Raine Laurila
Founder, Aimosti Oy. Oulu, Finland.
I studied information processing science at the University of Oulu, and I have spent my career with computers and information technology solving problems and making data useful to businesses. Aimosti is the same craft pointed at a harder problem.
I founded Aimosti in 2026 and built the pipeline that reads your file, the site you are reading now, and every gene, condition and drug page on it. I am not a doctor, a geneticist or a clinical laboratory scientist, which is exactly why nothing here rests on my authority. Every result and finding is grounded in actual, cited, peer-reviewed research. Aimosti is not a diagnostic test and does not provide medical advice; what it does is take your genome and cross-reference it against the latest knowledge.
Why I built this
Aimosti began in January 2025, when I bought myself a whole-genome sequencing kit. I had reached the age where I wanted a closer look at my own health. I already knew my lipid levels were less than optimal, and that lifestyle changes alone were not moving them much. I knew that some of the factors behind that are encoded in DNA, and thought I might be able to extract some of it myself. I believe the old adage that information is power, and I would rather know, even when the news is bad. I'm a proponent of informed choice and early prevention, which is exactly what having your genome sequenced and analysed gives you.
The results took about six months to arrive, by which point I had half stopped believing they ever would. In hindsight, the wait was useful. Fretting over whether I had lost both my money and my sample gave me months to think about what I had actually agreed to, and in particular about who owns my data.
I am still not sure whether I permanently lost control of my genome the day I signed up, or whether it has simply been drifting from one corporate owner to the next ever since, in a jurisdiction I do not live in, with no say from me, indefinitely. The policies were opaque enough that I could not tell, and I have wondered since whether that was the point. Then 23andMe went bankrupt, and a great many people discovered that the question of what happens to their genome would be settled by a bankruptcy court rather than by them.
What privacy-first actually means here
So Aimosti is privacy-first, and everything else is built around that. It's concrete rather than a slogan (it is in the product's DNA, if you like), because a slogan is exactly what I would have distrusted.
Aligned reads are deleted the moment the coverage computation finishes, on success or on failure. Your raw variant file goes within about 7 days. What is kept is your report. If you take the re-analysis subscription, a small panel-restricted variant file is kept alongside it so the report can be re-run as sources update.
Delete your account and the genome data, the reports and the uploaded files go with it. What remains is a de-identified record of your orders and consents, plus the accounting data Finnish law requires me to keep for about six years. That is the whole list, and the privacy policy states it in full and binds me to it. What I can state unequivocally is that no genome data, and nothing derived from your genome (this is the important part), is kept once you delete your account.
Why a data person reads genome files
Most of what decides whether an answer about your genome can be trusted is a data question before it is a medical one. Which reference build was the file written against? The contig names match across GRCh37 and GRCh38, so reading a file against the wrong build does not produce an error, it produces numbers at the wrong coordinates, which is considerably worse. Was a position actually examined, or is it merely absent from the file? In a gVCF the two look identical unless you check, and reading absence as a reference call turns a missing measurement into a clean result. How deep is the coverage at the exact spot where the claim is being made? And is this a short-variant file, which cannot see a copy-number change no matter how carefully it is read?
Those are the questions I ask, and I ask them before anything else. The common failure in consumer genetics is not misreading a variant. It is answering a question the file could never have supported and presenting that answer with the same confidence as the ones it could.
I do not diagnose, I do not interpret individual results, and I do not tell anyone what to do about their medication. Where clinical guidance exists, Aimosti restates it, names the guideline, and tells you to take it to your doctor or prescriber.
What I built
Aimosti re-reads a genome file you already own against current science. The parts I wrote:
The file check that runs in your browser. It reads the header locally, identifies the format and the reference build, and tells you which report the file can support. Nothing uploads, and you don't even need an account.
Build handling. A GRCh37 file is lifted to GRCh38 before analysis, under a guard that discards anything that does not survive the lift cleanly rather than letting it through at the wrong coordinates.
The 29-target pharmacogenomic panel, called against a dated CPIC snapshot, so the guidance a report quotes is pinned to a version, not to whatever the internet happened to say on the day it ran.
Deep Read for aligned files: CYP2D6 copy number, which no variant file can see; two recurrent deletions called from read depth; typed HLA-DQ; and a per-gene measurement of how well each gene was actually covered.
The refusals. A gene the file did not cover well enough is reported as not measured, never estimated. Modules that measurement showed could not be delivered with confidence from consumer data were cut, not shipped.
There is also excellent Finnish research to build on, and it would have been rude not to. Finns are a genetic isolate, which means a global product reads a Finn's genome against an allele-frequency baseline that is wrong for them. The Finnish Disease Heritage (a group of rare recessive conditions enriched in this population) is part of the carrier panel for exactly that reason. A re-analysis service is never finished, and at this stage there is a new piece, feature or fix to add most weeks. I intend to keep it that way.
The anti-hype train
The other thing this field has in abundance is hype. So findings here are tiered by how seriously they deserve to be taken, and the weaker material is shown and labelled weak, not quietly dropped, nor quietly promoted. The hyped and 'now trending' findings may be of interest to some, so they are included, but labelled explicitly as emerging.
No science originates here. Every claim on a gene, condition or drug page restates a published source: a CPIC guideline, GeneReviews, ClinVar, NCBI Gene, MedlinePlus, the FDA pharmacogenomic associations table. The source is named and linked where the claim is made, so you can check the reference yourself and keep digging further.
A claim I cannot source does not ship. Where the source is silent, the page says so, instead of filling the gap with confident adjectives.
Every page carries the date it was last reviewed, and I read each one end to end before it goes live. If you find something here that is wrong, write to me at [email protected] and I will correct it and re-date the page.
Closing
My background is in computer science rather than biology, and from where I sit this is largely a problem of organising information: connecting genes to something that means something to a person. Science is a global collaborative effort, and all of us are standing on the shoulders of giants, some of whom, thankfully, write excellent documentation. That is the only reason I could have started from curiosity about my own health and family history and end up running a service like this.