genetics: Loss-of-function mutations in SLC30A8 protect against type 2 diabetes (PMID 24584071) · NHGRI-EBI GWAS Catalog: rs13266634

Emerging scores

Emerging research

Not reportable Body mass index (BMI) 97-variant BMI score (Locke et al. 2015 loci, as scored by Dashti et al., BMC Medicine 2022) Emerging research

What it measures. Body mass index, a person's weight divided by the square of their height. The score adds up 97 variants that the GIANT consortium linked with BMI in 2015 (Locke et al.), as scored by Dashti et al. (2022).

Examined, not reportable: only 0 of 97 scoring variants (0%) were found in your file, below the 50% we require to place you on this score’s distribution. This is a coverage gap, not a result.

How much of the trait it explains. In 33,511 patients of European ancestry in the Mass General Brigham Biobank, the score accounted for 2.9 % of the differences in BMI between people. Each standard deviation of score went with a BMI about 0.83 kg/m² higher on average.

An emerging-research score adds up many small-effect variants into one number for a trait anyone can measure. It ranks your file against a reference group. It does not measure the trait in you, and a real measurement always outranks it. These scores are newer and less tested than the disease scores in your report, which is why they are kept apart from them.

2026-10-05 · PGS Catalog PGS002251 (restated; analytic European reference distribution under allele-frequency assumptions) · Dashti et al., BMC Medicine 2022 · 1000 Genomes Project, Nature 2015: the linkage between this score's variants, which widens its reference spread

Not reportable Chronotype (morning or evening type) 311-variant chronotype score (Jones et al. 2019 loci, as scored by Maukonen et al., Journal of Biological Rhythms 2020) Emerging research

What it measures. Chronotype: whether a person tends to be a morning type or an evening type, which is what a sleep-timing questionnaire asks about. The score adds up 311 variants that a 2019 study of 697,828 people (Jones et al., UK Biobank and 23andMe) linked with being a morning person, as Maukonen et al. (2020) scored them in Finnish adults.

Examined, not reportable: only 0 of 311 scoring variants (0%) were found in your file, below the 50% we require to place you on this score’s distribution. This is a coverage gap, not a result.

How much of the trait it explains. In 7,436 and 8,433 Finnish adults of the FINRISK study, the score accounted for roughly 1 to 2 % of the differences in chronotype between people (R² 0.013 to 0.019, as the PGS Catalog records the study's checks). In the 2019 study's activity-monitor data on 85,760 people, the 5 % carrying the most morning-linked variants slept on average 25 minutes earlier than the 5 % carrying the fewest.

2026-10-09 · PGS Catalog PGS000336 (restated; analytic European reference distribution under allele-frequency assumptions) · Maukonen et al., Journal of Biological Rhythms 2020: the score and its Finnish validation (CC BY 4.0) · Jones et al., Nature Communications 2019: the 351 chronotype loci and their effects (CC BY 4.0) · 1000 Genomes Project, Nature 2015: the linkage between this score's variants, which widens its reference spread

Not reportable Kidney filtering (eGFR) 147-variant eGFR score (Wuttke et al., Nature Genetics 2019) Emerging research

What it measures. How fast the kidneys filter blood, as estimated from a creatinine blood test (eGFR). The score adds up the 147 variants that a 2019 study of 765,348 people linked with eGFR and judged most likely to act on the kidney itself.

Examined, not reportable: only 0 of 147 scoring variants (0%) were found in your file, below the 50% we require to place you on this score’s distribution. This is a coverage gap, not a result.

How much of the trait it explains. In the source study, all 308 of its lead variants together accounted for 7.1 % of the differences in eGFR between people. This score uses 147 of them, so it accounts for less than that.

2026-10-05 · PGS Catalog PGS002810 (restated; analytic European reference distribution under allele-frequency assumptions) · Wuttke et al., Nature Genetics 2019 · 1000 Genomes Project, Nature 2015: the linkage between this score's variants, which widens its reference spread

Not reportable Height 3,286-variant height score (Yengo et al., Human Molecular Genetics 2018; GIANT consortium) Emerging research

What it measures. Adult height. The score adds up 3,286 variants that a 2018 GIANT consortium study of about 700,000 people of European ancestry linked with height, each weighted by its effect in that study.

Examined, not reportable: only 0 of 3286 scoring variants (0%) were found in your file, below the 50% we require to place you on this score’s distribution. This is a coverage gap, not a result.

How much of the trait it explains. In the source study's independent check, 8,552 unrelated adults in the US Health and Retirement Study, a score from these variants accounted for about a fifth of the differences in adult height between people (r² 0.197). A later study of the same variants in Dutch adolescents found between 8 and 14 %.

2026-10-05 · Yengo et al., Human Molecular Genetics 2018 (GIANT consortium): the variants and their weights · GIANT consortium data files: top 3,290 height SNPs, 2018 release · gnomAD v4 genomes, non-Finnish European: the effect-allele frequencies (CC0) · Xie et al., Circulation: Genomic and Precision Medicine 2020: the Dutch adolescent check · 1000 Genomes Project, Nature 2015: the linkage between this score's variants, which widens its reference spread

Not reportable Eye pressure (intraocular pressure) 101-variant eye-pressure score (MacGregor et al., Nature Genetics 2018; its 2 optic-nerve variants left out) Emerging research

What it measures. Pressure inside the eye, the number an eye examination measures (intraocular pressure). The score adds up 101 variants that a 2018 study of about 133,000 people linked with eye pressure.

Examined, not reportable: only 0 of 101 scoring variants (0%) were found in your file, below the 50% we require to place you on this score’s distribution. This is a coverage gap, not a result.

How much of the trait it explains. The source study checked its published version of the score against glaucoma rather than against measured pressure. That version adds two optic-nerve variants for which the source file gives no allele frequency, so they are left out here. In 1,734 people with advanced glaucoma and 2,938 without, it told the two groups apart only modestly (AUROC 0.65, where 0.5 is chance).

2026-10-05 · PGS Catalog PGS000124 (restated; analytic European reference distribution under allele-frequency assumptions) · MacGregor et al., Nature Genetics 2018 · 1000 Genomes Project, Nature 2015: the linkage between this score's variants, which widens its reference spread

The Fringe

Speculative · for curiosity only

Read from your file: 1 of 1. The other 0 had no variant reported in your file, so they’re read as the common (reference) genotype: an inference, not a measurement (see each card’s note).

ABCC11 Wet or dry earwax (ABCC11) Dry earwax Replicated · small effect

A single well-studied variant largely determines whether your earwax is wet or dry, and tracks the amount of underarm odour your glands produce. This is one of the cleanest single-gene human traits.

Your genotype: T/T (Body quirks)
Most-associated reading: Dry earwax

You carry two copies of the allele linked to dry, flaky earwax and fewer underarm odour precursors. Common in East Asian ancestry.

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: Functional and association studies, Yoshiura et al. 2006, Nature Genetics, the SNP in ABCC11 determines earwax type
Effect: The T allele (Gly180Arg) tracks dry, flaky earwax and reduced underarm odour; the reference C allele tracks wet, sticky earwax and more odour. Unusually for a trait, this single variant explains most of the difference. The T allele is common in East Asia and rare in Africa.

ABCC11 is a membrane transporter active in the glands that secrete earwax and underarm sweat; the T allele reduces transporter function, yielding drier wax and fewer odour precursors.

A genuinely strong single-gene effect for earwax, but underarm odour also depends heavily on skin bacteria, washing and clothing, so the body-odour link is looser than the earwax link.

What this isn’t: Not a diagnosis and not a hygiene verdict; it simply describes a gland-secretion type.

Yoshiura et al. 2006, Nature Genetics, ABCC11 earwax (PMID 16444273) · dbSNP, rs17822931

Entries we couldn't resolve from your file (26)

OR7D4 Androstenone: sweaty, sweet or nothing at all? (OR7D4) Replicated · small effect

Androstenone is a steroid found in sweat and in some pork. Some people find it rank and urine-like, some find it faintly sweet, and some cannot smell it at all. A single smell receptor, OR7D4, accounts for part of that split.

Your genotype: Not read (Taste & smell)

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: Functional receptor study plus association, Keller et al. 2007 (Nature); replicated in Norwegian meat-tasting work (Lunde 2012) and in a Han Chinese genome-wide scan (Li 2022)
Effect: The A allele marks the 'WM' version of OR7D4, which barely responds to androstenone in the lab. In the original study, people with one or two WM copies were, as a group, less sensitive to androstenone and found it less unpleasant than people with two working 'RT' copies. The effect is consistent across studies but explains only a few percent of why people differ.

