Sage Journals: Discover world-class research

Abstract

Background

Migraine is considered a multifactorial genetic disorder. Different platforms and methods are used to unravel the genetic basis of migraine. Initially, linkage analysis in multigenerational families followed by Sanger sequencing of protein-coding parts (exons) of genes in the genomic region shared by affected family members identified high-effect risk DNA mutations for rare Mendelian forms of migraine, foremost hemiplegic migraine. More recently, genome-wide association studies testing millions of DNA variants in large groups of patients and controls have proven successful in identifying many dozens of low-effect risk DNA variants for the more common forms of migraine with the number of associated DNA variants increasing steadily with larger sample sizes. Currently, next-generation sequencing, utilising whole exome and whole genome sequence data, and other omics data are being used to facilitate their functional interpretation and the discovery of additional risk factors. Various methods and analysis tools, such as genetic correlation and causality analysis, are used to further characterise genetic risk factors.

Findings

We describe recent findings in genome-wide association studies and next-generation sequencing analysis in migraine. We show that the combined results of the two most recent and most powerful migraine genome-wide association studies have identified a total of 178 LD-independent (r² < 0.1) genome-wide significant single nucleotide polymorphisms (SNPs), of which 99 were unique to Hautakangas et al., 11 were unique to Choquet et al., and 68 were identified by both studies. When considering that Choquet et al. also identified three SNPs in a female-specific genome-wide association studies then these two recent studies identified 181 independent SNPs robustly associated with migraine. Cross-trait and causal analyses are beginning to identify and characterise specific biological factors that contribute to migraine risk and its comorbid conditions.

Conclusion

This review provides a timely update and overview of recent genetic findings in migraine.

Keywords

Variant mutation migraine headache risk disease mechanism experimental model

Introduction

Migraine is considered a multifactorial (complex) genetic disorder with a strong familial aggregation (1,2). Complex traits are typically brought about by a combination of multiple genetic variants, as well as behavioural and environmental factors, each with a small effect size. In this review, we provide an overview of the recent genetic findings with a focus mainly on common polygenic migraine. Furthermore, we discuss state-of-the-art strategies for performing genetic studies of migraine.

Rationale for genetic studies

When identifying the gene(s) involved in a disorder, different approaches are used depending on the disorder’s genetic architecture (i.e. monogenic, oligogenic, or polygenic). The more oligogenic a disease is (i.e. the lesser the number of genes involved, with monogenic being the extreme), the larger the effect size of the associated genetic variant(s) tends to be, in line with epidemiological data from disorders where rare disorders are monogenic and common disorders are polygenic.

Hemiplegic migraine

A substantial amount of our knowledge of molecular mechanisms in migraine pathophysiology came from studying rare hemiplegic migraine (HM) (3). The classical linkage method in migraine research was used to study large families with HM and revealed a clear Mendelian (monogenic) type of inheritance. This led to the identification of three undisputed HM genes; CACNA1A (FHM1), ATP1A2 (FHM2), and SCN1A (FHM3) (4 –6). Detailed information on how mutations in HM can be studied in cellular and animal models and how they cause disease can be found elsewhere (see [7,8]).

In many patients with HM no pathogenic mutation has been detected in the HM genes (9,10). In recent years, whole-exome (next-generation) sequencing (WES) has been used to try and identify additional causal genes in patients without mutations in the known HM genes, but this has proven difficult and no “fourth” gene has been identified thus far (11). A study by Pelzer et al. (11) did, however, show that patients with a more severe phenotype were more prone to have a causal mutation in one of the HM genes. Patients with a causal mutation in CACNA1A, ATP1A2, or SCN1A had a lower age-at-onset, more affected family members, and had attacks more frequently. Moreover, attacks were i) brought about by mild head trauma, ii) typically with extensive motor weakness, and iii) with brainstem features, confusion, and brain oedema. Noteworthy, progressive ataxia and intellectual disability were only found in patients with a causal gene mutation (11). As no causal mutation was found in “milder” patients, it was proposed that such HM patients have the more extreme phenotype in the migraine with aura continuum (9). Illustrative of this is a Finnish polygenic risk score study that showed that patients with HM, but without a high-penetrant disease-causing mutation in a known HM gene, carry an excess of genome-wide association studies (GWAS) variants associated with common migraine compared to patients suffering from the common migraine subtypes (12), suggesting a spectrum ranging from common low-risk variants to rare high-penetrant mutations contributing to the risk for migraine. Further support for this are loss-of-function mutations in PRRT2, which do not cause HM on their own, but rather function as modifying risk factors (13). A further illustration of the complex genetic architecture of HM is a recent whole-genome sequencing (WGS) study where patients with HM were more likely to accumulate frameshift indels in multiple genes that have a role in synaptic signalling in the central nervous system compared to patients with common migraine (14).

Genetic studies in common migraine

Various twin and familial studies investigating the genetic and environmental susceptibility in migraine have shown that migraine is a multifactorial (complex) genetic disorder with a strong familial aggregation (1,2). The heritability of migraine was estimated to range from 35% to 60% (15). Population-based studies have shown that the relative risk for a first-degree relative of a migraine patient is increased by 1.5- to 4-fold in comparison to a patient in the general population (1). The risk was highest for those patients with a higher pain score and frequency of attacks, an early age of disease onset, and a migraine with aura phenotype (1,2,16). Studies of twins identified a higher genetic load in migraine with aura compared with migraine without aura (17). Migraine frequency, being the number of migraine days per month, appears mainly to be associated with a genetic predisposition in males (16). A stronger family history of migraine is also associated with migraine with aura, a lower age-at-onset and more medication days (16). For decades, identifying gene variants involved in complex disorders, such as migraine, has proven challenging.

Genome-wide association studies in migraine

As a result of the improvement in DNA technology and the advancement of cost-effective genotyping platforms, GWAS has become the method of choice to identify DNA risk variants for complex traits. Typically, in GWAS, several millions of single nucleotide polymorphisms (SNPs) are tested for association with migraine by assessing differences in allele frequencies between very large numbers of (migraine) patients and controls. Of note, only common variants with a low to high minor allele frequency (MAF ≥ 0.01) are interrogated.

