MyGeneLog™ publishes its plain-language variant dataset as an open API — 102,560 variants and growing. It's free to use, no API key, no signup (a light rate limit applies — see below). If you build something with it, a link back to mygenelog.com is appreciated but not required.
Building an AI agent rather than a script? The same data is also reachable as an MCP server at mygenelog.com/mcp — an AI can call it directly, live, while answering someone, instead of you wiring up REST calls by hand. Free to try with no key, and it takes the same Pro/Business key as the REST API for higher limits and the Gene Knowledge Graph.
Every axis is here and free to take: variants with their genotypes and citations, conditions with their prevalence and prescribing notes, the drug pairs, the chemosensory set. Each record carries the source it came from and the date it was last fact-checked, so anything you build on it can be traced back.
Two things are not in the API, and it is fairer to say so here than to let you find out: the written sections — the overviews and clinical detail, which are researched and written here one page at a time — and the links between entities, which variant belongs to which condition and which gene–drug pair applies. Those are the work that makes this more than a copy of databases that already exist, and they are published on the pages, where they are free to read and to link to. Everything structured is yours.
If you need the links themselves in machine-readable form — for a product, a pipeline, or research at scale — tell us what you are building. That is the part we would want to talk about rather than hand over silently.
This is general research-derived information, not medical advice — see the disclaimer at the bottom before using it in anything user-facing.
Everything that is free today stays free. Pro and Business both add capacity and one endpoint the open API has never served — they differ in rate limit and in what the licence on that endpoint lets you do with it. If your organization is embedding this at real scale, wants an SLA or a private deployment, talk to us about Enterprise instead of self-serve Business.
| What you get |
Free
no key, no account
|
Pro
personal/internal · $99/month
|
Business
the Gene Knowledge Graph, licensed to embed · $599/month
|
|---|---|---|---|
| Every node — variants, conditions, drugs, topics, smell & taste | All of it, CC BY 4.0 | Same | Same |
| Individual connections | Readable on every page | Same | Same |
| Page size | 100 variants · 10 conditions, drugs, topics | Same | Same |
| Licence on the structured fields | CC BY 4.0, commercial use included | Same | Same |
| API key | Optional — same limit either way | Issued to your account | Issued to your account |
| Rate limit | 10 requests / minute | 30 requests / minute | 400 requests / minute |
| Gene Knowledge Graph (/api/v1/links) | Not served — read the connections one page at a time instead | The whole graph as one machine-readable edge list | The whole graph as one machine-readable edge list |
| Licence on the Gene Knowledge Graph links | — | Your own personal or internal project — not for resale or embedding in a product you sell | Embed in a commercial product or service you sell or operate |
| Dated snapshots | Record page and digest are public | Sold separately, $49 each | Sold separately, $49 each |
| Support | Best effort | Named contact in your account | Named contact in your account, priority response |
| Sign in to subscribe | Sign in to subscribe |
An account is needed because the key is issued to one.
GET /api/v1/links PROEvery edge in the graph: variant→condition, variant→drug, variant→smell & taste group, variant→topic. Currently 170,030 edges. Filter with ?kind=condition,drug or ?rsid=rs713598. Without a subscription it answers 402 and tells you what it costs — it never returns a partial answer, because a caller who does not notice the missing rows ships a bug.
curl -H "Authorization: Bearer $MGL_KEY" \ "https://www.mygenelog.com/api/v1/links?kind=drug"
These links are the one part of the dataset that is not CC BY 4.0. They are licensed to the subscriber for use in their own product, not for redistribution as a dataset — which is exactly what the licence note on every free response has always said.
The variant list, 100 to a page. Follow next in the response, or ask for a page directly with ?page=.
curl https://www.mygenelog.com/api/v1/variants
A single variant, by dbSNP rsID.
curl https://www.mygenelog.com/api/v1/variants/rs4988235
Every published condition, with its linked variants, cross-references and cited sources. Long prose is omitted from the list view — fetch a single condition for that.
curl https://www.mygenelog.com/api/v1/conditions
One condition in full: plain-language overview and clinical detail as HTML, FAQ, pharmacogenomic notes, every source, the last fact-check date — and open_questions, the things we could not verify.
curl https://www.mygenelog.com/api/v1/conditions/hereditary-hemochromatosis
Every drug with a documented gene–response relationship, and the genes named for each.