OR7D4 is an odorant receptor that responds selectively to androstenone and the related steroid androstadienone. The WM version carries two amino-acid changes (R88W and T133M) that travel together almost perfectly and severely weaken the receptor's response, so less signal reaches the brain for the same amount of odour.

One receptor is part of the story, not all of it: many RT/RT people are still insensitive to androstenone, and how you describe a smell depends on concentration, context and experience. This card reads one of the two linked changes and infers the other.

What this isn’t: Not a test of your general sense of smell, not a statement about hormones or attraction, and not a health finding.

Keller et al. 2007, Nature: Genetic variation in a human odorant receptor alters odour perception (PMID 17873857) · Li et al. 2022, PLoS Genetics: From musk to body odor, decoding olfaction through genetic variation (PMID 35113854) · dbSNP: rs61729907

OR2M7 Asparagus pee: can you smell it? (OR2M7 cluster) Single study

Whether your genes tilt you toward being unable to smell the distinctive sulfurous odour in urine after eating asparagus. A variant in a dense cluster of smell-receptor genes shifts the odds; this is about perception, not anything about your health.

Your genotype: Not read (Taste & smell)

How seriously to take this: the evidence

Strength: Single study
Source of the claim: GWAS, Pelchat et al. 2011, Chemical Senses, a psychophysical and genetic study of asparagus urine odour
Effect: The reference A allele was associated with reduced ability to smell the asparagus urine odour (asparagus anosmia). The G allele tracks the ability to smell it. The per-allele effect is modest.

The variant sits at the end of chromosome 1 inside a cluster of about fifty olfactory-receptor genes, nearest OR2M7; the cluster is thought to house the receptors for the sulfurous odour compounds.

A modest, single-study signal. Whether you notice the smell also depends on how much asparagus you ate and how recently, and most of the variation is other genes.

What this isn’t: Not a diagnosis and not a sign of any smell disorder; it is one quirk at one of many smell receptors.

Pelchat et al. 2011, Chemical Senses, asparagus urine odour (PMID 20876394) · dbSNP, rs4481887

TAS2R38 Bitter taste: PTC taster or not (TAS2R38) Replicated · small effect

Whether your genes make you sensitive to certain bitter compounds (PTC and PROP) found in foods like brassicas, coffee and tonic water. This bitter-taste receptor is one of the best-studied taste genes; the marker below is its single best-characterised site.

Your genotype: Not read (Taste & smell)

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: Functional and association studies, Kim et al. 2003, Science, positional cloning of the PTC bitter-taste receptor
Effect: At this site the reference C allele (Pro49) and the G allele (Ala49) tag the two main TAS2R38 forms. The taster form tracks greater sensitivity to PTC and PROP bitterness; the non-taster form tracks reduced sensitivity. The full receptor is defined by three linked sites, of which this is the lead.

TAS2R38 encodes a bitter-taste receptor on the tongue; the amino-acid differences change how strongly the receptor responds to thiourea compounds like PTC and PROP.

A real, well-replicated effect, but bitter perception of real foods also depends on the other linked sites in this gene, on saliva and on tongue papilla density, plus habit. A single site only tags the haplotype.

What this isn’t: Not a diagnosis and not a verdict on what you should eat; it describes sensitivity to a narrow class of bitter compounds.

Kim et al. 2003, Science, PTC bitter-taste receptor (PMID 12595690) · dbSNP, rs713598

KITLG Blond vs darker hair (KITLG) Established biology · inferred call

A well-understood 'highlights' variant: the lighter allele tilts hair blonder and is one of the top contributors to classic Northern-European blondness. It's a probabilistic nudge layered on top of the bigger eye-and-hair-colour machinery.

Your genotype: Not read (Hair & appearance)

How seriously to take this: the evidence

Strength: Established biology · inferred call
Source of the claim: GWAS Catalog (Sulem 2007 / Han 2008) + Guenther et al. 2014 (Nat Genet) functional enhancer mechanism: replicated, mechanistically demonstrated
Effect: The C allele tilts hair lighter/blonder, and sits among the top contributors to classic Northern-European blondness. It works as a probabilistic nudge layered on the bigger HERC2/OCA2 eye-and-hair-colour machinery, and is characterised in European-ancestry samples. Lighter: not a guarantee of blond.

rs12821256 lies ~350 kb upstream of KITLG (KIT ligand) inside a hair-follicle enhancer; the C allele weakens an LEF1 transcription-factor binding site, lowering enhancer activity and KITLG expression in the follicle. Less KIT-ligand signalling to melanocytes yields less melanin in the hair shaft (a lighter, blonder tone) without affecting skin or eyes the way broad pigment genes do.

Hair colour is strongly polygenic; this SNP is one modest-effect regulatory tweak, and both its frequency and calibrated effect are European-ancestry-specific.

What this isn’t: Not 'the blond gene': it does not by itself dictate hair colour or override the larger pigment-gene network.

dbSNP: rs12821256 · Guenther et al. 2014, Nat Genet: a molecular basis for classic blond hair in Europeans

OR6A2 Cilantro tastes like soap (OR6A2) Single study

Whether your genes nudge you toward perceiving a soapy note in fresh coriander (cilantro). A variant in a cluster of smell-receptor genes shifts the odds a little, but exposure, cuisine and plain habit matter far more.

Your genotype: Not read (Taste & smell)

How seriously to take this: the evidence

Strength: Single study
Source of the claim: GWAS (23andMe research cohort), Eriksson et al. 2012, Flavour, a genome-wide association study of cilantro preference
Effect: The A allele was associated with slightly lower odds of reporting cilantro as soapy. The reference C allele therefore tracks the soapy perception. The effect is small and explains only a sliver of the trait.

The variant sits within a cluster of olfactory-receptor genes on chromosome 11; OR6A2, which binds several of the aldehydes that give coriander its aroma, is the proposed receptor behind the soapy note.

A small, single-cohort signal. Many people learn to enjoy cilantro with exposure regardless of genotype, and most of the variation is not this one marker.

What this isn’t: Not a diagnosis, not a verdict that you must dislike coriander, and not destiny in the kitchen.

Eriksson et al. 2012, Flavour, cilantro preference GWAS · dbSNP, rs72921001

HERC2 Blue or brown eyes (HERC2) Replicated · small effect

A single regulatory variant near the OCA2 pigment gene is the strongest single predictor of blue versus brown eyes. It is not the whole story (several genes fine-tune the shade), but this one site does most of the work.

Your genotype: Not read (Appearance)

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: Functional and association studies, Eiberg et al. 2008, Human Genetics; Sturm et al. 2008, the HERC2 blue-brown eye switch
Effect: The reference A allele tracks brown eyes; the G allele tracks blue. G/G individuals are usually blue-eyed, while one or two A alleles usually means brown or hazel. This single site is strongly predictive but not absolute.

The variant lies in an intron of HERC2 in a regulatory element controlling the neighbouring OCA2 pigment gene; the G allele reduces OCA2 expression and iris melanin, yielding lighter eyes.

Strongly predictive for the blue-brown axis, but green, hazel and intermediate shades depend on additional genes such as OCA2, TYR and SLC24A4, so the exact colour is polygenic.

What this isn’t: Not a diagnosis and not a paternity or ancestry test; it predicts a pigment tendency, not a guaranteed colour.

Eiberg et al. 2008, Human Genetics, HERC2 blue eye colour (PMID 18172690) · dbSNP, rs12913832

CD36 Can you taste fat? (CD36) Mixed evidence

Whether your genes tune how well you detect fat as a taste on the tongue, separate from texture or smell. A common variant near the fat-sensing receptor gene shifts oral fat sensitivity a little; habit and overall diet dominate.

Your genotype: Not read (Taste & smell)

How seriously to take this: the evidence

Strength: Mixed evidence
Source of the claim: Association studies, Pepino et al. 2012, Journal of Lipid Research, CD36 and oral fat perception
Effect: The A allele has been linked in several studies to lower CD36 expression and reduced oral sensitivity to fat (a higher detection threshold), while the reference G allele tracks keener fat detection. Results vary across cohorts, so the direction is not universal.

CD36 is a receptor on taste cells that binds long-chain fatty acids; lower receptor levels are thought to raise the threshold at which fat is detected as a taste.

An inconsistent association across studies and populations. Whether you notice fat in food is mostly about texture, aroma and habit, not this single marker.

What this isn’t: Not a diagnosis and not a diet prescription; it describes a small difference in oral fat detection.

Pepino et al. 2012, J Lipid Res, CD36 and oral fat perception (PMID 22045638) · dbSNP, rs1761667

IRF4 Freckles & premature greying (IRF4) Established biology · inferred call

One of the loudest single-SNP voices in cosmetic genetics: a variant linked to facial freckling, fair sun-sensitive skin, lighter childhood hair, and a tendency toward earlier hair greying. Mostly characterised in European-ancestry people, and still only one voice in a big choir.