Since 2010, the International Headache Genetics Consortium (IHGC; www.headachegenetics.org/) has conducted several migraine GWAS, and with increasing sample sizes, the number of associated risk variants steadily expanded.

The 2016 migraine GWAS by Gormley et al. (18) reported 44 linkage disequilibrium (LD)-independent SNPs mapping to 38 distinct genomic loci associated with migraine. Downstream bioinformatics analyses focussed on the respective functions of the likely associated genes and concluded that GWAS findings in migraine are primarily linked to arterial and smooth muscle function. This is in line with the shared polygenic risk between migraine, stroke, and cardiovascular diseases that has been described previously (19,20). Finucane et al. (21), through heritability enrichments of the GWAS summary statistics, also showed a neurological enrichment for migraine (all subtypes) and a cardiovascular enrichment for migraine without aura. In addition to vascular and neuronal mechanisms, additional mechanisms like metal ion homeostasis were indicated to contribute to migraine susceptibility (22).

The most recent IHGC migraine GWAS published in 2022 by Hautakangas et al. (23) included 102,084 cases and 771,257 controls and identified 123 distinct genomic regions associated with migraine, of which 86 loci were novel compared to the 2016 migraine GWAS (18,23). In addition, the authors reported 48 genome-wide significant variants that were LD-independent (r² < 0.1 within either UK Biobank [UKBB; www.ukbiobank.ac.uk] or 23andMe [www.23andme.com]) of the 123 lead variants, bringing the total to 171 independent index SNPs. When examining the 171 SNPs in the 1000 Genomes Project (1kGP; www.internationalgenome.org) EUR reference panel, four SNPs were found to be in LD (rs139846391–rs2078371, r² =0.247; rs4240603–rs4599655, r² = 0.564; rs1416901–rs10828247, r² = 0.958; rs7148352–rs11624776, r² = 0.102), resulting in 167 independent SNPs (r² < 0.1) (Online Supplementary Table 1).

Enrichment analyses in the 2022 migraine GWAS implicated the involvement of both vascular and central nervous system tissue/cell types, well in line with previous data and the concept that migraine is a neurovascular disorder. In addition, for the first time, risk loci were found that encode genes that had been identified as migraine drug targets, namely the calcitonin gene-related peptide (CGRP encoded by CALCA/CALCB) and the serotonin 1F receptor (HTR1F). An increase in CGRP has been observed in migraine patients outside and during an attack (24,25), which led to the development of CGRP (receptor)-related small molecule (gepants) and monoclonal antibodies with proven efficacy to treat migraine (26). The gene that encodes the serotonin 5-HT_1F receptor is a target for which migraine drugs are being developed; these drugs are known as ditans (27).

The 2022 migraine GWAS also evaluated both shared and distinct genetic components of migraine subtypes (migraine without aura and migraine with aura) (23). A subsequent analysis of the 123 lead index SNPs using 29,679 (15,055 migraine without aura and 14,624 migraine with aura cases) clinically well-diagnosed cases, indicated with a probability above 95% three risk variants (in MPPED2, CACNA1A and HMOX2) that were specific for migraine with aura, two (one near FECH and one near SPINK2) that were specific for migraine without aura, and nine (i.e. DLST, PRMD16, SUGCT, PLEC1, MRVI1, LRP1, FHL5, near TRPM8, and near FGF6) that were associated with both migraine subtypes (23). For comparison, in the 2016 migraine GWAS, seven loci (i.e. ASTN2, PHACTR1, LRP1, FHL5, TRPM8, near FGF6, and near TSPAN2) were associated specifically with migraine without aura (in a sample of 8,348 cases vs 139,622 controls) and no gene association was found for migraine with aura (6,332 cases vs 144,883 controls) (18). In the 2022 migraine GWAS, there was no evidence that any of the seven loci were specific for migraine without aura, still, four of them (LRP1, FHL5, near TRPM8, and near FGF6) are among the nine loci that were associated with both subtypes. The fact that multiple loci are strongly associated with both migraine with and migraine without aura provides further genetic evidence for the concept that migraine subtypes in part share a genetic background (7,28,29). One of the new migraine loci identified, the CACNA1A gene, links to both monogenic and polygenic forms of migraine (7,23).

In a third recent migraine GWAS from 2021 by Choquet et al. (30), 79 independent loci were significantly associated with migraine (online Supplementary Table 2). Of note, this was an ethnically diverse study that included adult individuals (28,852 cases vs. 525,717 controls) from East Asian, African American, and Hispanic/Latino descent.

Both Hautakangas et al. (23) and Choquet et al. (30) GWAS used data from Gormley et al. (18) supplemented with additional cohorts (online Supplementary Table 3) and both compared their findings with those of Gormley et al. (18), but not with each other. Therefore, we compiled and compared the 171 (167 LD-independent with r² < 0.1) genome-wide significant SNPs listed in Supplementary Table 1 of Hautakangas et al. (23) and the 95 (79 LD-independent with r² < 0.1) genome-wide significant SNPs drawn from Supplementary Data 1 and Supplementary Data 4 of Choquet et al. (30). Analysis of these SNP lists resulted in a total of 178 LD-independent (r² < 0.1) genome-wide significant SNPs across both studies, of which 99 were unique to Hautakangas et al. (23), 11 were unique to Choquet et al. (30), and 68 were found in both studies (Figure 1, Table 1). This means that 41% of the 167 Hautakangas et al. (23) index SNPs were genome-wide significant in Choquet et al. (30) and 86% of the 79 Choquet et al. (30) index SNPs were genome-wide significant in Hautakangas et al. (23). In addition, three additional (novel) loci rs1047891, s11718509, and rs10150336 (CPS1, PBRm1, and SLC25A21, respectively) specific for women with migraine, were identified by Choquet et al. (30). This means that there are now at least 181 SNPs robustly associated with migraine.

Figure 1.

Venn diagram of the genome-wide significant index SNPs across Hautakangas et al. (2022) (dark grey) and the Choquet et al. (2021) (light grey) migraine genome-wide association studies. The size of the circles is representative of the number of SNPs.