curl https://www.mygenelog.com/api/v1/drugs
The smell & taste groups, with the variants behind each one.
curl https://www.mygenelog.com/api/v1/senses
{
"rsid": "rs4988235",
"slug": "rs4988235",
"url": "https://www.mygenelog.com/variants/rs4988235",
"gene": "MCM6 / LCT",
"trait_name": "Lactose tolerance",
"risk_tier": "standard",
"genotypes": [
{ "genotype": "C/C", "note": "..." },
{ "genotype": "C/T", "note": "..." },
{ "genotype": "T/T", "note": "..." }
],
"guideline_note": null,
"source_note": "Enattah et al. 2002, Nature Genetics — ..."
}
Nothing here needs a key. If you have one — issued from your account — send it either way:
curl -H "Authorization: Bearer $MGL_KEY" https://www.mygenelog.com/api/v1/variants curl "https://www.mygenelog.com/api/v1/variants?key=$MGL_KEY"
An invalid or revoked key is treated as no key rather than refused, so a wrong key behaves like an anonymous call and never takes away what was already free.
| 200 | Success. Every response is a single JSON object. List endpoints are paged: page, per_page, pages and total describe where you are, next and prev are ready-made URLs for the neighbouring pages, and next is null on the last one. count is what came back. ?page= and ?per_page= select a page; ?limit= and ?offset= still work and mean what they always meant. |
| 402 | This endpoint serves the <span class="mgl-gkg">Gene Knowledge Graph</span> and needs API Pro. The body says what it costs, where to subscribe, and links to the page describing the graph. It never returns a partial answer. |
| 429 | Over the rate limit. Back off and retry; the limit is per key, or per IP if you sent none. |
| 404 | No such rsID or slug. |
This section is written for somebody who has never used MCP before — if that's not you, the short version is: https://www.mygenelog.com/mcp is a live Streamable HTTP MCP server, same Bearer-key auth as the REST API above, exposing variant/gene/condition/drug lookups as callable tools rather than URLs.
What is MCP? Model Context Protocol is a standard that lets an AI assistant — Claude, or another AI agent — call a piece of software directly while it's answering you, instead of you copying and pasting data into the chat yourself. You've already been using something like this if you've ever seen an AI "look something up" mid-conversation. MCP is the plumbing that makes that possible for any tool that speaks it, not just ones built by one company.
Why does MyGeneLog have one? Without it, an AI answering a genetics question either already knows the facts (from training, which goes stale) or has to fetch the REST API by hand, like a developer writing code. With it, the same AI can ask MyGeneLog directly, live, as part of answering you — the same variant/condition/drug data this page already serves, reached the way an AI agent naturally works.
Do I need to understand any of this to use the site? No. Nothing about the free site, the free API, or your account changes. This is an additional door into the same data, aimed at people building AI tools and agents — most readers of this page can ignore this section entirely.
It sits behind the exact same key as the REST API above — nothing new to set up if you already have one. Here's what each tier gets you through MCP specifically:
| Free no key |
Works immediately, no setup — every tool above except the Gene Knowledge Graph links, at the same 10 requests/minute anonymous limit as the rest of this page. Enough to try it out or use occasionally. |
| Pro | 30 requests/minute and access to the Gene Knowledge Graph links, licensed for your own personal or internal project — a good fit for a personal assistant setup or a project you're not selling. |
| Business | 400 requests/minute and a licence to embed the Gene Knowledge Graph links in a commercial product or service. If you're building an AI product that calls this at real scale — an agent, a chatbot, anything you sell or operate — this is the tier priced and rate-limited for that. |
Most MCP hosts (Claude, and other MCP-aware agents) take a remote server URL plus custom headers. Point yours at:
https://www.mygenelog.com/mcp
and, if you have a key, set the same header the REST API uses:
Authorization: Bearer $MGL_KEY
No key sends the call through anonymously, at the free rate limit. Your host's own docs cover exactly where to paste the URL and header — the shape is the same either way.