Your genotype: Not read (Appearance & pigment)

How seriously to take this: the evidence

Strength: Established biology · inferred call
Source of the claim: GWAS Catalog (Han et al. 2008) + Praetorius et al. 2013 (Cell) enhancer mechanism: a rare case where one pigment SNP carries an outsized, mechanistically nailed-down effect
Effect: The T allele is linked to more facial freckling, fairer and more sun-sensitive skin, lighter (blonder) childhood hair, and a tendency toward earlier greying. Effect sizes are real but modest and overwhelmingly characterised in European-ancestry samples, so the read-out is most informative there. Think of it as one loud voice in a large choir, not a verdict.

rs12203592 sits in an IRF4 intron inside a melanocyte enhancer; the T allele weakens IRF4/TFAP2A-driven activation of tyrosinase (TYR), the rate-limiting melanin enzyme. Less TYR signalling means less melanin: hence lighter pigment, freckling, and (because the same machinery maintains hair-follicle melanocytes) earlier loss of hair colour.

Pigmentation and greying are highly polygenic; this one SNP explains only a slice of the variance and its allele frequencies and effects are calibrated to European-ancestry populations.

What this isn’t: Not a disease marker and not a deterministic predictor: a probabilistic, cosmetic-trait nudge.

dbSNP: rs12203592 · Praetorius et al. 2013, Cell: IRF4 enhancer regulates pigmentation via TYR

OR2J3 Smell of cut grass: cis-3-hexen-1-ol (OR2J3) Mixed evidence

The green, freshly-mown smell of cut grass comes largely from one compound, cis-3-hexen-1-ol, which is also used to give foods a 'green' note. A change in one smell receptor was linked to how easily people detect it, though later studies have been less consistent.

Your genotype: Not read (Taste & smell)

How seriously to take this: the evidence

Strength: Mixed evidence
Source of the claim: Candidate-region association plus receptor assays, McRae et al. 2012 (Chemical Senses, 52 people); nominal replication in a Han Chinese cohort but not in a second validation cohort (Li et al. 2022)
Effect: The G allele (alanine at position 113) weakened the receptor's response to cis-3-hexen-1-ol in cell assays, and in the discovery study it was associated with needing a higher concentration to detect the smell. A haplotype carrying this change plus a second one (R226Q) explained about a quarter of detection differences in that small study. A larger 2022 study saw the association in one cohort only.

OR2J3 is an odorant receptor that responds to cis-3-hexen-1-ol. The threonine-to-alanine change at position 113 reduces that response, and together with a second change (R226Q) it nearly abolishes it in the lab. Fewer working receptors means a weaker signal for the same amount of odour.

The discovery study was very small and the follow-up evidence is mixed: treat this as a lead, not a settled result. Several other receptors in the same chromosome region can also respond to this compound, and this card does not read the second change (R226Q).

What this isn’t: Not a test of your sense of smell in general, and not a health finding of any kind.

McRae et al. 2012, Chemical Senses: OR2J3 and detection of the grassy-smelling odour cis-3-hexen-1-ol (PMID 22714804) · Li et al. 2022, PLoS Genetics: replication attempt across two cohorts (PMID 35113854) · dbSNP: rs28757581

OR11H7 Sweaty feet and ripe cheese: isovaleric acid (OR11H7) Preliminary

Isovaleric acid is the pungent note in sweaty feet and some strong cheeses. About half of the copies of the smell-receptor gene OR11H7 in the population are broken by a stop signal. People with two broken copies were rarely among the most sensitive smellers in the one study that looked.

Your genotype: Not read (Taste & smell)

How seriously to take this: the evidence

Strength: Preliminary
Source of the claim: Candidate-gene association plus receptor expression assay, Menashe et al. 2007 (PLoS Biology), 377 people; not genome-wide significant and not independently replicated
Effect: At this site the reference T creates a stop signal that truncates the receptor; the C allele restores a working receptor. In the study, people with two broken copies were markedly under-represented among those with heightened sensitivity to isovaleric acid. Having one or two working copies was associated with a better chance of detecting it at low concentrations.

OR11H7 is a 'segregating pseudogene': some people carry a working version of the receptor and some carry a broken one. When the authors repaired the stop codon and expressed the receptor in frog eggs, the intact version responded to isovaleric acid, which fits the receptor being one of the sensors for this smell.

An early, single-study finding in a few hundred people. The same study found that people sensitive to one smell tend to be sensitive to several, so general smell acuity also plays a large part. The rsID for the stop change was matched to the paper by its coordinate; the paper itself describes it by position.

What this isn’t: Not a hygiene or health finding, and not a test of your sense of smell in general.

Menashe et al. 2007, PLoS Biology: Genetic elucidation of human hyperosmia to isovaleric acid (PMID 17973576) · dbSNP: rs1953558

TENM2 Rage at chewing sounds: misophonia (TENM2) Single study

The marker behind 23andMe's famous 'filled with rage by the sound of someone chewing' study: the genetics of misophonia. About 1 in 5 surveyed people reported that reaction, and this spot was the standout hit. A fun 'is it just me?' curiosity, not a diagnosis.

Your genotype: Not read (Senses & quirks)

How seriously to take this: the evidence

Strength: Single study
Source of the claim: One very large 23andMe self-report GWAS (Front. Neurosci. 2022; ~80k+ European-ancestry participants): genome-wide significant, but a single consumer cohort, self-reported, intronic
Effect: This variant sits near TENM2, the gene 23andMe flagged in its 'rage at the sound of chewing' study. About 1 in 5 surveyed customers reported that reaction, and this marker came up as the standout hit. It's a fun curiosity, not a clinical result, and almost entirely European-ancestry data.

TENM2 (teneurin-2) is a cell-adhesion protein highly active in neurons during brain development and wiring. The hypothesis is that variation here subtly tunes circuits handling sound and emotion, possibly heightening reactions to repetitive noises, but the variant is intronic and its functional effect is unproven.

This rests on a single self-reported GWAS of one symptom, not a clinical misophonia diagnosis, and almost entirely European-ancestry data. Association is not causation, and the intronic variant's biology is unknown.

What this isn’t: Not a misophonia test or diagnosis: a statistical nudge tied to one self-reported sound-rage question.

dbSNP: rs2937573 · 23andMe misophonia GWAS (Front. Neurosci. 2022)

IL21 Mosquito magnet? Self-reported bites (IL21 region) Single study

Some people swear mosquitoes single them out. A large 23andMe survey found that the same stretch of DNA near immune genes on chromosome 4 was linked to how big bites get, how much they itch and whether people think mosquitoes leave them alone.

Your genotype: Not read (Body quirks)

How seriously to take this: the evidence

Strength: Single study
Source of the claim: GWAS of self-reported mosquito traits, Jones et al. 2017 (Human Molecular Genetics), 23andMe research participants; NHGRI-EBI GWAS Catalog GCST004864 (16,576 people for perceived attractiveness)
Effect: Each copy of the T allele raised the odds of describing yourself as rarely bitten by about a quarter (odds ratio 1.27). The same haplotype was also the strongest signal for smaller bite reactions in about 85,000 people. The authors' analysis suggests people who react less to bites may simply notice fewer of them, so 'less attractive' may partly mean 'less reactive'.

The variant sits between IL21, an immune signalling gene, and BBS12, and it affects how strongly nearby genes are expressed. A plausible reading is that it shapes the skin's immune reaction to mosquito saliva, which determines how visible and itchy a bite becomes. Whether it changes how attractive you actually are to mosquitoes was not measured.

One cohort, entirely self-reported, and never repeated by an independent group. People judge 'how often I get bitten' mostly by the welts they notice afterwards, which is exactly what this region affects, so the genetics may be telling you about your skin reaction more than about the mosquitoes.

What this isn’t: Not a measure of protection against mosquito-borne infections, not an allergy test, and not a health finding.

Jones et al. 2017, Human Molecular Genetics: GWAS of self-reported mosquito bite size, itch intensity and attractiveness to mosquitoes (PMID 28199695) · NHGRI-EBI GWAS Catalog: rs309403

6p24 Motion sickness (carsick / seasick tendency) Single study

Whether your genes tilt you toward feeling carsick, seasick or queasy reading in a moving vehicle. One of many common variants nudges the odds; motion sickness is highly polygenic and very situational.

Your genotype: Not read (Quirks & reflexes)

How seriously to take this: the evidence

Strength: Single study
Source of the claim: GWAS (NHGRI-EBI GWAS Catalog): Hromatka et al. 2015, 23andMe research cohort (PMID 25628336)
Effect: The G allele was associated with modestly higher odds of motion sickness in a very large self-reported study. It is one of many small-effect markers: each one, including this, explains only a sliver of why some people get carsick and others don't.