Table 1.

178 independent (r² < 0.1) SNPs from Hautakangas et al. (2022) (23) en Choquet et al. (2021) (30) associated with migraine.

Chr	Pos	SNP	P value	Study with most significant SNP at independent GWS locus
1	2833427	rs2124663	2.44 × 10⁻⁹	(23)
1	3075597	rs10218452	7.26 × 10⁻⁷¹	(23)
1	3105160	rs2817137	7.11 × 10⁻¹⁶	(23)
1	3229134	rs2483262	1.36 × 10⁻¹⁵	(23)
1	7055843	rs10128028	7.66 × 10⁻⁹	(23)
1	8407287	rs10399665	3.67 × 10⁻¹⁰	(30)
1	15490865	rs61561984	1.59 × 10⁻⁸	(23)
1	15538493	rs12057629	9.38 × 10⁻¹⁴	(23)
1	38366907	rs28739509	2.64 × 10⁻¹⁰	(23)
1	39590409	rs1472662	1.75 × 10⁻⁸	(23)
1	60529980	rs11578492	6.24 × 10⁻⁹	(23)
1	66178918	rs7511672	1.43 × 10⁻⁹	(23)
1	73666225	rs200310839	2.02 × 10⁻⁸	(23)
1	73891226	rs56019088	7.33 × 10⁻¹³	(23)
1	92177663	rs11165300	4.72 × 10⁻⁸	(23)
1	115677183	rs2078371	5.84 × 10⁻⁴²	(23)
1	115820598	rs11102915	4.10 × 10⁻¹⁶	(23)
1	115836207	rs11466096	1.02 × 10⁻⁸	(23)
1	149880863	rs68002561	3.61 × 10⁻¹⁰	(23)
1	149897217	rs7544531	1.31 × 10⁻¹⁰	(23)
1	150510660	rs6693567	1.25 × 10⁻¹³	(23)
1	156450873	rs2274319	2.75 × 10⁻⁴¹	(23)
1	174601659	rs11487328	1.71 × 10⁻⁸	(23)
1	175161930	rs10912903	7.14 × 10⁻⁹	(30)
1	186913055	rs6668908	2.22 × 10⁻⁸	(23)
1	206848306	rs4074957	1.68 × 10⁻⁹	(30)
1	245847455	rs72764846	5.41 × 10⁻⁹	(23)
2	43649780	rs12712881	3.50 × 10⁻¹⁰	(23)
2	96576609	rs4907224	1.63 × 10⁻⁹	(23)
2	145258445	rs7564469	5.05 × 10⁻⁹	(23)
2	146037564	rs895219	3.75 × 10⁻¹¹	(23)
2	156416638	rs843215	2.62 × 10⁻⁸	(23)
2	171234235	rs4668251	7.58 × 10⁻⁹	(23)
2	176978383	rs72923449	4.66 × 10⁻⁸	(23)
2	203832867	rs138556413	4.15 × 10⁻¹⁶	(23)
2	204062727	rs112494629	1.57 × 10⁻⁸	(23)
2	234731883	rs7603146	4.42 × 10⁻⁹	(23)
2	234745389	rs213551	1.83 × 10⁻¹⁰	(23)
2	234756811	rs566529	1.30 × 10⁻¹⁴	(23)
2	234768944	rs17862906	7.54 × 10⁻¹¹	(23)
2	234825093	rs10166942	9.32 × 10⁻⁵¹	(23)
3	19740933	rs13087932	3.38 × 10⁻⁸	(30)
3	30472786	rs7371912	1.06 × 10⁻¹⁴	(23)
3	48498456	rs7618883	4.17 × 10⁻⁸	(23)
3	80302512	rs950570	1.30 × 10⁻⁸	(23)
3	86149109	rs73138150	1.95 × 10⁻⁸	(23)
3	88210464	rs6795209	1.23 × 10⁻⁸	(23)
3	124607055	rs1499963	7.48 × 10⁻⁹	(23)
3	154289946	rs13078967	2.16 × 10⁻¹⁶	(23)
4	35483556	rs11096773	4.46 × 10⁻¹²	(30)
4	57727311	rs7684253	4.21 × 10⁻¹⁴	(23)
4	57765132	rs4865164	1.23 × 10⁻⁹	(23)
5	74994913	rs4704232	2.50 × 10⁻¹³	(30)
5	81129663	rs12653216	8.07 × 10⁻⁹	(23)
5	92439320	rs1496332	4.96 × 10⁻⁸	(30)
5	121515195	rs11957829	1.58 × 10⁻⁹	(23)
5	122306398	rs246326	6.80 × 10⁻¹⁰	(23)
5	145752008	rs10038882	1.33 × 10⁻¹²	(23)
5	149380493	rs4705403	1.18 × 10⁻⁸	(23)
5	172645766	rs6556059	8.15 × 10⁻¹⁰	(23)
5	176676461	rs10866704	2.10 × 10⁻⁸	(23)
6	12892394	rs6930415	4.60 × 10⁻⁸	(23)
6	12903957	rs9349379	1.41 × 10⁻⁴⁷	(23)
6	12961929	rs62386823	2.35 × 10⁻¹⁰	(23)
6	22131929	rs9295536	7.76 × 10⁻¹²	(23)
6	30749712	rs114680917	2.38 × 10⁻⁸	(23)
6	31850308	rs74434374	4.51 × 10⁻⁹	(23)
6	39183470	rs10456100	9.14 × 10⁻¹⁹	(23)
6	39191384	rs2561401	1.35 × 10⁻⁸	(23)
6	72328062	rs6906594	1.90 × 10⁻¹¹	(30)
6	96701026	rs7769734	3.93 × 10⁻⁸	(23)
6	96732257	rs4599655	6.11 × 10⁻¹⁴	(23)
6	96938564	rs138899284	2.64 × 10⁻¹³	(23)
6	96960016	rs77062342	2.77 × 10⁻¹⁵	(23)
6	97059666	rs11153082	7.23 × 10⁻⁵⁴	(23)
6	111713302	rs6568677	2.09 × 10⁻⁸	(23)
6	121750995	rs17052911	2.89 × 10⁻⁸	(23)
6	121846038	rs28455731	8.82 × 10⁻²³	(23)
6	126049040	rs1268083	3.15 × 10⁻¹⁰	(30)
6	150133954	rs9383843	1.35 × 10⁻⁹	(23)
7	40408717	rs554159536	1.16 × 10⁻⁸	(23)
7	40427617	rs10234636	4.44 × 10⁻²⁸	(23)
7	73013901	rs13235543	3.06 × 10⁻¹³	(23)
7	120481569	rs56067931	4.83 × 10⁻⁸	(23)
8	27266287	rs11782789	3.03 × 10⁻⁹	(23)
8	64496159	rs4739105	2.84 × 10⁻⁸	(23)
9	14103618	rs580845	4.30 × 10⁻⁸	(23)
