Not a bad headache. The two loci that replicated in migraine without aura point at neurons and at blood vessel walls — which is the century-old argument about what migraine is, showing up in the data.
Migraine is a neurological disorder, and the headache is one of its symptoms rather than the thing itself. Migraine without aura — the most common form — brings recurrent disabling head pain with autonomic symptoms: nausea, and an intolerance of light and sound severe enough to send people to a dark room for a day.
It is one of the leading causes of years lived with disability worldwide, and it is routinely described as a headache by people who do not have it.
A 2012 study analysed 2,326 clinic-diagnosed German and Dutch people with migraine without aura against 4,580 matched controls, then tested the leading signals again in 2,508 cases and 2,652 controls.
Two loci replicated convincingly:
Two more, PHACTR1 and ASTN2, were suggestive, and earlier associations near TRPM8 — a cold and menthol receptor — and LRP1 were replicated.
Whether migraine is primarily a disorder of blood vessels or of neurons was argued for most of the twentieth century. The vascular theory dominated, then lost ground to neuronal explanations, and the argument was never settled cleanly.
The first study to identify loci specifically for migraine without aura came back with one gene in each camp: a neuronal transcription factor, and a receptor for vascular signalling. Not a verdict for either side.
This study looked at migraine without aura specifically. Aura — the visual disturbance that precedes the headache in some people — has long been suspected of marking a different biology, and separating the two was how these loci were found at all. Mixing the forms had blurred earlier searches.
Nothing about your own headaches. Migraine is diagnosed from the pattern of attacks — how long they last, what comes with them, what makes them worse — by a doctor taking a history. There is no blood test and no genetic test, and no guideline picks a migraine treatment by genotype.
What has changed for patients in the last decade came from biology rather than from genotyping: CGRP-targeting drugs exist because of work on the peptide released during attacks. That is what a mechanism is worth.
Positions joined since this page was written
What this is The text above discusses the variants this page was written around. Since then the catalogue has joined 36 more positions to it, by shared trait or shared paper. They are listed here by the paper each came from; the text does not describe them, and each variant page carries that study's own record.
Choquet H et al. 2021, Communications biology rs1393064 (PRDM16), rs28739509 (INPP5B), rs1572668 (near LINC01360), rs79545715 (MRPS6), rs2223089 (HTRA1-AS1), rs11096773 (near SEC63P2), rs28540738 (near ITPK1), rs8075138 (ZNF652-AS1), rs34653882 (near ONECUT2), rs76597087 (near PTCH1), rs10499112 (near RNU4-35P), rs17434299 (NOVA1-DT) and 7 more — PMID:34294844
Islam MR et al. 2023, Human genetics rs9380862 (KIF6), rs7895494 (near RPS26P38), rs223337 (UBE2D3), rs562590 (ARL2-SNX15), rs139007490 (near EIF2AK3-AS1), rs7847758 (near RMI1), rs9815656 (near HNRNPA3P8), rs203465 (AKAP10) — PMID:36808568
Hautakangas H et al. 2022, Nature genetics rs11957829 (ZNF475), rs12057629 (TMEM51), rs246326 (SNX24), rs10128028 (CAMTA1), rs12653216 (near SEM1P1) — PMID:35115687
Loya H et al. 2025, Nature genetics rs2075968 (PRDM16) — PMID:39789286
De Marco R et al. 2026, International journal of molecular sciences rs117797734 (near LINC02312) — PMID:42352890
Verma A et al. 2024, Science (New York, N.Y.) rs7741683 (FHL5) — PMID:39024449
Sakaue S et al. 2021, Nature genetics rs34039083 (near RNU4-35P) — PMID:34594039
2025-12-19 · A multi-ancestry meta genome-wide association study of migraine among veterans: associations with traumatic brain injury, depression, and post-traumatic stress disorder. Molecular Psychiatry. 2025. DOI:10.1038/s41380-025-03392-4