Motion sickness is thought to arise from conflicting signals between the inner ear and the eyes; the associated markers point loosely toward inner-ear development and neurological processing, with no single clear mechanism.

A modest, real association from a large GWAS, but motion sickness is highly polygenic and dominated by the situation: the boat, the back seat, the book. This SNP barely moves the needle for any one person.

What this isn’t: Not a diagnosis and not a prediction that you will (or won't) be sick on the next ferry.

NHGRI-EBI GWAS Catalog: rs2153535 (motion sickness) · Hromatka et al. 2015, Human Molecular Genetics: GWAS of motion sickness (PMID 25628336) · dbSNP: rs2153535

OR4D6 Can you smell laundry musk? Galaxolide (OR4D6) Single study

Galaxolide is a synthetic musk used in laundry detergents, fabric softeners and perfumes. A noticeable share of people barely smell it. A 2022 study traced much of that to one smell receptor, OR4D6.

Your genotype: Not read (Taste & smell)

How seriously to take this: the evidence

Strength: Single study
Source of the claim: Genome-wide scan, Li et al. 2022 (PLoS Genetics): discovery in 1,000 Han Chinese, validated in an ethnically diverse cohort of 364; not yet repeated by an independent group
Effect: The C allele (threonine at position 263) was associated with weaker perceived intensity of Galaxolide, at genome-wide significance in the discovery cohort and again in the validation cohort. In the study, people homozygous for the variant haplotype were, with few exceptions, unable to smell Galaxolide at all. The top variants explained about 13 percent of the differences between people.

OR4D6 is an odorant receptor; the study proposes it as a receptor for musk compounds. The variant travels with a second change in the same gene (S151T), and the two together appear to reduce or remove the receptor's contribution, which would leave Galaxolide faint or undetectable.

One study, though it did include its own validation cohort. The receptor did not respond to Galaxolide in the authors' cell assay, so how the variant works is still unexplained. Musk perception also varies with concentration and with the specific musk compound.

What this isn’t: Not a test of your sense of smell overall and not a health finding. It says nothing about other musks.

Li et al. 2022, PLoS Genetics: From musk to body odor, decoding olfaction through genetic variation (PMID 35113854) · dbSNP: rs1453541

ZEB2 Sun sneeze: the ACHOO reflex (near ZEB2) Replicated · small effect

Whether your genes tilt you toward sneezing when you step into bright light, sometimes called the photic sneeze reflex or ACHOO. A common variant near a neural-development gene nudges the odds; the reflex is harmless and situational.

Your genotype: Not read (Quirks & reflexes)

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: GWAS (23andMe research cohort), Eriksson et al. 2010, PLoS Genetics; replicated in a Chinese cohort, Wang et al. 2019
Effect: The reference C allele was associated with higher odds of the photic sneeze reflex. The effect is modest and the reflex is a curiosity.

The variant lies in an intergenic region of 2q22 near ZEB2, a gene involved in neural-crest and cranial-nerve development; the precise mechanism for light-triggered sneezing is not settled.

A modest, replicated association, but light-triggered sneezing is influenced by many factors and the variant only nudges the odds.

What this isn’t: Not a diagnosis and not a medical condition; the photic sneeze reflex is a harmless quirk.

Eriksson et al. 2010, PLoS Genetics, web-based GWAS of common traits (PMID 20585627) · dbSNP, rs10427255

TAS2R19 How bitter is tonic water? Quinine taste (TAS2R19) Replicated · small effect

Quinine is what makes tonic water bitter. How intense it tastes varies from person to person, and a common change in the bitter-taste receptor TAS2R19 accounts for a small part of that.

Your genotype: Not read (Taste & smell)

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: Genome-wide association, Reed et al. 2010 (Human Molecular Genetics), 1,457 Australian twins and siblings, confirmed in an independent US twin sample; a separate lab study tied a TAS2R19 variant to grapefruit-juice bitterness and liking (Hayes et al. 2011)
Effect: The A allele (cysteine at position 299) was associated with tasting quinine as more intense, in both the discovery and the replication sample. The region explained about 6 percent of the differences in quinine ratings, so most of the variation lies elsewhere.

TAS2R19 is one of about 25 bitter-taste receptors on the tongue. It sits in a tightly linked cluster of bitter-receptor and salivary-protein genes on chromosome 12, and the authors note that this linkage stops them naming a single causal variant: the R299C change may do the work itself or travel with a neighbour that does.

A modest, replicated effect from a region where several similar genes are inherited together; a later study pointed at the neighbouring receptor TAS2R31 as a possible driver. How bitter tonic water tastes also depends on sugar, temperature and familiarity.

What this isn’t: Not a health finding, and not a statement about any medicine that contains quinine.

Reed et al. 2010, Human Molecular Genetics: Quinine taste intensity and a bitter receptor cluster on chromosome 12 (PMID 20675712) · Hayes et al. 2011, Chemical Senses: TAS2R variation and bitter beverages, including grapefruit juice (PMID 21163912) · Hayes et al. 2015, Chemical Senses: Quinine bitterness and grapefruit liking associate with allelic variants in TAS2R31 (PMID 26024668) · NHGRI-EBI GWAS Catalog: rs10772420

MC1R Red hair & freckling (MC1R R151C) Established biology · inferred call

One of the famous 'ginger' switches in MC1R, the master dial for red hair, pale skin and freckles. One of the strongest single-SNP cosmetic effects in the genome, but the very same variant also means more sun-sensitivity, so it carries a real health note, not just a fun fact.

Your genotype: Not read (Appearance & pigment)

How seriously to take this: the evidence

Strength: Established biology · inferred call
Source of the claim: Large GWAS + decades of MC1R pigmentation genetics and functional receptor studies: R151C is one of the classic, heavily replicated red-hair 'R' alleles (red-hair OR ~12)
Effect: Carrying two loss-of-function copies strongly tilts toward red or strawberry-blond hair, fair freckly skin, and a tendency to burn rather than tan; one copy often shows up as freckles, lighter colouring, or red highlights in beard/body hair. It is one of the strongest single-SNP cosmetic effects in the genome, though other MC1R variants and genes also vote (it acts roughly recessively for red hair).

MC1R is a receptor on melanocytes that, when activated, switches pigment production toward dark brown/black eumelanin. The R151C change cripples receptor signalling, so cells default to making reddish-yellow pheomelanin instead: the pigment behind red hair and freckles. Less eumelanin also means less natural UV shielding in the skin.

Health note: the same loss-of-function alleles that produce red hair and fair skin also carry meaningfully higher sun-sensitivity and elevated melanoma (and other skin-cancer) risk, even in carriers who don't look obviously red-haired. This is a real reason that sun protection matters more, not just a cosmetic curio.

What this isn’t: Not a complete red-hair test and not a melanoma diagnosis: one major variant among several MC1R alleles, and hair, skin tone and cancer risk each depend on more than this site.

ClinVar/dbSNP: rs1805007 (MC1R R151C) · dbSNP: rs1805007

TRPV1 Chili-heat sensitivity (TRPV1) Mixed evidence

A variant in the actual protein that fires when capsaicin (the 'heat' in chili) hits your tongue. A fun candidate for 'why is everyone at the table sweating but me?', though the genetic effect is small and the studies disagree.

Your genotype: Not read (Taste & smell)

How seriously to take this: the evidence

Strength: Mixed evidence
Source of the claim: Small candidate-gene/functional studies on capsaicin sensitivity (not a large GWAS): direction inconsistent across studies
Effect: TRPV1 is literally the protein that fires when capsaicin hits your tongue, so it's a fun candidate for chili tolerance. Some studies suggest the 585Val (C) form is associated with slightly different sensitivity to capsaicin and warmth, but studies disagree on which allele raises or lowers sensitivity. Treat any personal read as a curiosity, not a verdict on your hot-sauce tolerance.

TRPV1 is a heat- and capsaicin-gated ion channel on sensory neurons; activating it produces the burning sensation. The Ile585Val substitution sits in the channel protein and may subtly tweak how readily it opens, plausibly nudging perceived heat intensity.

Studies disagree on which allele raises or lowers sensitivity, and effects are small versus huge non-genetic factors (habituation, diet, the specific chili). Don't over-read your genotype.

What this isn’t: Not a measure of how much you 'like' spicy food or your overall pain tolerance: only one input into capsaicin perception.

dbSNP: rs8065080 · Ensembl: rs8065080 (TRPV1 Ile585Val)

TAS1R3 How sweet is sweet? Sucrose sensitivity (TAS1R3) Single study

Whether your genes make table sugar taste more or less intense to you. A variant in the promoter of the sweet-taste receptor gene tunes how sensitive your tongue is to sucrose. This is about perceived intensity, not how much you like sweets.