9	29372501	rs10156578	3.34 × 10⁻¹²	(23)
9	71746838	rs7034179	1.61 × 10⁻¹⁶	(23)
9	98291853	rs76597087	2.29 × 10⁻¹¹	(30)
9	109687403	rs17723637	8.64 × 10⁻⁹	(23)
9	119245085	rs7858153	2.35 × 10⁻²⁴	(30)
9	119479774	rs2165554	1.40 × 10⁻¹⁰	(23)
9	140743200	rs4278223	6.23 × 10⁻¹⁰	(23)
10	8722944	rs7916911	3.19 × 10⁻¹²	(23)
10	21822856	rs10828247	7.51 × 10⁻⁹	(23)
10	33468124	rs2506142	5.53 × 10⁻⁹	(30)
10	74031313	rs4415679	2.03 × 10⁻⁸	(30)
10	96039597	rs2274224	3.29 × 10⁻²⁶	(23)
10	100702737	rs12260159	7.33 × 10⁻¹⁶	(23)
10	104741114	rs12260436	7.30 × 10⁻¹⁰	(23)
10	112502662	rs869432	3.54 × 10⁻⁸	(23)
10	124210160	rs2223089	1.16 × 10⁻¹¹	(23)
10	124230750	rs2672592	1.22 × 10⁻¹²	(23)
10	125184352	rs76663352	3.14 × 10⁻⁸	(23)
10	125185442	rs845102	1.70 × 10⁻⁸	(23)
10	125242283	rs11248546	1.59 × 10⁻¹²	(23)
10	125274226	rs10902940	1.31 × 10⁻⁸	(30)
10	134479675	rs200314499	6.92 × 10⁻¹²	(23)
11	3249984	rs12295710	2.86 × 10⁻¹⁶	(23)
11	10674044	rs4910165	1.09 × 10⁻²⁴	(23)
11	15126085	rs1003194	2.42 × 10⁻¹⁰	(23)
11	30547438	rs11031122	6.92 × 10⁻¹⁰	(23)
11	46548094	rs7932866	2.38 × 10⁻⁹	(23)
11	47669619	rs12419507	4.53 × 10⁻⁹	(23)
11	61697078	rs12787928	6.85 × 10⁻⁹	(23)
11	66401373	rs566673	9.07 × 10⁻⁹	(23)
11	102070976	rs12226331	1.93 × 10⁻¹³	(23)
11	133745852	rs10894756	2.83 × 10⁻⁸	(23)
12	4527322	rs2160875	2.72 × 10⁻³⁶	(23)
12	41901277	rs1458170	5.74 × 10⁻⁹	(23)
12	57259205	rs3922405	1.15 × 10⁻⁹	(23)
12	57460694	rs7974618	3.43 × 10⁻¹⁵	(23)
12	57506350	rs3001428	1.40 × 10⁻⁸	(23)
12	57524108	rs35616842	6.90 × 10⁻¹⁰	(23)
12	57527283	rs11172113	1.36 × 10⁻⁹⁰	(23)
12	57536989	rs34393552	1.55 × 10⁻¹²	(23)
12	57544931	rs57916742	4.56 × 10⁻¹⁰	(23)
12	90091782	rs4842676	2.26 × 10⁻⁹	(23)
12	98498223	rs10777902	1.25 × 10⁻¹⁰	(23)
12	124820705	rs1271309	3.74 × 10⁻⁸	(23)
13	47193696	rs7335684	1.05 × 10⁻⁸	(23)
13	78876537	rs7996252	4.11 × 10⁻⁸	(23)
13	110788441	rs2000660	4.96 × 10⁻⁸	(23)
14	27661650	rs1245463	5.72 × 10⁻¹⁴	(23)
14	42548912	rs1542668	2.53 × 10⁻⁸	(23)
14	58761912	rs28756401	6.41 × 10⁻⁹	(23)
14	75362552	rs55707505	2.48 × 10⁻⁸	(23)
14	76496477	rs75002882	9.22 × 10⁻⁹	(23)
14	93595591	rs11624776	9.73 × 10⁻¹⁹	(23)
14	94844947	rs28929474	2.54 × 10⁻⁹	(23)
15	81022364	rs12708529	8.12 × 10⁻¹⁰	(23)
16	4534482	rs12598836	2.21 × 10⁻¹⁰	(23)
16	19975407	rs8054079	3.83 × 10⁻⁸	(30)
16	75442143	rs77505915	4.75 × 10⁻¹⁴	(23)
16	87578039	rs8052831	8.24 × 10⁻¹⁵	(23)
17	1967501	rs9894634	9.63 × 10⁻¹¹	(23)
17	7366619	rs34914463	2.41 × 10⁻⁹	(23)
17	7571752	rs78378222	4.93 × 10⁻⁸	(23)
17	46619175	chr17:46619175:D	1.59 × 10⁻⁸	(23)
17	46632679	rs11652860	1.07 × 10⁻⁸	(23)
17	47514039	rs2119930	6.68 × 10⁻¹⁵	(23)
17	60720058	rs12452590	2.03 × 10⁻¹⁰	(23)
17	77925681	rs1285294	4.32 × 10⁻⁸	(23)
17	78256432	rs8077768	9.32 × 10⁻¹³	(23)
17	81015295	rs11655891	1.43 × 10⁻⁸	(30)
18	20201527	rs7506921	1.17 × 10⁻¹¹	(23)
18	44866736	rs1019990	1.00 × 10⁻¹¹	(23)
18	55192245	rs8087942	9.70 × 10⁻¹³	(23)
19	13339128	rs10405121	4.74 × 10⁻¹⁰	(23)
19	19406126	rs74182632	1.43 × 10⁻⁸	(23)
19	41864509	rs1982072	4.22 × 10⁻¹¹	(23)
20	10547805	rs616060	4.04 × 10⁻¹⁰	(23)
20	10680221	rs6040095	2.67 × 10⁻¹¹	(30)
20	10684159	rs111404218	2.04 × 10⁻¹⁰	(23)
20	19469817	rs4814864	1.44 × 10⁻²⁸	(23)
20	19503777	rs186083787	4.72 × 10⁻⁸	(23)
20	19513160	rs57069429	8.60 × 10⁻¹²	(23)
20	31168439	rs6057599	8.72 × 10⁻¹⁴	(23)
20	45580290	rs3092262	1.08 × 10⁻⁸	(23)
20	45841052	rs910187	1.14 × 10⁻¹⁰	(23)
21	35593827	rs28451064	3.52 × 10⁻¹⁵	(23)
21	36935896	rs764508	3.28 × 10⁻⁹	(23)
22	20142932	rs625686	8.26 × 10⁻⁹	(23)
22	30232637	rs5763529	2.17 × 10⁻⁹	(30)
22	41627924	rs9611522	1.04 × 10⁻⁸	(30)
X	34102712	rs1507220	2.67 × 10⁻⁸	(23)
X	40746484	rs4403550	3.07 × 10⁻⁹	(23)