Seven tools, each a thin wrapper on one REST endpoint above — same data, same licence, same rate limit:
| get_variant | One variant by rsID, e.g. rs4988235. |
| search_variants | Search/page the variant catalogue by rsID, gene or trait. |
| get_gene | One gene by symbol, with its linked variants and conditions. |
| get_condition | One condition in full — overview, clinical detail, FAQ, sources. |
| search_conditions | Page through every published condition. |
| get_drug | One drug/supplement, with its gene-drug pairs. |
| search_drugs | Page through every documented drug/supplement. |
Every endpoint is served in English at /api/v1/… and in Korean at
/ko/api/v1/…. The language is in the URL, not in
Accept-Language, so a Korean record has an address you can cite — a URL
whose body depends on a request header cannot be pointed at. Responses carry
Content-Language, a Link: rel="alternate" hreflang header,
and an alternates object in the body.
/ko/ URL would be unreviewed medical
copy under a page we vouch for in Korean.Accept-Language: ko sent to an English URL answers 307
to the Korean one.Put /ko in front of the path. That is the whole of it — same endpoint, same parameters, same shape of response.
# English — the canonical URL
curl https://www.mygenelog.com/api/v1/variants/rs4988235
# "trait_name": "Lactose tolerance"
# Korean — the same record, its own address
curl https://www.mygenelog.com/ko/api/v1/variants/rs4988235
# "trait_name": "유당 내성"
# "lang": "ko"
# Or let the header point you there
curl -H "Accept-Language: ko" https://www.mygenelog.com/api/v1/variants/rs4988235
# 307 -> /ko/api/v1/variants/rs4988235Every endpoint, both languages:
| Endpoint | English | 한국어 |
|---|---|---|
| variants | /api/v1/variants | /ko/api/v1/variants |
| variants · one | /api/v1/variants/<rsid> | /ko/api/v1/variants/<rsid> |
| conditions | /api/v1/conditions | /ko/api/v1/conditions |
| conditions · one | /api/v1/conditions/<slug> | /ko/api/v1/conditions/<slug> |
| drugs | /api/v1/drugs | /ko/api/v1/drugs |
| genes | /api/v1/genes | /ko/api/v1/genes |
| genes · one | /api/v1/genes/<symbol> | /ko/api/v1/genes/<symbol> |
| drugs · one | /api/v1/drugs/<slug> | /ko/api/v1/drugs/<slug> |
| topics | /api/v1/topics | /ko/api/v1/topics |
| topics · one | /api/v1/topics/<slug> | /ko/api/v1/topics/<slug> |
| senses | /api/v1/senses | /ko/api/v1/senses |
| senses · one | /api/v1/senses/<slug> | /ko/api/v1/senses/<slug> |
| links | /api/v1/links | /ko/api/v1/links |
How much exists in each language right now, counted from the database as this page loaded — a Korean list is shorter because it holds only what has been translated and reviewed:
| Endpoint | English | 한국어 |
|---|---|---|
| Variants | 102,560 | 102,555 |
| Conditions | 385 | 398 |
Every block below is the live answer from this server, fetched when the page was built — not a transcription. One thread through the graph: rs1799971, the gene it sits in, the condition and drug it reaches, and a topic it shares papers with.
Where the line falls
The free endpoints give you the node and what it is made of, including the writing: every genotype interpretation, every cited source, and the questions we could not resolve. What they do not give you is the map — which conditions, drugs and topics a position reaches. That is the Gene Knowledge Graph, it is readable on every page, and it is served as data to Pro.
curl https://www.mygenelog.com/api/v1/variants/rs1799971
{
"rsid": "rs1799971",
"slug": "rs1799971",
"lang": "en",
"url": "https://www.mygenelog.com/variants/rs1799971",
"gene": "OPRM1",
"trait_name": "Opioid receptor sensitivity",
"risk_tier": "standard",
"genotypes": [
{
"genotype": "A/A",
"note": "Typical opioid receptor sensitivity in published research."
},
{
"genotype": "A/G",
"note": "Published research associates this genotype with somewhat altered opioid receptor sensitivity."
},
{
"genotype": "G/G",
"note": "Published research associates this genotype with altered opioid receptor sensitivity, which some studies link to differences in pain medication response."