Your genotype: Not read (Taste & smell)

How seriously to take this: the evidence

Strength: Single study
Source of the claim: Genetic and functional study, Fushan et al. 2009, Current Biology, a TAS1R3 promoter variant and sucrose sensitivity
Effect: The C allele tracks higher promoter activity and greater sensitivity to sucrose; the reference T allele tracks lower activity and reduced sweet sensitivity. The site was reported to explain roughly a sixth of the population variability in sucrose perception.

TAS1R3 is part of the sweet-taste receptor; the variant sits in the gene's promoter and changes how much receptor is made, tuning the strength of the sweet signal.

A notable single-cohort signal, but how sweet a food seems also depends on temperature, the other sweet-receptor subunit and what you are used to. Sensitivity is not the same as preference.

What this isn’t: Not a diagnosis and not a measure of how much you crave sugar; it describes perceived intensity.

Fushan et al. 2009, Current Biology, TAS1R3 and sucrose sensitivity (PMID 19559618) · dbSNP, rs35744813

FGF21 Sweet tooth (FGF21 'sugar hormone') Replicated · small effect

Whether your genes nudge you toward a sweet tooth. FGF21 is a liver hormone that helps regulate sugar intake; a common variant near it shifts the odds a little, but habit, culture and appetite dominate.

Your genotype: Not read (Taste & appetite)

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: GWAS (NHGRI-EBI GWAS Catalog): FGF21 macronutrient / sweet-food-liking studies (Søberg et al. 2017 and later relative-sugar-intake GWAS)
Effect: The A allele was associated with a slightly higher relative sugar intake, and the G allele with slightly lower liking of sweet foods. The per-allele effect is tiny: a faint nudge on a behaviour that is overwhelmingly about habit, availability and appetite.

FGF21 is a hormone released by the liver that acts on the brain to help regulate sugar and macronutrient preference; common variation near the gene appears to fine-tune that signal.

A small, real, but easily-overwhelmed effect. Your sweet tooth is mostly your kitchen and your habits, not this one SNP.

What this isn’t: Not a diagnosis, not a diet prescription, and not a verdict on your willpower or your weight.

NHGRI-EBI GWAS Catalog: rs838133 · dbSNP: rs838133

CA6 Tongue taste-bud density: gustin (CA6) Single study

Whether your genes nudge how many fungiform papillae (the little taste-bud-bearing bumps) you have on your tongue, which tracks loosely with being a stronger or weaker taster. A variant in the salivary gustin gene shifts the odds; it pairs with, but is separate from, the bitter-taste gene.

Your genotype: Not read (Taste & smell)

How seriously to take this: the evidence

Strength: Single study
Source of the claim: Functional and association studies, Melis et al. 2013, PLoS One, gustin (CA6) and fungiform papilla density
Effect: The reference A allele (Ser90) tracks the more functional protein, higher fungiform papilla density and a stronger-taster tendency; the G allele (Gly90) tracks fewer papillae and a weaker-taster tendency. The effect is modest and from a small number of cohorts.

CA6 (gustin) is a protein in saliva thought to support the growth and maintenance of fungiform papillae; the amino-acid change alters its activity, affecting papilla density and so taste-bud number.

A modest, single-group signal. Papilla counts vary for many reasons and the link to overall taste strength is loose; the bitter-taste receptor gene plays a larger role for bitterness itself.

What this isn’t: Not a diagnosis and not a verdict on whether you are a supertaster; it is one small input to taste-bud density.

Melis et al. 2013, PLoS One, gustin (CA6) and fungiform papillae · dbSNP, rs2274333

EDAR Hair thickness and shovel-shaped teeth (EDAR) Replicated · small effect

A single variant that arose in ancient East Asia and spread to Native American populations shifts several body traits at once: thicker, straighter hair fibres, scooped (shovel-shaped) front teeth, more sweat glands and subtle face shape. It is rare in European and African ancestry.

Your genotype: Not read (Appearance)

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: Functional and association studies, Kimura et al. 2009, American Journal of Human Genetics; mouse-model work by Kamberov et al. 2013
Effect: The G allele (the derived 370A form) tracks thicker and straighter hair shafts, shovel-shaped incisors, more eccrine sweat glands and subtle changes in chin and ear shape. The reference A allele tracks the more typical European and African pattern. The G allele is at very high frequency in East Asian and Native American ancestry and near-absent elsewhere.

EDAR is a receptor that guides the development of hair follicles, teeth and sweat glands; the derived allele increases EDAR signalling, producing the linked set of traits.

A real, well-replicated set of effects, but each trait also depends on other genes, age and grooming. The associations are clearest in populations where the allele is common.

What this isn’t: Not a diagnosis and not an ancestry test, though the allele frequency varies strongly by ancestry.

Kimura et al. 2009, Am J Hum Genet, EDAR and shovel incisors (PMID 19804850) · dbSNP, rs3827760

TAS1R1 Umami (savory) taste sensitivity (TAS1R1) Single study

Whether the savory 'fifth taste' of broth, parmesan, soy sauce and ripe tomatoes leaps out at you or barely registers. This TAS1R1 spot is one nudge in that dial: a fun hint, not a verdict on your palate.

Your genotype: Not read (Taste & smell)

How seriously to take this: the evidence

Strength: Single study
Source of the claim: Functional cell-based + small psychophysical taste-threshold studies (Shigemura 2009); later cohorts give mixed results: suggestive rather than established
Effect: The A (Thr372) version of the umami receptor has been linked in lab and small taste-test studies to a more sensitive umami receptor: picking up savoury flavours at lower concentrations. The effect is modest and one of many ingredients, so treat it as a fun hint, not a verdict. Replication has been inconsistent across populations.

TAS1R1 pairs with TAS1R3 to form the tongue's umami receptor, which detects glutamate (and is boosted by nucleotides like those in dried mushrooms and cured fish). The Ala372Thr swap sits in the receptor's large extracellular domain, and the Thr version appears to make the receptor respond at lower glutamate levels.

Umami perception is polygenic and strongly shaped by culture, diet and age, so a single SNP explains only a sliver. Replication has been inconsistent across populations.

What this isn’t: Not a measure of how much you like savoury food or your overall 'supertaster' status: only one small genetic input to umami sensitivity.

dbSNP: rs34160967 · Shigemura et al. 2009, PLOS ONE: individual differences in human umami taste

OR51B2 How strongly you smell body odour's signature acid (OR51B2) Single study

Trans-3-methyl-2-hexenoic acid is one of the compounds that gives underarm odour its characteristic smell, and roughly a quarter of people are reported to have a specific blind spot for it. A 2022 study tied part of the difference to a smell receptor called OR51B2.

Your genotype: Not read (Taste & smell)

How seriously to take this: the evidence

Strength: Single study
Source of the claim: Genome-wide scan, Li et al. 2022 (PLoS Genetics): discovery in 1,000 Han Chinese, validated in an ethnically diverse cohort of 364; fine-mapping left this missense change as the only variant in the credible set
Effect: The A allele (a leucine-to-phenylalanine change in OR51B2) was associated with rating the compound as more intense, at genome-wide significance in the discovery cohort and again in the validation cohort. The novel variants at this locus explained about 4 percent of the differences between people.

OR51B2 is an odorant receptor in a cluster on chromosome 11 that responded to the compound in the authors' cell assay. Unexpectedly, the variant version that goes with stronger perception did not respond in the same assay, so the link between the receptor change and what people smell is not yet explained.

One study with its own validation cohort, and a mechanism that runs against expectation. Body odour is a mixture of more than a hundred compounds, so this one receptor covers only a small part of how sweat smells to you. The more-intense allele is the majority one in Europeans.

What this isn’t: Not a statement about how you smell to others, not a hygiene or health finding, and not a test of your sense of smell in general.

Li et al. 2022, PLoS Genetics: From musk to body odor, decoding olfaction through genetic variation (PMID 35113854) · dbSNP: rs10837814

PAX3 Eyebrow convergence: the 'unibrow' nudge (PAX3) Single study

A marker the big Latin-American facial-features GWAS tied to how close together your eyebrows grow: your odds of a faint (or frank) middle-of-the-forehead bridge. A gentle statistical nudge, not a verdict, and grooming always has the final say.