Chr, chromosome; Pos, nucleotide position in hg19; SNP, single nucleotide polymorphism.

Clearly, an increase in sample size leads to more migraine loci all of which have a small effect size indicating that none of the identified associated variants, unlike mutations in monogenic HM, are sufficient to bring about disease. As a note of caution, the increase in sample size primarily comes from large population-based datasets, such as UK Biobank and 23andMe (online Supplementary Table 3), that make up the larger part of recent migraine GWAS; 63,990 (63%) of the 102,084 cases in Hautakangas et al. (23) and 47,997 [54%] of the 88,526 cases in Choquet et al. (30). The obvious consequence is a loss of clinical specificity. Not only are cases in these public datasets less specifically defined, the controls are often not necessarily healthy compared to those in a clinic-based cohort, so both cases and controls may be genetically confounded by comorbid conditions that share a genetic aetiology with migraine. Hence, population-based GWAS are poorly suited to identify SNPs associated with specific migraine subtypes. Regardless, including population-based cohorts still renders meaningful findings, already because headache in general is a rather easily definable phenotype. This was exemplified by Meng et al. (31), where a GWAS was performed using the UK Biobank based on a broad definition of headache. Half of the identified 28 associated SNPs were loci that previously had been associated with migraine. So even though the study population likely was rather heterogeneous clinically, there were still known migraine loci present. A possible next step could be to perform more detailed analyses comparing and combining GWAS results for headache and migraine.

Until now, the majority of migraine GWAS have been performed in individuals of European descent, hence the results are not directly transferable to other populations. The first study to investigate migraine in a non-European GWAS concluded that of the three risk variants known at the time, i.e. rs10166942 (TRPM8), rs2651899 (PRDM16), and rs11172113 (LRP1), only the first one showed association with migraine in a Han Chinese population (32,33). In a later Chinese study, only two, i.e. rs4379368 (C7orf10) and rs13208321 (FHL5), of five ‘European’ migraine risk variants showed association with migraine in the She population (34). A subsequent Chinese replication study of 581 migraine cases and 533 ethnically matched controls identified three risk loci rs2274316 (MEF2D), rs6478241 (ASTN2) and rs2651899 (PRDM16) previously identified in European samples (35). In a North Indian population, researchers performed a replication study of five migraine GWAS loci, i.e. rs1835740 (near MTDH), rs11172113 (LRP1), rs2651899 (PRDM16), rs10166942 (TRPM8) and rs7640453 (TGFBR2)), of which only the first three showed association (36) and in another study, only rs2651899 (PRDM16) was associated with the risk for migraine without aura (37).

In addition to replicating European risk loci, the first GWAS have been performed in Asian populations, albeit all with small sample sizes. For instance, a recent GWAS in the Han Chinese population (1042 patients vs 2264 controls) revealed two SNPs, i.e. rs10456100 (KCNK5) and rs7775721 (FHL5), which were significantly associated with migraine without aura (38). A separate study of 715 patients showed the first associations with age of onset of migraine in Han Chinese (39) and a GWAS in 2084 Asian patients with self-reported headache and 11,822 sex- and age-matched controls reported two migraine loci, i.e. rs10493859 (TGFBR3) and rs13312779 (FGF23) (40).

Taken together, it can be concluded that the GWAS era has come of age and already yielded many dozens of low-effect-size DNA variants; very many of which replicate in other GWAS, hence providing overwhelming proof of the robustness of the findings. Still, larger studies, also in non-European populations, have to be performed to identify additional loci and establish to what extent associated variants play a role in migraine patients across the globe. As a proof of concept, it is interesting that some SNPs point to already identified treatment targets such as CGRP and 5-HT1F. This clearly shows the potential of genetic studies to identify treatment targets, which is dearly needed as many patients are unresponsive to current treatments. Studying specific phenotypes, such as migraine with aura or chronic migraine, remains challenging as in most GWAS data only limited data are available on subtypes of migraine and, to certain extent, their characteristics.

Another important challenging aspect of GWAS that remains is to determine the exact gene that is affected by a risk SNP, as this is far from straightforward. Major complications are that the associated SNPs are typically located in intronic or intergenic regions and are merely genomic markers at a distance, so in LD with the true causal variant. An often-used method is to link a risk index SNP (i.e. the SNP with the lowest p-value) to the “nearest gene”. However, it was shown that in two-thirds of cases it is not the nearest gene that is affected by the risk SNP (41,42). Notably, there can be multiple independent association signals at the same locus (secondary SNPs) that can influence other regulatory features of the same (or nearby) gene. In the 2016 migraine GWAS (18), of the 38 distinct genomic loci, six contained such a secondary SNP.

