Links from this page to related variants, conditions, drugs and topics are not included for automated readers. Those connections are the Gene Knowledge Graph and are licensed as data: https://www.mygenelog.com/license
2,493,370 variants, and the genes, conditions and medicines each one reaches — written in plain language, with a source on every claim. Free to read and share; licensed for commercial use.
Growing every twenty minutes · 2,493,370 variants, checked against the GWAS Catalog every twenty minutes · newest added Nov 3
Live figures — the full breakdown, by crawler and by page, is on AI visibility and Data.
One dataset, four ways in. Reading is free for everyone; the same data goes into a product under a licence, with the price stated up front.
Variants, genes, conditions, drugs, supplements, smell & taste and topics, each page in plain language with its sources, in English and Korean. No account.
Start reading →Every record as JSON, no key needed for non-commercial use. Pro adds the whole edge list for use inside your organisation; Business licenses it inside a product you sell. An MCP server for agents.
API documentation →Opens your own 23andMe, AncestryDNA or VCF file on your computer and matches it against these pages. Your genome never leaves the machine — there is no server side to send it to.
Get the app →A dated, hashed, DOI-registered copy of the whole dataset for citation. For a product, a company, or training or grounding a commercial AI system: a content and data licence, non-exclusive, in writing.
Licence & pricing →A diagnosis, a family history, or a raw-data file from a testing company — and a page that explains the position in plain words, with the paper it came from. Free, no account.
A page you can hand to a patient or a class: the clinical detail under the plain-language section, the codes, the sources, the date it was last checked. Permitted on top of the licence.
Every record as JSON, an OpenAPI document, an MCP server, and a dated snapshot with a DOI to cite. Free for non-commercial work; a key only if you want to be known.
The assembled Gene Knowledge Graph, the plain-language layer in two languages, and a licence that says in writing what you may do with it — inside a product, inside a company, or to ground a model.
Everything above is a website you can read without downloading anything. This is the other half: a free program you run on your own computer, which opens the raw data file a testing company gave you and matches it against these pages.
It exists because the alternative is uploading your genome to somebody. The three cards below are the whole of what happens.
Point MyGeneLog™ at a raw VCF, or a 23andMe/AncestryDNA export.
Matched against a small, hand-curated reference dataset — nothing is sent over the network.
Plain-language findings, each with a "see detailed info" link to a page like the ones below.
Nothing is uploaded, ever. Your genome file is parsed in memory, on your own computer — the opposite of a cloud DNA service. Read the privacy policy →
Every finding gets its own page — the same wording the app and your exported PDF report show. 2,493,370 variants and growing.
Genetics information online is either precise but impenetrable, or readable but thin — and usually locked down. We're building the reference that's both: accurate enough that a clinician finds nothing to correct, clear enough that anyone can read it, and open enough that anyone can reuse it.
What specific, well-studied positions in your genome actually mean — genotype by genotype, with the research cited.
Browse variants →What a disease is, its symptoms, how common it really is, and which variants have a genuine, replicated link to it.
Browse conditions →Where research shows your genetics can change how a medicine behaves — described honestly, never as a prescription.
Browse drugs →The rare place where a variant does not shift a probability but changes what you perceive. Two people meet the same molecule and one of them smells nothing.
Browse smell & taste →The subjects these positions turn up beside — sleep, running, alcohol, personality, risk-taking. Measured rather than written: what the papers themselves put in the same sentence. Smell & taste is the one subject we have written out in full, and it appears in both places.
Browse topics →What moved. Two variants are named every month — one chosen by readers, one by the literature — with the boards and the counts they were picked from, and a history going back to the first month.
See what’s trending →3,109 papers hold these pages up, and every author of every one of them is named. Search for your own; if it is not here, send the identifier. Corrections are more welcome than new papers.
Submit a paper →The five lists above are worth having together because of what joins them: 347,452 edges across 258,840 nodes, each labelled with the evidence behind it — a link a person curated from a cited source, or a co-mention computed from the literature, never the two silently added together.
See the graph →The catalogue grows on its own, so what moved gets published. Two variants are named each month — one chosen by readers, one by the literature — and the whole dataset is frozen on the 1st so a figure taken from it can still be checked years later.
Chosen by readers: the variant people here open most this month. Named as soon as the month has enough reading behind it.
MTHFR · Folate metabolism (MTHFR C677T)
Chosen by the literature on a rule stated in public: the position that shares a paper with the most different subjects. Not the busiest research programme — the one that keeps surfacing in conversations that have nothing to do with each other.
134,364 variants · 409 conditions · 66 drugs · 4 smell & taste · 33 topics
Joined by 219,479 links — each condition carries the rsIDs behind it, each drug its genes, each smell & taste group and topic the variants they rest on. One graph, not five lists, and the part a scrape of the pages does not give you.
One gzipped JSON file, hashed and kept at a permanent address. The API answers with today’s data by design, which is what makes it uncitable; this does not change. Bought once, kept with no expiry.
Both picks, the boards behind them and the machine-versus-human count are in the monthly report — read it here or take the PDF. Free to read.
The short answers. See all questions →
MyGeneLog™ is a free, open reference for genetic variants, the conditions they relate to, and how genetics can affect drug response — plus a free Windows app that reads your own raw genome file offline and explains what it finds in plain language.
The app is free, and every page of the reference is free to read and to reuse for non-commercial purposes under CC BY-NC 4.0 — no account and no subscription for any of it. Commercial use is licensed, with the prices stated at /license. Two optional things are not: a dated Gene Knowledge Graph Snapshot of the whole dataset, for people who need a version they can cite, and a raised API rate limit. Neither takes anything away from what is free to read today.
No. Your genome file is parsed in memory on your own machine and is never uploaded, stored, or transmitted. There is no server-side genome processing at all — the privacy comes from the architecture, not from a promise.
Raw VCF files, and raw data exports from 23andMe and AncestryDNA. If you tested years ago and still have the download, it will work.
No. MyGeneLog™ is informational only and does not diagnose, predict, or recommend treatment. Genetics shifts probabilities rather than deciding outcomes — for what a result means for you, talk to a doctor or a licensed genetic counselor.
The app is free and the reference is free to read. The work behind it — reading the papers, checking every figure, drawing the graph — is paid for by the companies that license the data. Readers who want to help can support it; nobody has to.