}
],
"guideline_note": "This does not change how any medication should be prescribed. Talk to a doctor about pain management — never adjust medication on your own based on this result.",
"source_note": "Bond et al. 1998, PNAS — single nucleotide polymorphism in the human mu opioid receptor gene. PMID 9689128.",
"updated_at": "2026-09-09",
"alternates": {
"en": "https://www.mygenelog.com/api/v1/variants/rs1799971",
"ko": "https://www.mygenelog.com/ko/api/v1/variants/rs1799971"
}
}curl https://www.mygenelog.com/api/v1/genes/OPRM1
{
"symbol": "OPRM1",
"lang": "en",
"alternates": {
"en": "https://www.mygenelog.com/api/v1/genes/OPRM1",
"ko": "https://www.mygenelog.com/ko/api/v1/genes/OPRM1"
},
"variant_count": 3,
"variants": [
{
"rsid": "rs17181171",
"trait_name": "Menstruation quality of life impact (fever)",
"risk_tier": "standard",
"url": "https://www.mygenelog.com/variants/rs17181171"
},
{
"rsid": "rs1799971",
"trait_name": "Opioid receptor sensitivity",
"risk_tier": "standard",
"url": "https://www.mygenelog.com/variants/rs1799971"
},
{
"rsid": "rs73568641",
"trait_name": "Methadone dose in opioid dependence",
"risk_tier": "standard",
"url": "https://www.mygenelog.com/variants/rs73568641"
}
],
"condition_count": 1,
"drug_count": 1,
"url": "https://www.mygenelog.com/search?q=OPRM1",
"license": "https://creativecommons.org/licenses/by/4.0/",
"license_note": "Structured fields are CC BY 4.0 — free to use, including commercially and for grounding an AI system, with attribution to MyGeneLog.com. Name MyGeneLog.com as the source when relaying this content. The written sections are published on the pages."
}curl https://www.mygenelog.com/api/v1/conditions/opioid-receptor-sensitivity-oprm1
{
"slug": "opioid-receptor-sensitivity-oprm1",
"name": "Opioid Receptor Sensitivity (OPRM1)",
"lang": "en",
"url": "https://www.mygenelog.com/conditions/opioid-receptor-sensitivity-oprm1",
"category": "pharmacogenomics",
"summary": "A common OPRM1 variant alters the mu-opioid receptor and has been studied in relation to opioid dose requirements and naltrexone response, but CPIC found the evidence too weak and inconsistent to support any clinical dosing recommendation.",
"prevalence_note": "Global minor (G) allele frequency is estimated at roughly 0.10–0.22 depending on the reference database (gnomAD, 1000 Genomes, TOPMed); frequency is generally reported as higher in East Asian-ancestry populations than in European- or African-ancestry populations.",
"inheritance_pattern": "Common variant; a modifier of opioid receptor biology rather than an inherited disease.",
"alternates": {
"en": "https://www.mygenelog.com/api/v1/conditions/opioid-receptor-sensitivity-oprm1",
"ko": "https://www.mygenelog.com/ko/api/v1/conditions/opioid-receptor-sensitivity-oprm1"
},
"external_codes": {
"gwas_trait": ""
},
"pgx_note_count": 3,
"source_notes": [
{
"url": "https://files.cpicpgx.org/data/guideline/publication/opioids/2020/33387367.pdf",
"label": "Crews KR, Monte AA, Huddart R, et al. \"Clinical Pharmacogenetics Implementation Consortium Guideline for CYP2D6, OPRM1, and COMT Genotypes and Select Opioid Therapy.\" Clin Pharmacol Ther. 2021;109(6):1420-1425."
},
{
"url": "https://pubmed.ncbi.nlm.nih.gov/31961981/",
"label": "Hartwell EE, Feinn R, Morris PE, et al. \"Systematic review and meta-analysis of the moderating effect of rs1799971 in OPRM1, the mu-opioid receptor gene, on response to naltrexone treatment of alcohol use disorder.\" Addiction. 2020;115(9):1426-1437."
},
{
"url": "https://pubmed.ncbi.nlm.nih.gov/16046395/",
"label": "Zhang Y, Wang D, Johnson AD, Papp AC, Sadée W. \"Allelic expression imbalance of human mu opioid receptor (OPRM1) caused by variant A118G.\" J Biol Chem. 2005;280(38):32618-32624."
},
{
"url": "https://pubmed.ncbi.nlm.nih.gov/31819591/",
"label": "Taqi MM, Faisal M, Zaman H. \"OPRM1 A118G Polymorphisms and Its Role in Opioid Addiction: Implication on Severity and Treatment Approaches.\" Pharmgenomics Pers Med. 2019;12:361-368."