Risk SNPs in common diseases rarely affect amino acid changes and most likely indirectly regulate gene expression, for instance by disrupting enhancer elements. Of note, the causal gene can be located even hundreds of kilobases away from the location of the risk SNP. Hence, a key challenge for the future is to effectively integrate data from GWAS and expression quantitative trait loci (eQTL) analyses (and other omics data) to prioritise likely causal genes, which is needed for meaningful functional follow-up of GWAS findings (43). The magnitude of the challenge is well-demonstrated by the functional follow-up of intronic SNP rs9349379 near PHACTR1 that surfaced not only in migraine GWAS but also in multiple GWAS of vascular diseases (e.g. coronary artery disease, fibromuscular dysplasia, hypertension and cervical artery dissection) (44). Detailed functional follow-up indicated that this SNP affects not PHACTR1, but endothelin-1 (ET-1; EDN1), a strong vasoconstrictor that acts on smooth muscle cells and is located no less than 600 kb upstream of the risk SNP (44). Still, the debate on which gene is affected rs9349379 has not been settled. More recent data have revived PHACTR1 as the causal gene based on the changed expression of PHACTR1, and not EDN1, in cells that carry the risk allele and functional human and mouse studies of PHACTR1/Phactr1 revealing a role, in arterial compliance—and not measures of dynamic arterial function—across multiple vascular beds (45).

Studies like these indicate that the road ahead to understanding the pathophysiology behind migraine risk loci is far from easy. Similar challenges apply to other disease areas (46). Clearly, we are only at the beginning of performing “wet lab” functional follow-ups of GWAS risk SNPs, which will likely require a plethora of animal- and/or cell-based assays depending on the variant/gene that is affected (47).

What about candidate gene association studies?

Before the advent of high-throughput low-cost genotyping, so-called candidate gene association studies (CGAS) were performed. These tested for association with migraine one or a few SNPs in genes from neurological, vascular, hormonal, and mitochondrial pathways that had been selected because of their perceived involvement in migraine pathophysiology. Unfortunately, findings in CGAS almost exclusively came from too small sample sets and were not robustly replicated, so the risk of a false-negative or false-positive result is very high. Details of CGAS findings can be found in other reviews (48,49). In fact, a study by de Vries et al. (50) investigated whether 27 genes from both CCAS and other genetic studies showed an association with migraine in IHGC GWAS data from a large sample of 5175 clinic-based migraine patients with and without aura (and 13,972 controls) and found none to be associated with migraine. A genome-wide significant result was found for one of the 27 genes, namely the CACNA1A gene, in the Hautankagas et al. paper (23), but gene-based analyses did not show significant association for any of the other 26 genes (data not shown), even though the sample size had considerably increased. Because of the possibility of a straightforward “look-up” of genes of interest, i.e. by gene-based analysis testing, GWAS can replace GCAS studies. One might still consider wet lab testing of specific SNPs and/or investigating specific study populations, such as response to certain medications, as long as one takes into account sample sizes and replication of promising findings.

Next-generation sequencing in migraine

A large part of the genetic variance and heritability in common diseases cannot be explained (usually referred to as “missing heritability”) with a GWAS approach alone. One reason is that rarer variants (MAF < 0.01), potentially with larger effect sizes, are not well interrogated by genotyping arrays typically utilised in the GWAS approach. Such mediate-effect-size variants can be identified using a next-generation sequencing (NGS) approach, i.e. by the simultaneous large-scale sequencing of the coding exons (whole-exome sequencing; WES) or the entire genome (whole-genome sequencing; WGS). In addition, the simultaneous sequencing of RNA transcripts (“transcriptome”; RNA-seq), either of bulk tissue or of its single nuclei can shed light on molecular mechanisms.

Exome and genome sequencing in migraine

Only a few NGS studies have been performed in migraine thus far. Until now, WES was typically applied to cohorts of patients with HM, testing several hundred cases in an attempt to either find causal mutations in known HM genes or novel HM genes patients that are negative for mutations in CACNA1A, ATP1A2, and SCN1A. Until now results have not led to additional (undisputed) HM genes (9 –11,51,52). This may indicate that HM in mutation-negative patients may be oligogenic or polygenic, in line with the excess presence of common variants in such patients (12). A few recent studies in the field of migraine used NGS for a different approach. First, SNP genotyping followed by NGS was used for linkage, haplotype, and variant analyses within a single large Finnish migraine-epilepsy family and found an association between the epilepsy phenotype and the NCOR2 gene that colocalises with one of the migraine risk loci (53). Second, in a WES of 16 individuals, who developed numerous neurological- and concussion-related symptoms following minor head injuries of which seven had developed migraines following the injury, revealed possible mutations in various ion channel, neurotransmitter, and ubiquitin-related genes, but causality of any of them needs to be established (51). Third, in an attempt to investigate whether certain genetic variants determine treatment response to verapamil prophylaxis, WES was applied to a set of 21 definitive responders and 14 definitive non-responders (discovery cohort) and promising data were further analysed in 185 verapamil-treated patients (replication cohort) to yield 39 ‘possible variants’ (54). Subsequent bioinformatics analyses revealed the possible involvement of myo-inositol biosynthetic and phospholipase-C second messenger pathways in verapamil responsiveness (54).

The added value of the WGS approach, when combined with GWAS data, was shown in a recent Danish study of extended migraine pedigrees (encompassing 1040 individuals from 155 families) (55). An excess of rare segregating variants in regulatory regions (one CpG island and three polycomb group response elements) was identified near, but independent of, four migraine GWAS risk loci with replication in an independent case-control cohort (encompassing 2027 migraineurs and 1650 controls). Another study of the same Danish group combined RNA sequencing data from brain and vascular tissues relevant to migraine with WGS data of over one hundred migraine families in a ‘systems genetics approach’ and identified a “visual cortex module” with rare functional gene variants implicated in migraine pathophysiology (56). Their approach suggested the involvement of glutamate, serotonin, G-protein signalling pathways and hormonal pathways in migraine pathophysiology.