Your genotype: Not read (Face & features)

How seriously to take this: the evidence

Strength: Single study
Source of the claim: Adhikari et al. 2016 (Nat Commun): genome-wide significant at the PAX3 monobrow locus in ~6,000 admixed Latin Americans; biologically coherent (PAX3→Waardenburg synophrys), not broadly replicated across ancestries
Effect: This marker sits in DNA downstream of PAX3 that the GWAS tied to how close together your eyebrows grow. The A allele nudges toward greater eyebrow convergence. The effect is a gentle statistical nudge, not a verdict: plenty of people with the 'more-convergent' allele have perfectly tidy brows. Discovered in admixed Latin Americans, so the effect size in European/Finnish brows may differ.

PAX3 is a master regulator of neural-crest derivatives, including pigment cells and the patterning of facial/hair structures; rare coding PAX3 mutations cause Waardenburg syndrome, in which joined brows (synophrys) is common. This common variant likely tweaks PAX3 regulation subtly rather than breaking it, dialling brow-hair distribution up or down.

A regulatory/intergenic common variant, not the Waardenburg disease mutation: it carries no disease implication on its own, and the effect estimates come from one ancestry group.

What this isn’t: Not a diagnosis of any syndrome, and it does not predict hearing, eye colour or any health outcome.

Adhikari et al. 2016, Nat Commun: facial & scalp-hair GWAS in Latin Americans · dbSNP: rs2395845

OR5A1 Can you smell violets? Beta-ionone (OR5A1) Replicated · small effect

Whether your genes make you highly sensitive to beta-ionone, the floral, violet-like aroma used in perfumes and added to some foods and wines. A single variant in one smell receptor explains most of this near-on-or-off difference.

Your genotype: Not read (Taste & smell)

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: GWAS and functional study, Jaeger et al. 2013, Current Biology, a Mendelian olfactory-sensitivity trait
Effect: The reference G allele tracks high sensitivity: carriers of one or two G alleles detect beta-ionone at far lower concentrations than A/A individuals, and tend to describe it as fragrant, floral and violet-like. The A allele (Asn183) tracks low sensitivity. This single site explains most of the variation.

OR5A1 is an odorant receptor; the amino-acid change at position 183 alters how strongly the receptor responds to beta-ionone, behaving almost like a single-gene on-or-off switch.

An unusually strong single-gene effect for a smell trait, but how you describe an aroma also depends on context, concentration and experience.

What this isn’t: Not a diagnosis and not a general verdict on your sense of smell; it is specific to one floral compound.

Jaeger et al. 2013, Current Biology, Mendelian olfactory sensitivity (PMID 23910657) · dbSNP, rs6591536

Ancestry

Your ancestry composition

The autosomal picture: how much of your genome resembles each of our reference groups, read across 27,936 ancestry-informative markers. This is a resemblance to sampled reference populations, not a nationality and not a family tree.

Aimosti · aadr-v66.p1-HO/k31-chip-aj/2026-10 · 27,936 markers

This is the mix of reference populations that matches your file most closely: each region shows its share, and the populations inside it show a range only.

brighter is more of your ancestry softer edges mean a wider range the dot's area is your percentage

Western and Southern Europe 80% 72–92
Northwest European about 65–95 %
Other Western and Southern Europe references, which your file cannot tell apart
Finland, the Baltics and Eastern Europe 15% 6–28
Baltic Finnic about 0–30 %
Other Finland, the Baltics and Eastern Europe references, which your file cannot tell apart

0255075100

On the lighter end of each bar. The top of each range includes the 5 % we could not place. Some of it may belong to any region above, so the ranges admit it rather than hide it.

Unresolved 5% —

The panel is a closed list: a population it doesn't carry shows up as the references nearest to it. In tests on genomes built from real reference people, a region's range held the true share about four times in five on chip data and two times in three on whole genomes.

Every figure above, and the reference people behind each region.
Region / reference	Estimate	Range	Reference populations
Western and Southern Europe	80 %	72–92 %
Northwest European	—	about 65–95 %	354 people: British (Kent, England) · English · French · Icelandic · Northern and Western European (Utah) · Norwegian · Orcadian · Scottish
Italian–Balkan	—	not separable	333 people: Albanian · Bulgarian · Croatian · Gagauz · Greek · Italian Central · Italian North · Italian South · Maltese · Moldavian · Romanian · Serb (Bosnia) · Serb (Montenegro) · Serb (Serbia) · Sicilian · Tuscan (Italy)
Iberian	—	not separable	283 people: Iberian (Spain) · Spanish · Spanish North
Basque	—	not separable	54 people: Basque
Finland, the Baltics and Eastern Europe	15 %	6–28 %
Baltic Finnic	—	about 0–30 %	142 people: Estonian · Finnish · Finnish (Finland) · Karelian · Veps
East European	—	not separable	173 people: Belarusian · Czech · Hungarian · Lithuanian · Polish · Russian · Ukrainian · Ukrainian North
Volga–Ural	—	not separable	170 people: Bashkir · Besermyan · Chuvash · Khanty · Komi Zyrian · Mansi · Mordovian · Tatar Kazan · Tatar Mishar · Udmurt
Unresolved	5 %	—	Not placed; admitted at the top of every range above

Your two deepest lines

Your deep maternal (mitochondrial) and paternal (Y-chromosome) lines as haplogroups, restated from the published PhyloTree (mtDNA) and ISOGG (Y-DNA) trees. A haplogroup traces one single line, a single thread out of the thousands of ancestors who make you. It is not your ethnicity, your nationality, or a "percent ancestry" figure. It answers a different question from the percentages above, and neither is a better version of the other.

Maternal line · mtDNA Haplogroup H1a1

Your maternal line belongs to haplogroup H, the most common maternal lineage in Europe.

Lineage: N ▸ R ▸ R0 ▸ HV ▸ H ▸ H1 ▸ H1a1
Placement quality: 1.00 · 7 of 7 expected markers found
Resolution: genotyping chip · 7 positions measured

Your chip measures 7 mitochondrial positions; Haplogroup H1a1 is the finest branch it can separate (H1a1, H1a1a, H1a1a1 and 2 more read the same on it).

H1 is the largest branch inside H, close to a third of it in the Loogvali survey, and it carries a single coding-region change at position 3010. It is thickest in the southwest. Achilli and colleagues put its frequency peak among the Basques of Spain, better than a quarter of the samples there, and read that pattern as late-glacial hunter-gatherers spreading out of the Franco-Cantabrian refuge from about 15,000 years ago. Pereira and colleagues date the branch to roughly 14,000 years on coding-region data. Frequencies thin out to the north and east; Tambets and colleagues judged that H1 reached Fennoscandia by a western route, and its small Finnish sub-branch H1f is almost absent elsewhere in Europe.

H expanded across Europe after the last Ice Age and today accounts for roughly four in ten maternal lines in many European populations, Finland among them. It is also found across the Near East, North Africa and Central Asia. Its subclades are numerous and regional: H1 and H3 are thickest in the southwest, H5 and H11 in central Europe, and several smaller ones concentrate in Fennoscandia.

This describes one single thread of your ancestry, the unbroken maternal line of your mother's mother's mother, back through deep time. It is not your ethnicity or a percentage of where your family is "from"; it is one lineage out of the thousands that make you.

Haplogroup H is a branch of the mitochondrial tree defined by a set of changes that everyone on it carries. It sits inside HV and holds so much of the European maternal tree that the finer branch below it, rather than H itself, is what distinguishes one European line from another.

A haplogroup traces one maternal line only: it is not ethnicity, nationality, or a percent-ancestry figure.

Haplogrep 3.3.2 · PhyloTree Build 17.2 · PhyloTree: the mtDNA tree (Build 17) · MedlinePlus Genetics: Mitochondrial DNA · Achilli et al. 2004, Am J Hum Genet: The molecular dissection of mtDNA haplogroup H confirms that the Franco-Cantabrian glacial refuge was a major source for the European gene pool · Pereira et al. 2005, Genome Res: High-resolution mtDNA evidence for the late-glacial resettlement of Europe from an Iberian refugium · Loogvali et al. 2004, Mol Biol Evol: Disuniting uniformity: a pied cladistic canvas of mtDNA haplogroup H in Eurasia · Tambets et al. 2004, Am J Hum Genet: The western and eastern roots of the Saami

Paternal line · Y-DNA Haplogroup I1

Your paternal line belongs to haplogroup I1, a characteristically Nordic branch within the older haplogroup I.

Lineage: IJ ▸ I ▸ I1
Diagnostic markers: 2 of 2 present

I1 reaches its highest frequencies in Scandinavia and Finland and is common across northern Europe. It is a relatively young, rapidly expanded branch of the deeper, older haplogroup I.

This traces one single line of your ancestry (your father's father's father, back through deep time) and nothing more. It is not your ethnicity or a percentage of where your family is "from"; it is one lineage among the many that make you.

Haplogroup I1 is a branch of the older haplogroup I: your line carries I's defining M170 change and the additional M253 change that marks I1, which is why we can place you in the more specific clade. Finer I1 subclades need markers beyond this v1 set and are a planned addition.