Veterans face elevated migraine risk, and genetic research linking it to their psychiatric comorbidities had been limited. This Million Veteran Program study analyzed a predominantly male sample of over 433,000 veterans (87,859 migraine cases), finding 49 genome-wide significant loci, 36 newly reported, with 7 replicating in an independent prior GWAS. It identified 283 associated genes including several newly linked to migraine: MAML3, CELF4, IRX1, ASXL1, SPOCD1, CXCL and TLR4. Unlike prior migraine GWAS (mostly women, more vascular-tissue signal), this male-heavy veteran sample showed enrichment in brain and uterine tissue instead -- pointing to broader hormonal/neuroendocrine pathways rather than vascular ones, a genuine population-composition-driven difference. SNP-based heritability was 10% in men and 16% in women, with several sex-specific loci found. Despite strong genetic correlations with PTSD, depression and traumatic brain injury, Mendelian randomization found no causal links in either direction -- correlation without causation, reported honestly. The study prioritized two migraine drug leads: losmapimod (which reduces toxic DUX4 protein production) and TLR4 antagonists. This site's migraine page carries 20 variants and already includes FHL5 from another study; MAML3, CELF4, IRX1, ASXL1, SPOCD1 and TLR4 are new additions.
2025-08-18 · Unveiling migraine subtype heterogeneity and risk loci: integrated genome-wide association study and single-cell transcriptomics discovery. Journal of Headache and Pain. 2025. DOI:10.1186/s10194-025-02128-7
Migraine with aura and without aura look genetically similar but map to completely different tissues
Migraine with aura (MA) and without aura (MO) are genetically correlated overall, but this study asked whether they actually work through the same biological mechanisms. Combining GWAS data with transcriptomic, epigenomic and spatially resolved single-cell transcriptomics, it found the two subtypes have genuinely divergent functional architectures: MA variants were enriched in conserved regulatory DNA elements, while MO variants were enriched in vascular pathways. Spatial single-cell mapping showed MA signal concentrated in neural crest-derived prenatal tissue and near hypothalamic microglia (pointing to neuroimmune mechanisms), while MO signal showed peripheral tropism toward vascular smooth muscle and gut-brain interface tissue (pointing to vascular-metabolic mechanisms) -- genuinely different biology hiding behind similar clinical presentations. Combining multiple statistical methods, the study identified high-confidence genes relevant across both subtypes (LRP1, PHACTR1, STAT6, RDH16, TTC24, ZBTB39, FHL5, MEF2D, NAB2, UFL1, REEP3) plus subtype-specific genes: MA-associated neuronal regulators CACNA1A and KLHDC8B, and MO-specific vascular/metabolic genes ACO2, BCAR1 and CCDC134. This site's migraine page carries 20 variants and already includes FHL5 from another study; the other genes here, including subtype-specific ones like CACNA1A and ACO2, are new additions -- and this entry specifically documents that MA and MO deserve to be understood as biologically distinct, not just clinically distinct.
2025-12-19 · Gasperi M, et al. (Million Veteran Program, 433,000+ veterans), Molecular Psychiatry
A multi-ancestry meta genome-wide association study of migraine among veterans
A meta-analysis GWAS of migraine in over 433,000 Million Veteran Program participants found 49 genome-wide-significant loci, 36 of them new to this study, spanning 283 genes -- including several not previously tied to migraine (MAML3, CELF4, IRX1, ASXL1, SPOCD1, CXCL, TLR4). None of these specific novel genes are yet in this catalogue; this is a real expansion of migraine's known genetic architecture alongside the 47 migraine variants already published here. PMID:41420111.