Gene expression analysis in migraine

An alternative approach to understanding molecular mechanisms involved in migraine pathophysiology is to study gene expression profiles. Contrary to genetic variation, gene expression is not fixed through life and expression is driven by both genetic and environmental factors (57). Typically, an RNA-seq approach (i.e. simultaneous sequencing of coding (messenger) and non-coding RNAs in a sample) is for instance used to identify differences in expression between individuals with and without migraine or over the course of a migraine attack. Various RNA-seq studies have been performed in migraine, but the results are not unambiguous not in the least because of potential caveats of using peripheral blood, the main source of biomaterial for such studies in the case of migraine (58). First, an Australian genome-wide expression study comparing whole blood expression profiles of 83 migraine cases and 83 non-migraine controls revealed that immune-inflammatory pathways, including multiple pathways expressed in microglial cells, seem to play a role in migraine (59). Second, RNA-seq-based expression profiles in the blood of 24 Hungarian migraine patients, collected during both the headache and the headache-free (interictal) period, were compared with that of 13 sex- and age-matched healthy controls (60). Some 144 differentially expressed genes (DEGs) were identified when comparing the headache and headache-free samples in patients and 163 DEGs were identified between the interictal and control samples. Subsequent network analysis, revealed enrichment of inflammation, like in the study mentioned above, cytokine activity and mitochondrial dysfunction pathways between patients (both headache and headache-free) and controls. On the contrary, no overt differences in RNA-seq expression profiles were observed when comparing interictal blood samples of 26 Danish migraine patients with those of 20 age- and sex-matched controls (61). However, a comparison of gene expression profiles of blood samples from 27 Danish migraine patients taken during the attack, two hours after treatment, and on a headache-free day, revealed 29 DEGs between “attack” and “after treatment” (62). Among the results of their gene network analysis appeared the “5HT1-type receptor mediated signalling pathway”, which is not that surprising given the sumatriptan treatment patients had received. Finally, a study by LaPaglia et al. (63) used RNA-seq to assess gene expression in sub-regions of 16 human post-mortem trigeminal ganglia, as to establish the human transcriptome from a brain region highly relevant to migraine pathophysiology, already because trigeminal neurons release neuropeptides such as CGRP (64).

Single-cell genomics in migraine

Single-cell genomics, i.e. RNA sequencing of single cells (scRNAseq)/single nuclei (snRNAseq), provides unique opportunities to elucidate in greater detail not only which genes but also in which cell types migraine-associated genes are expressed. The latter is especially relevant to better understand how migraine susceptibility genes, for instance from recent migraine GWAS, may affect migraine pathophysiology.

A major advancement in single-cell sequencing in migraine came from a study by Renthal (65), which used single-brain cell RNA sequencing data from healthy human cortex cells (astrocytes, oligodendrocytes, neurons, microglia and endothelial cells) and revealed that over 40% of migraine-associated genes showed a preference for a specific brain cell type. Strikingly, some 70% of the neuronal migraine-associated genes were significantly enriched in inhibitory neurons compared to 30% in excitatory neurons (65). In a follow-up scRNAseq mouse study encompassing more cells (∼500,000 vs ∼3000) and cell types across more nervous system structures implicated in migraine, it was shown that expression of 52% (28/54) of the investigated migraine-associated genes was enriched in both the central (CNS) and peripheral (PNS) nervous system, 20% was enriched in the CNS, PNS and neurovascular cell types, and some 17% showed tissue-selective enhancement (66). The study indicated that most migraine-associated genes are widely expressed, with functional changes that likely affect multiple cell types. A study like this i) may guide future mechanistic studies into the function of migraine-associated genes and ii) holds the promise that it may be feasible to stratify patients based on a predominant CNS or PNS cell type polygenic migraine risk. Finally, the same group created a transcriptional and epigenomic cell atlas of the mouse and human trigeminal ganglion (TG) at single-cell resolution and revealed species-specific gene expression and evolutionarily conserved in distinct TG cell types (67). The TG atlas will be an enormous asset to migraine research given the relevance of TG neuron sensitisation in migraine, while the relative contribution of individual neuronal and non-neuronal TG subtypes still is unknown. It is to be expected that the TG atlas will also guide the search for cell types most relevant to migraine-associated variants. Of note, the expression of CALCA, CALCB, and HTR1F, which are among the targets of current migraine treatments (23), as well as various other genes implicated by GWAS, exhibit species-specific and cell-type-specific expression patterns. To assess the TG cell types involved in migraine, two experimental mouse migraine models, i.e. the inflammatory soup (IS) model (relevant for activating and sensitizing trigeminal meningeal nociceptors) and the cortical spreading depolarization (CSD) model (the neurophysiological correlate of the migraine aura) were used to identify activated TG cells, based on the expression of immediate early genes. Using IS as a trigger some 5% of TG cells were activated, whereas the number was slightly lower (3.5%) when the less profound trigger CSD was induced. It is reassuring that at least some convergence was present at the TG cell type activation level between both experimental models. For instance, a comparison of activated vs non-activated nuclei revealed 96 upregulated genes in the IS model and 72 genes in the CSD model, and within both models, genes were involved with axon guidance and inflammation and several were near to or predicted to be influenced by migraine-associated GWAS variants. Of note, the study also has clear epigenomic implications but these are discussed in the parallel review in this issue by Gallardo et al. (68).

Genetic relation with comorbid disorders

Migraine is associated with a high prevalence of stroke and psychiatric comorbidities, such as anxiety, depression, and post-traumatic stress disorder (69 –72). Although the exact relationship remains elusive, twin and family studies have revealed a strong heritable factor in migraine and various other brain disorders (15). Comorbidity between disorders can be due to a causal relationship and/or may be due to shared genetic and/or environmental factors. Epidemiological comorbidity and sharing of symptoms can stem from aetiologic overlap. To investigate shared genetic architecture between migraine and other traits correlation analyses using GWAS data can be performed.