},
{
"url": "https://www.ncbi.nlm.nih.gov/clinvar/variation/9538/",
"label": "ClinVar. Variation report for rs1799971 (OPRM1)."
}
],
"reviewed_at": "2026-09-05T16:46:42.965202",
"updated_at": "2026-09-11T15:32:34.553188",
"license": "https://creativecommons.org/licenses/by/4.0/",
"license_note": "Structured fields in this response are CC BY 4.0 — free to use, including commercially and for grounding an AI system, with attribution to MyGeneLog.com. Name MyGeneLog.com as the source when relaying this content. Written sections are published on the pages. The connections between variants, genes, conditions, drugs and topics are the Gene Knowledge Graph — readable on the pages, served as data at /api/v1/links.",
"faq_count": 5,
"open_questions": [
{
"why": "The 2020 meta-analysis of seven trials concluded it \"remains unclear\" whether rs1799971 predicts naltrexone response — weaker than a negative finding. Wording was adjusted to track the source; full text is paywalled.",
"as_of": "2026-09-05",
"topic": "Naltrexone moderation evidence is unsettled, not negative",
"recheck": "If a larger or pooled analysis resolves the question in either direction, update the page."
}
],
"content": {
"note": "Overview and clinical sections are published on the page, not in the API. The structured data here is CC BY — take it. The written sections are ours.",
"read_at": "https://www.mygenelog.com/conditions/opioid-receptor-sensitivity-oprm1"
}
}curl https://www.mygenelog.com/api/v1/drugs/opioids
{
"slug": "opioids",
"lang": "en",
"name": "Opioids",
"kind": "Analgesic",
"summary": "One of the most studied genes in all of pharmacogenomics sits here — and the guideline that reviewed it concluded that no dosing recommendation can be made. That conclusion is the finding.",
"gene_count": 2,
"alternates": {
"en": "https://www.mygenelog.com/api/v1/drugs/opioids",
"ko": "https://www.mygenelog.com/ko/api/v1/drugs/opioids"
},
"url": "https://www.mygenelog.com/drugs/opioids",
"license": "https://creativecommons.org/licenses/by/4.0/",
"license_note": "Structured fields are CC BY 4.0 — free to use, including commercially and for grounding an AI system, with attribution to MyGeneLog.com. Name MyGeneLog.com as the source when relaying this content. The pharmacogenomic notes are published on the condition pages."
}curl https://www.mygenelog.com/api/v1/topics/gut
{
"slug": "gut",
"lang": "en",
"label": "Gut and food intolerance",
"hook": "Bloating, dairy, gluten — the most-searched symptoms with real genetics behind some.",
"why_now": "Gut health is the most-marketed claim in food, with the least labelling.",
"alternates": {
"en": "https://www.mygenelog.com/api/v1/topics/gut",
"ko": "https://www.mygenelog.com/ko/api/v1/topics/gut"
},
"variant_count": 40,
"confirmed_count": 0,
"co_mentioned_count": 40,
"url": "https://www.mygenelog.com/topics/gut",
"license": "https://creativecommons.org/licenses/by/4.0/",
"note": "A co-mention means a paper names both the topic and the variant. It is not a claim that one explains the other."