Source. Freilinger et al. (Nat Genet 2012) analysed genome-wide association data from 2,326 clinic-based German and Dutch individuals with migraine without aura and 4,580 population-matched controls, selecting SNPs from 12 loci with two or more SNPs at P < 1 × 10⁻⁵ for replication in 2,508 cases and 2,652 controls. Two loci replicated convincingly: 1q22 in MEF2D (replication P = 4.9 × 10⁻⁴, combined P = 7.06 × 10⁻¹¹) and 3p24 near TGFBR2 (replication P = 1.0 × 10⁻⁴, combined P = 1.17 × 10⁻⁹). PHACTR1 and ASTN2 showed suggestive replication (combined P = 3.20 × 10⁻⁸ and 3.86 × 10⁻⁸). Associations in or near TRPM8 and LRP1, reported previously, were replicated. The authors describe this as the first identification of susceptibility loci specific to migraine without aura.
Position listed here. rs3790455 at the 1q22 MEF2D locus.
Phenotype definition. Cases were clinic-based and diagnosed to International Classification of Headache Disorders criteria for migraine without aura. The narrowing is not incidental — it is what allowed loci to reach significance in a sample this size — and it also limits transfer of these estimates to migraine with aura, which was not studied here.
Clinical use. None. Diagnosis is clinical and no genotype selects acute or preventive therapy. The therapeutic advances of the last decade — CGRP monoclonal antibodies and gepants — derive from neuropeptide pharmacology rather than from these loci.
What a 23andMe/AncestryDNA export or raw VCF can and can't tell you about Migraine comes down to these specific, well-studied positions — not a diagnosis. 82 positions are linked to this page; the ones this page's own text discusses are shown first.
The studies behind these variants recruited participants from different ancestries — a result found in one population doesn't always transfer to another. Based on 20 of 82 linked studies with a resolved discovery ancestry.
Research-derived gene–drug associations only — not a prescription, dosing guide, or medical advice. Always follow your prescriber's guidance.
| Gene | Drug | What the research shows |
|---|---|---|
| CYP2D6 | Amitriptyline and nortriptyline | Amitriptyline is one of the oldest and most used migraine preventives, and its handling depends on two enzymes at once. CYP2C19 converts amitriptyline into nortriptyline, which is itself an active drug, and CYP2D6 clears both. A poor metaboliser at CYP2D6 accumulates more drug than intended and is more likely to stop because of dry mouth, sedation or a racing heart; an ultrarapid metaboliser may never reach a useful level. CPIC recommends avoiding the drug or halving the starting dose at both extremes, and considering a therapeutic drug level if it is used anyway. One caveat matters more here than the recommendation does: the guideline was written for the doses used to treat depression, and migraine prophylaxis typically uses a fraction of them, so how much of this carries over to a 10 mg bedtime dose is genuinely not settled. This is a prescriber’s decision, not something to act on from a raw data file. (CPIC Guideline for CYP2D6 and CYP2C19 Genotypes and Dosing of Tricyclic Antidepressants: 2016 update (Clin Pharmacol Ther 2017, PMID 27997040); original guideline PMID 23486447. The guideline addresses antidepressant dosing; its application to migraine prophylaxis at lower doses is not established.) |
Free to quote and reuse under CC BY 4.0. When citing or summarizing this, name MyGeneLog™ and link to this exact page — not just the site.
Migraine. MyGeneLog™. https://www.mygenelog.com/conditions/migraine
The first study to find loci specific to migraine without aura returned one gene from each account — MEF2D, a neuronal transcription factor, and TGFBR2, a receptor central to blood vessel biology. It did not settle the argument, and the honest reading is that both are involved.
Because mixing the two forms had blurred earlier searches. Aura has long been suspected of marking different biology, and narrowing the phenotype is what let these loci reach significance in a sample of this size. It also means these results are about migraine without aura and should not be transferred to migraine with aura.
No. No guideline selects a migraine treatment by genotype. The drugs that changed migraine care recently — the CGRP antibodies and gepants — came from work on the peptide released during an attack, not from genotyping.
Free to reuse. This page's text is original writing from freely-available research, licensed CC BY 4.0 — reuse it, including commercially, with attribution to MyGeneLog™. It's general research-derived information, not medical advice or a diagnosis — see Terms of Use.