A haplogroup traces one paternal line only, not ethnicity, nationality, or a percent-ancestry figure. I1 is a branch of I; finer I1 subclade detail is a planned addition.

ISOGG 2019–2020 · ISOGG Y-DNA Haplogroup Tree (2019–2020) · MedlinePlus Genetics: Y chromosome

What this module does not do (3)

Named on purpose: the gaps are the trust signal. A haplogroup is one line, not a whole family tree.

The written story of every branch. Haplogrep places the maternal line on the full PhyloTree, which is thousands of branches deep, and we have authored prose for a few dozen of them so far. A lineage outside that set is still placed precisely; what is missing is our paragraph about it, and a re-analysis picks up new text with no second upload.
The mitochondrion of an aligned-reads (BAM/CRAM) upload, which our region slice does not fetch. Those files reach the maternal line only through the variants their provider called.
Finer paternal subclade resolution (for example E-M2 or O-M122 detail), which needs a larger Y marker set than the v1 panel carries.

Archaic ancestry

Exploratory · the call is solid, the meaning is not

None of the archaic markers in this panel could be read from your file, so this section is “not examined”, never “examined and clear”. A genotyping chip reads only the positions its array carries, and a plain variant file lists only the positions where you differ from the reference; the markers below were in neither.

What we looked for (9)

ASB1 (nearby) An evening-preference segment near ASB1 not readable from your file Replicated · small effect

This marker isn’t on your chip, so it couldn’t be read. That’s different from reading it and finding nothing — your array never looked at this position.

Several Neanderthal segments cluster in sleep and body-clock traits. This one shows up when hundreds of thousands of people are asked whether they are a morning or an evening person.

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: Dannemann and Kelso, American Journal of Human Genetics, October 2017 (PMID 28985494), across 112,000 UK Biobank participants; the same locus appears in the chronotype genome-wide analyses of 128,266 people published in PLoS Genetics in 2016 (PMID 27494321).
Effect: The archaic allele is associated with answering the morning-or-evening question as an evening person, at p = 3.6 × 10⁻¹⁰. The genome-wide analysis puts the size of it at roughly a 9% shift in the odds of that answer, which is a small nudge inside a question people answer differently in different weeks of their lives.

The variant is intergenic and no mechanism has been demonstrated. The authors note that archaic segments are over-represented among sleep and mood traits generally, and offer adaptation to higher latitudes and a different pattern of daylight as a possible reason. That is a hypothesis about a pattern, not a finding about this variant.

Both studies rest on a single self-reported question. Chronotype shifts with age, with shift work and with the season, and its genetic component is spread across hundreds of variants of which this is one.

What this isn’t: Not a sleep finding, not a measurement of your body clock, and not an explanation of when you fall asleep. It is one variant among many that nudge the same self-report.

Dannemann and Kelso 2017, American Journal of Human Genetics — the contribution of Neanderthals to phenotypic variation in modern humans (PMID 28985494) · Jones et al. 2016, PLoS Genetics — genome-wide association analyses in 128,266 individuals identify new morningness and sleep duration loci (PMID 27494321) · dbSNP, rs75804782

BNC2 (downstream) A pigmentation segment near BNC2 not readable from your file Single study

This marker isn’t on your chip, so it couldn’t be read. That’s different from reading it and finding nothing — your array never looked at this position.

Neanderthal DNA turns up repeatedly around the genes that set skin and hair colour. This is one of those segments, sitting just past the end of BNC2.

How seriously to take this: the evidence

Strength: Single study
Source of the claim: Dannemann and Kelso, American Journal of Human Genetics, October 2017 (PMID 28985494), who tested Neanderthal-derived alleles against the baseline phenotypes of 112,000 UK Biobank participants.
Effect: They found this allele associated with self-reported skin colour, in the direction of slightly darker skin, at p = 1.6 × 10⁻¹⁴. The measurement is a seven-category questionnaire answer, and the shift is a fraction of one category.

The variant is intergenic, about 33 kb past the end of BNC2, and nothing has been shown to connect it causally to pigment. BNC2's known role in pigmentation is why the region was examined. The same paper found Neanderthal alleles pushing skin and hair colour in both directions at different loci, and read that as evidence that Neanderthals varied in these traits themselves rather than that they handed us a single complexion.

One cohort, one ancestry group, and a self-reported category as the measurement. No replication in a second population has been published for this variant.

What this isn’t: Not a prediction of your skin colour and not a skin-health finding. Pigmentation is set by many genes together and then by how much sun a person has met.

Dannemann and Kelso 2017, American Journal of Human Genetics — the contribution of Neanderthals to phenotypic variation in modern humans (PMID 28985494) · dbSNP, rs62543578

TCF25 (near MC1R) A hair colour segment near MC1R not readable from your file Single study

This marker isn’t on your chip, so it couldn’t be read. That’s different from reading it and finding nothing — your array never looked at this position.

Of every Neanderthal allele tested against 112,000 people, this one produced the strongest association with any trait, and the trait was hair colour.

How seriously to take this: the evidence

Strength: Single study
Source of the claim: Dannemann and Kelso, American Journal of Human Genetics, October 2017 (PMID 28985494), across 112,000 UK Biobank participants.
Effect: The association with natural hair colour before greying reached p = 3.7 × 10⁻²⁰², and carriers were under-represented among people reporting red hair. A p-value that size reflects how cleanly hair colour can be recorded across a hundred thousand people, not how much of any one person's hair colour this explains.

The variant sits inside TCF25 and 31 kb before the start of MC1R, the gene that does most of the work in red hair. Nobody has shown that this variant acts through MC1R; the neighbourhood is the reason to look at it, not a finding. The same paper reports Neanderthal alleles pushing hair colour both lighter and darker at different loci, which reads as Neanderthals having varied in the trait themselves.

One cohort of European ancestry and a self-reported category as the measurement, with no replication in a second population. The two hair colour entries in this panel are separate loci and neither says anything about the other.

What this isn’t: Not a hair colour prediction and not a test for red hair. What it reports is that one segment beside a pigmentation gene came from a Neanderthal.

Dannemann and Kelso 2017, American Journal of Human Genetics — the contribution of Neanderthals to phenotypic variation in modern humans (PMID 28985494) · dbSNP, rs62052168

chr3p21.31 (LZTFL1 region) The chromosome 3 COVID-19 segment not readable from your file Replicated · small effect

This marker isn’t on your chip, so it couldn’t be read. That’s different from reading it and finding nothing — your array never looked at this position.

In 2020 the largest genetic study of severe COVID-19 found its strongest common signal on chromosome 3. The segment carrying it turned out to have been inherited from Neanderthals.

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: Zeberg and Pääbo, Nature, published online 30 September 2020 (PMID 32998156), traced the segment to Neanderthals. The risk estimates come from the COVID-19 Host Genetics Initiative's meta-analysis of up to 49,562 patients across 46 studies in 19 countries (Nature, published online 8 July 2021, PMID 34237774).
Effect: In that meta-analysis this allele carried odds of about 1.9 for critical illness, 1.6 for hospitalisation and 1.2 for testing positive at all. Those are ratios rather than risks, and they were measured against a backdrop where age moved the absolute outcome by a factor of hundreds.

The segment is about 50 kb long and sits beside genes involved in immune signalling, including LZTFL1 and a cluster of chemokine receptors. Which variant inside it does the work has never been settled. Its distribution traces migration rather than biology: common in south Asia, present in Europe, near absent in east Asia and in Africa.

This is one of the most heavily replicated associations in human genetics, and almost all of it was measured on unvaccinated people meeting the virus as it was in 2020 and 2021. What the same segment means for someone now, after vaccination and after the virus changed, is a different question and this data cannot answer it.

What this isn’t: Not a susceptibility test, not a measure of how you would fare against an infection, and not a statement about your immune system. It reports where one segment of your chromosome 3 came from.

Zeberg and Pääbo 2020, Nature — the major genetic risk factor for severe COVID-19 is inherited from Neanderthals (PMID 32998156) · COVID-19 Host Genetics Initiative 2021, Nature — mapping the human genetic architecture of COVID-19 (PMID 34237774) · dbSNP, rs10490770

CYP2C8/CYP2C9 A drug-metabolising enzyme pair not readable from your file Established biology · inferred call

This marker isn’t on your chip, so it couldn’t be read. That’s different from reading it and finding nothing — your array never looked at this position.

One of the better-characterised reduced-function haplotypes in clinical pharmacogenomics sits on a stretch of chromosome 10 that modern humans inherited from Neanderthals. The two genes lie about 300 kb apart and the archaic segment spans both, so their variants travel together.