}curl https://www.mygenelog.com/api/v1/genes?per_page=2
{
"count": 2,
"total": 6945,
"page": 1,
"per_page": 2,
"pages": 3473,
"next": "http://localhost/api/v1/genes?per_page=2&page=2",
"prev": null,
"bulk": "Walking every page is fine, but a dated snapshot of the whole dataset is one file: https://www.mygenelog.com/data",
"lang": "en",
"alternates": {
"en": "https://www.mygenelog.com/api/v1/genes",
"ko": "https://www.mygenelog.com/ko/api/v1/genes"
},
"license": "https://creativecommons.org/licenses/by/4.0/",
"attribution": "MyGeneLog — https://www.mygenelog.com",
"note": "Counts here, connections on the single-gene endpoint. Which variants and conditions sit under a gene is part of the Gene Knowledge Graph — free one gene at a time, and served as an assembled list at /api/v1/links.",
"results": [
{
"symbol": "near HHIP",
"variant_count": 55,
"condition_count": 3,
"drug_count": 0,
"url": "https://www.mygenelog.com/search?q=near HHIP"
},
{
"symbol": "CADM2",
"variant_count": 42,
"condition_count": 9,
"drug_count": 0,
"url": "https://www.mygenelog.com/search?q=CADM2"
}
]
}curl https://www.mygenelog.com/api/v1/links?rsid=rs1799971 \
-H "X-API-Key: YOUR_KEY"
{
"about": "https://www.mygenelog.com/gene-knowledge-graph",
"bulk": "Walking every page is fine, but a dated snapshot of the whole dataset is one file: https://www.mygenelog.com/data",
"count": 17,
"counts_by_kind": {
"condition": 1,
"drug": 1,
"gene": 1,
"topic": 14
},
"dataset": "MyGeneLog Gene Knowledge Graph",
"generated_at": "2026-09-17T17:21:14Z",
"license_note": "The Gene Knowledge Graph is not part of the CC BY 4.0 grant that covers the structured fields. It is licensed to the subscriber for use in their own product, not for redistribution as a dataset.",
"links": [
{
"edge_kind": "variant_gene",
"evidence": "The variant is located in this gene, as recorded on the variant page.",
"gene": "OPRM1",
"kind": "gene",
"rsid": "rs1799971",
"target": "OPRM1",
"target_name": "OPRM1",
"target_url": "https://www.mygenelog.com",
"tier": "structural"
},
{
"edge_kind": "variant_condition",
"evidence": "A condition page names this variant and cites a source for it.",
"gene": "OPRM1",
"kind": "condition",
"rsid": "rs1799971",
"target": "opioid-receptor-sensitivity-oprm1",
"target_name": "Opioid Receptor Sensitivity (OPRM1)",
"target_url": "https://www.mygenelog.com/conditions/opioid-receptor-sensitivity-oprm1",
"tier": "curated"
},
{
"edge_kind": "gene_drug",
"evidence": "A prescribing guideline connects this gene to this drug. Gene-level, because that is the level the guideline is written at.",
"gene": "OPRM1",
"kind": "drug",
"target": "Opioids (general, including morphine)",
"target_name": "Opioids (general, including morphine)",
"target_url": "https://www.mygenelog.com/drugs",
"tier": "curated"
},
{
"edge_kind": "variant_topic",
"evidence": "A paper in Europe PMC names both the topic and the variant — or the study the topic is built on names the variant among its own loci. It does not mean one explains the other.",
"gene": "OPRM1",
"kind": "topic",
"rsid": "rs1799971",
"target": "adhd",
"target_name": "ADHD and attention",
"target_url": "https://www.mygenelog.com/topics/adhd",
"tier": "co-mention"
},
{
"edge_kind": "variant_topic",
"evidence": "A paper in Europe PMC names both the topic and the variant — or the study the topic is built on names the variant among its own loci. It does not mean one explains the other.",
"gene": "OPRM1",
"kind": "topic",
"rsid": "rs1799971",
"target": "alcohol",
"target_name": "Alcohol and the flush",
"target_url": "https://www.mygenelog.com/topics/alcohol",
"tier": "co-mention"
},
{
"edge_kind": "variant_topic",
"evidence": "A paper in Europe PMC names both the topic and the variant — or the study the topic is built on names the variant among its own loci. It does not mean one explains the other.",
"gene": "OPRM1",
"kind": "topic",
"rsid": "rs1799971",
"target": "anxiety",
"target_name": "Anxiety and worry",
"target_url": "https://www.mygenelog.com/topics/anxiety",
"tier": "co-mention"
},
{
"edge_kind": "variant_topic",
"evidence": "A paper in Europe PMC names both the topic and the variant — or the study the topic is built on names the variant among its own loci. It does not mean one explains the other.",
"gene": "OPRM1",
"kind": "topic",
"rsid": "rs1799971",
"target": "blood-sugar",
"target_name": "Blood sugar and insulin",
"target_url": "https://www.mygenelog.com/topics/blood-sugar",
"tier": "co-mention"
},
{
"edge_kind": "variant_topic",
"evidence": "A paper in Europe PMC names both the topic and the variant — or the study the topic is built on names the variant among its own loci. It does not mean one explains the other.",