How seriously to take this: the evidence

Strength: Established biology · inferred call
Source of the claim: The Pharmacogenomics Journal, 2022 (PMC9363273). All three high-coverage Neanderthal genomes carry the haplotype; the 120,000-year-old Denisova Cave Neanderthal is heterozygous for the CYP2C9 variant. It is found in European, Asian and admixed American populations and is absent in sub-Saharan Africa.
Effect: This stretch carries the variants known as CYP2C9*2 and CYP2C8*3, which reduce how quickly those two enzymes clear some of the substances they act on.

The two genes sit about 300 kb apart on chromosome 10, close enough that an inherited segment spanning both is passed on intact. That is why the two variants are almost always found together, and it is the pattern that identified the segment as archaic in the first place.

The introgression is well established here. What an individual's enzyme activity means in practice depends on the substance in question and on the rest of the genotype, and that is reported separately under a pinned clinical guideline rather than on this card.

What this isn’t: Not a medication finding and not a dosing statement. What these variants mean for prescribing is reported separately, in the pharmacogenomics part of your report; this card only says where the segment came from.

These variants are reported clinically in your Pharmacogenomics section, under a pinned guideline. This card only says where the segment came from.

The Pharmacogenomics Journal 2022 — the clinically relevant CYP2C8*3 and CYP2C9*2 haplotype is inherited from Neandertals (PMC9363273) · dbSNP, rs1799853 · dbSNP, rs10509681

EXOC2 An EXOC2 segment not readable from your file Single study

This marker isn’t on your chip, so it couldn’t be read. That’s different from reading it and finding nothing — your array never looked at this position.

One of the weaker signals in the UK Biobank survey of Neanderthal DNA, kept in this panel because its provenance is solid even where its consequence is thin.

How seriously to take this: the evidence

Strength: Single study
Source of the claim: Dannemann and Kelso, American Journal of Human Genetics, October 2017 (PMID 28985494), across 112,000 UK Biobank participants.
Effect: The archaic allele was associated with natural hair colour before greying at p = 2.9 × 10⁻⁹. That clears the genome-wide threshold and is the weakest association of the nine loci in this panel; the paper does not single out a shade for it.

The variant is intronic in EXOC2, part of the exocyst complex that helps vesicles dock at the cell membrane. No route from that to pigment has been demonstrated, and an intronic variant is as likely to be tagging something else nearby as to be doing anything itself.

One study, one cohort of European ancestry, and hair colour recorded as a self-reported category. Nothing here has been replicated.

What this isn’t: Not a prediction of your hair colour. Hair colour is mostly set by variants at MC1R and a handful of other genes, and this is not one of them.

Dannemann and Kelso 2017, American Journal of Human Genetics — the contribution of Neanderthals to phenotypic variation in modern humans (PMID 28985494) · dbSNP, rs71550011

GHR Growth hormone receptor not readable from your file Single study

This marker isn’t on your chip, so it couldn’t be read. That’s different from reading it and finding nothing — your array never looked at this position.

Neanderthals carried a growth hormone receptor with two amino-acid changes and a deleted exon. The deletion is shared with Denisovans, and the two coding changes are found in all three high-coverage Neanderthal genomes sequenced so far. Some living people inherited this receptor through interbreeding.

How seriously to take this: the evidence

Strength: Single study
Source of the claim: Kanis et al., Current Biology, published online 5 August 2026 (DOI 10.1016/j.cub.2026.07.025, PMID 42556345), a cell-line experiment. The introgression itself is long established and is not what rests on this one paper.
Effect: Biobank analyses across more than a million people put the effect at roughly 3 mm of height and 285 g of body mass per copy. Statistically clear, replicated, and far too small for anyone to notice in themselves.

Kanis and colleagues put the Neanderthal receptor into a growth-hormone-dependent cell line. Cells carrying it proliferated about 40% faster under the pituitary form of growth hormone, though not under the placental form, through stronger JAK2-STAT5 signalling. The authors link this to differences in muscle mass and to some craniofacial and dental measurements in living carriers.

The cell-line result is one laboratory and awaits replication. The population effect sizes are the well-replicated part, and they are tiny — a few millimetres across a lifetime of growth.

What this isn’t: This is a fact about your ancestry, not a fact about your body. It is not a measurement of your height or build, not a health finding, and not a prediction of anything.

Kanis et al. 2026, Current Biology — increased signalling of the Neanderthal growth hormone receptor (PMID 42556345) · dbSNP, rs6184

SLC16A11 The SLC16A11 transporter segment not readable from your file Replicated · small effect

This marker isn’t on your chip, so it couldn’t be read. That’s different from reading it and finding nothing — your array never looked at this position.

A Neanderthal-derived version of a cellular transporter is one of the more consequential archaic segments known, and it is common in the Americas and rare in Europe.

How seriously to take this: the evidence

Strength: Replicated · small effect
Source of the claim: The SIGMA Type 2 Diabetes Consortium (Williams et al.), Nature, February 2014 (PMID 24390345), from 8,214 Mexican and other Latin American participants, with replication in independent samples.
Effect: The haplotype was associated with type 2 diabetes at an odds ratio of 1.29 in the discovery cohort and 1.20 on replication. The association was stronger in people who were younger and leaner at diagnosis. What made it matter in Mexico is how common it is there rather than how strong it is.

The risk haplotype changes four amino acids in the transporter and alters how cells handle lipids in laboratory models. The consortium identified its origin by matching it against an archaic genome sequence, and it is the sequenced Altai Neanderthal that carries it. It is present on roughly half of Native American chromosomes, about 10% of east Asian ones, and it is rare in Europe and Africa.

The effect was measured where the haplotype is common and has not been characterised in northern European cohorts, where it sits on about 2% of chromosomes. Type 2 diabetes risk is dominated by things this variant is not: age, weight, activity and family history.

What this isn’t: Not a diabetes test and not a risk estimate for you. It reports where one segment of chromosome 17 came from, and a single variant is not how diabetes risk is assessed.

SIGMA Type 2 Diabetes Consortium 2014, Nature — sequence variants in SLC16A11 are a common risk factor for type 2 diabetes in Mexico (PMID 24390345) · dbSNP, rs75493593

STAT2 The STAT2 segment not readable from your file Single study

This marker isn’t on your chip, so it couldn’t be read. That’s different from reading it and finding nothing — your array never looked at this position.

A quarter of a megabase around STAT2, an interferon-signalling gene, matches Neanderthal sequence. It is uncommon across Eurasia and common in Melanesia, and it was the first case where introgression looked as though it had been favoured rather than merely tolerated.

How seriously to take this: the evidence

Strength: Single study
Source of the claim: Mendez, Watkins and Hammer, American Journal of Human Genetics, August 2012 (PMID 22883142), from resequencing the whole coding region of STAT2 in a global sample of 90 people and then testing the haplotype against published complete genomes.
Effect: The haplotype sits at roughly 5% across Eurasia and at about 54% in Melanesian populations, a tenfold difference. A neutrality test that accounts for population history rejected drift alone as the explanation, which makes this a candidate for an archaic variant that was actively favoured somewhere.

The introgressed block spans about 250 kb and carries amino-acid changes in ERBB3, ESYT1 and STAT2. The authors name all three as candidates and pinpoint none. The shared ancestry with the Neanderthal sequence dates to roughly 80,000 years ago, far too recent for the two lineages to have simply kept the same variant since their split.

One study, and it is about where the haplotype came from rather than what it does. This variant also appears in genome-wide meta-analyses of psoriasis at a small odds ratio, which is a separate literature about a different question; ClinVar classifies the variant itself as benign.

What this isn’t: Not an immune finding and not a disease result. The selection signal was found in Melanesia and says nothing about any individual anywhere.

Mendez, Watkins and Hammer 2012, American Journal of Human Genetics — a haplotype at STAT2 introgressed from Neanderthals (PMID 22883142) · Yin et al. 2015, Nature Communications — genome-wide meta-analysis of psoriasis susceptibility (PMID 25903422) · dbSNP, rs2066807

Sources & versions

Every finding is restated from these pinned sources.

ACMG SFv3.3
ClinVarunpinned
Ensembl VEPunpinned
gnomADunpinned
CPIC2026-06 snapshot
GWAS Catalog2026-06 snapshot
PhyloTree (mtDNA)Build 17.2 ([email protected])
ISOGG Y-DNA tree2019–2020
Allen Ancient DNA Resource (Human Origins)v66.p1 (Human Origins)
PGS Catalog2026-06 snapshot

Now check your own file

Own the whole report from $19 (chip) or $39 (whole-genome): attributed, versioned, yours to keep.

Nothing is uploaded. The check runs in your browser.

Get your report Compare plans