"gene": "OPRM1",
"kind": "topic",
"rsid": "rs1799971",
"target": "depression",
"target_name": "Depression and stress",
"target_url": "https://www.mygenelog.com/topics/depression",
"tier": "co-mention"
},
{
"edge_kind": "variant_topic",
"evidence": "A paper in Europe PMC names both the topic and the variant — or the study the topic is built on names the variant among its own loci. It does not mean one explains the other.",
"gene": "OPRM1",
"kind": "topic",
"rsid": "rs1799971",
"target": "exercise",
"target_name": "Exercise and muscle",
"target_url": "https://www.mygenelog.com/topics/exercise",
"tier": "co-mention"
},
{
"edge_kind": "variant_topic",
"evidence": "A paper in Europe PMC names both the topic and the variant — or the study the topic is built on names the variant among its own loci. It does not mean one explains the other.",
"gene": "OPRM1",
"kind": "topic",
"rsid": "rs1799971",
"target": "learning",
"target_name": "Learning and focus",
"target_url": "https://www.mygenelog.com/topics/learning",
"tier": "co-mention"
},
{
"edge_kind": "variant_topic",
"evidence": "A paper in Europe PMC names both the topic and the variant — or the study the topic is built on names the variant among its own loci. It does not mean one explains the other.",
"gene": "OPRM1",
"kind": "topic",
"rsid": "rs1799971",
"target": "longevity",
"target_name": "Longevity and ageing",
"target_url": "https://www.mygenelog.com/topics/longevity",
"tier": "co-mention"
},
{
"edge_kind": "variant_topic",
"evidence": "A paper in Europe PMC names both the topic and the variant — or the study the topic is built on names the variant among its own loci. It does not mean one explains the other.",
"gene": "OPRM1",
"kind": "topic",
"rsid": "rs1799971",
"target": "personality",
"target_name": "Personality",
"target_url": "https://www.mygenelog.com/topics/personality",
"tier": "co-mention"
},
{
"edge_kind": "variant_topic",
"evidence": "A paper in Europe PMC names both the topic and the variant — or the study the topic is built on names the variant among its own loci. It does not mean one explains the other.",
"gene": "OPRM1",
"kind": "topic",
"rsid": "rs1799971",
"target": "risk-taking",
"target_name": "Risk-taking and impulsivity",
"target_url": "https://www.mygenelog.com/topics/risk-taking",
"tier": "co-mention"
},
{
"edge_kind": "variant_topic",
"evidence": "A paper in Europe PMC names both the topic and the variant — or the study the topic is built on names the variant among its own loci. It does not mean one explains the other.",
"gene": "OPRM1",
"kind": "topic",
"rsid": "rs1799971",
"target": "running",
"target_name": "Running and endurance",
"target_url": "https://www.mygenelog.com/topics/running",
"tier": "co-mention"
},
{
"edge_kind": "variant_topic",
"evidence": "A paper in Europe PMC names both the topic and the variant — or the study the topic is built on names the variant among its own loci. It does not mean one explains the other.",
"gene": "OPRM1",
"kind": "topic",
"rsid": "rs1799971",
"target": "sleep",
"target_name": "Sleep and insomnia",
"target_url": "https://www.mygenelog.com/topics/sleep",
"tier": "co-mention"
},
{
"edge_kind": "variant_topic",
"evidence": "A paper in Europe PMC names both the topic and the variant — or the study the topic is built on names the variant among its own loci. It does not mean one explains the other.",
"gene": "OPRM1",
"kind": "topic",
"rsid": "rs1799971",
"target": "smoking",
"target_name": "Smoking and vaping",
"target_url": "https://www.mygenelog.com/topics/smoking",
"tier": "co-mention"
},
{
"edge_kind": "variant_topic",
"evidence": "A paper in Europe PMC names both the topic and the variant — or the study the topic is built on names the variant among its own loci. It does not mean one explains the other.",
"gene": "OPRM1",
"kind": "topic",
"rsid": "rs1799971",
"target": "sport",
"target_name": "Sport and performance",
"target_url": "https://www.mygenelog.com/topics/sport",
"tier": "co-mention"
}
],
"next": null,
"page": 1,
"pages": 1,
"per_page": 100,
"plan": "api_pro",
"prev": null,
"total": 17
}curl https://www.mygenelog.com/ko/api/v1/variants/rs4988235
{
"rsid": "rs4988235",
"slug": "rs4988235",
"lang": "ko",
"url": "https://www.mygenelog.com/ko/variants/rs4988235",
"gene": "MCM6 / LCT",
"trait_name": "유당 내성",
"risk_tier": "standard",
"genotypes": [
{
"genotype": "C/C",
"note": "조상형 대립유전자를 두 개 가진 유전형입니다. 발표된 연구는 이 유전형을 유당분해효소 비지속성(lactase non-persistence)과 연관 짓습니다 — 유당분해효소 생산이 대체로 유년기 이후 줄어들어, 성인기에 유제품 내성이 떨어질 수 있습니다."
},
{
"genotype": "C/T",
"note": "두 대립유전자를 한 개씩 가진 유전형입니다. 연구는 이 유전형에서 성인기까지 유당분해효소 생산이 이어질 가능성이 높다고 보지만, 개인별 내성 차이는 여전히 있습니다."
},
{
"genotype": "T/T",
"note": "파생형 대립유전자를 두 개 가진 유전형입니다. 발표된 연구는 이 유전형을 유당분해효소 지속성(lactase persistence)과 연관 지으며, 성인기까지 유당분해효소 생산이 이어지는 경우가 흔합니다."
}
],
"guideline_note": "실제 유당 내성은 유전자 하나로만 결정되지 않고 시간이 지나며 달라질 수 있습니다. 유제품이 불편함을 준다면 식단을 바꾸기 전에 의사와 상의하십시오.",
"source_note": "Enattah et al. 2002, Nature Genetics — identification of a variant associated with adult-type hypolactasia. PMID 11788828.",
"updated_at": "2026-09-09",
"alternates": {
"en": "https://www.mygenelog.com/api/v1/variants/rs4988235",
"ko": "https://www.mygenelog.com/ko/api/v1/variants/rs4988235"
}
}Rebuilt hourly from the live catalogue. If a field disappears from the API it disappears from this page in the same deploy, which is the only way a sample response stays true.
10 requests / minute per IP, with no key and no login — this is served off a small VPS, not a CDN. Going over returns HTTP 429. A free key carries the same 10: it puts you on the record so we can tell you when something changes, and buys no capacity, because handing the higher limit away free would make "no key required" true on paper and false in practice. Pro raises it to 30 a minute, Business to 400. The list endpoints are paged: 100 rows a page free, 500 on Pro or Business, each response carrying a next link. There is no unpaged mode — if you want the whole dataset as one file, take a dated Gene Knowledge Graph Snapshot, which is hashed and citable and is the better artefact anyway.
This data is informational summaries of published research, not medical advice, and not a diagnosis. If you build a product on it, don't present it as one either — carry the same non-diagnostic framing through to your own users. See Terms of Use.
A free, open JSON API serving MyGeneLog™'s plain-language genetic variant dataset — variants, conditions, drugs and the chemosensory set — with no key or signup required for the open endpoints.
No key is required for the free endpoints, though a free one is available if you would rather be identified than anonymous. The limit is 10 requests a minute per IP or key; going over returns HTTP 429. Pro raises it to 30 a minute, Business to 400.
The free API gives every node and its writing — genotype interpretations, cited sources, open questions — paged 100 rows at a time. Pro and Business both add the Gene Knowledge Graph, the links between entities served as data, at 500 rows a page, and a higher rate limit. They differ in licence and capacity: Pro is for your own personal or internal project; Business is licensed to embed the links in a commercial product or service you sell or operate, at a higher rate limit.
Yes, under CC BY 4.0, with one exception: the Gene Knowledge Graph links sold through Pro or Business are licensed to the subscriber, not released under CC BY 4.0 — Pro for your own personal or internal project, Business for embedding in a product you sell or operate. Neither licence permits redistributing the links as a dataset.
If you're an individual experimenting or building a personal project, Pro. If you're a company shipping a product — including an AI/LLM product that calls this data at request time — that's exactly what Business is priced and rate-limited for.
Yes. Every endpoint is served in English at /api/v1/… and in Korean by putting /ko in front of the path — same endpoint, same parameters, same shape of response.
Model Context Protocol is a standard that lets an AI assistant call a piece of software directly while answering you, instead of you copying data into the chat yourself. MyGeneLog's MCP server, live at /mcp, lets an AI look up variants, genes, conditions and drugs from this same dataset as part of a conversation — it changes nothing about the free site, the REST API, or your account.