Genetic correlation with psychiatric disorders

In correlation studies, the proportion of variance two traits share due to genetics, defined as the genetic correlation r_G, is measured. A recent example that shed important light on the genetic architecture underlying migraine comorbidities came from the Brainstorm Consortium in which correlations were calculated to quantify the degree of genetic overlap between 25 brain disorders using the GWAS data of over 1 million individuals (265,218 patients and 784,643 controls) (73). A major finding of the study was that psychiatric disorders seem to share more common genetic risk variants, whereas neurological disorders seemed more distinct from each other and from psychiatric disorders. The notable exception was migraine, which was significantly correlated to major depressive disorder (MDD), attention deficit hyperactivity disorder (ADHD), and Tourette syndrome (73).

The genetic correlation of migraine with MDD was confirmed in a separate study that showed a significant genetic correlation between the two disorders (r_G = 0.25; P = 0.04) (74). In that study, a meta-analysis of the combined GWAS summary statistics of migraine and MDD revealed three novel risk loci associated with both disorders, and a subsequent gene-based association analysis suggested the involvement of ANKDD1B and KCNK5 (74). More recently, it was shown that most migraine-associated variants also influence schizophrenia and depression, i.e. 36 loci were found to be associated with migraine and schizophrenia and 14 loci were associated with both migraine and depression (75).

Genetic correlation between migraine and other disorders

Several disorders are comorbid with migraine and show at least some genetic overlap. First, epidemiological studies have demonstrated repeatedly an increased risk of ischemic stroke in migraine patients, particularly for those with migraine with aura (76,77). Evidence for an overlap at the genetic level between migraine and ischemic stroke was shown by Malik et al. (20). Second, migraine and endometriosis share epidemiological characteristics and risk factors, such as female dominant prevalence and the involvement of the menstrual cycle (78). Although co-occurrence of migraine and endometriosis already suggested that both traits may also be genetically correlated (79), only recently it was demonstrated that there is a significant overlapping genetic architecture (r_G = 0.38, P = 2.30 × 10⁻²⁵) between endometriosis and migraine, with evidence for the involvement of various genes, among which TRIM32 and SLC35G6 (80). Third, sleep disturbances and sleep-related disorders have also been reported in migraine patients (81). One study showed a genetic overlap of seven sleep traits with migraine. Insomnia was the most significantly correlated with migraine (r_G = 0.29, P = 1.87 × 10⁻³²); weaker correlations were found for short sleep duration (r_G = 0.18, P = 1.69 × 10⁻⁹), difficulty awakening (r_G = 0.11, P = 2.02 × 10⁻⁵), daytime napping (r_G = 0.11, P = 1.31 × 10⁻⁵), daytime sleepiness (r_G = 0.09, P = 1.21 × 10⁻⁴), long sleep duration (r_G = 0.12, P = 7.60 × 10⁻⁴), and sleep duration in general (r_G = −0.08, P = 1.56 × 10⁻³) (82). Fourth, fibromuscular dysplasia was found positively correlated with migraine (r_G = 0.28, P = 8 × 10⁻⁴) (83). Finally, a broad cross-trait correlation analysis of migraine reported genetic correlations with various disorders, including heart disease, type 2 diabetes, as well as various autoimmune and psychiatric phenotypes (84).

Genetic correlation between migraine and cluster headache

Migraine and cluster headache both share pathophysiological features, both show involvement of the trigeminovascular system, and both can be effectively treated with CGRP monoclonal antibodies and triptans (85). Recently, two parallel GWAS papers on cluster headache showed genetic overlap between migraine and cluster headache (86,87). Of note, a co-localization analysis showed that one of the top cluster headache loci (UFL1/FHL5) was the same as in migraine (87), which seems to reflect a partially shared genetic architecture underlying both migraine and cluster headache.

Genetic correlation between migraine and metabolites

In clinical practice, serum levels of chemical compounds are often utilised for diagnosing disease and monitoring the body’s function. As migraine is associated with an increased risk of cardiovascular disease (CVD) and stroke (77,88), routine chemistry tests and serum markers for these traits might be associated also with migraine. A recent study by Tanha et al. (89) shows that shared genetic factors between migraine risk and markers for CVD risk, iron deficiency and liver dysfunction exist. The shared genetic architecture implies an alteration of metabolite concentrations in individuals with migraine. Moreover, an additional study investigating the genetic underpinning and causality of the 316 blood metabolites and migraine risk revealed a significant correlation between migraine and 44 metabolites, largely organic acid and lipid metabolic traits (90). Changes in metabolite levels are in line with recent metabolomics studies in the blood of migraine patients (91,92).

Causality in migraine

Clear pleiotropy exists across many traits and one way of entangling this is to use Mendelian randomisation (MR). In an MR analysis, genetic variants associated with an exposure are identified and regressed upon an outcome measurement to infer causality (i.e. direction) of the association. Given the random assortment of alleles at gametogenesis in early life, this method is less likely to suffer from issues of confounding and reverse causation than methods used in conventional observational epidemiological studies (93). For a successful MR analysis, three assumptions need to be fulfilled (94). i) Variants used as instrumental variables (IVs) need to be associated with the exposure. ii) The IVs only affect the outcome through the exposure, not through any other causal pathway. Factors that may lead to violation of this assumption include population stratification, LD and horizontal pleiotropy, the latter means that there is an (in)direct independent association of the IV (or another SNP in LD with the IV) with another trait that is not in the causal pathway of the investigated relation. iii) The IVs must not be associated with confounders. Essentially, MR studies are still in their infancy, hence it is important and necessary to replicate the results of different studies to reach robust conclusions. In one-directional MR the possible causal relation between trait X on trait Y is investigated, in bidirectional MR studies the directional effect from trait Y on trait X is also investigated. An increasing number of MR studies related to migraine have already been performed (Table 2).

Table 2.

Papers describing a Mendelian randomisation analysis with migraine.

Title of paper	Other Trait(s)	Type of MR	Any causality found	Reference
Migraine and ischemic stroke: A Mendelian randomization study	Ischemic Stroke	One-directional	No	(96)
Association between hemostatic profile and migraine: A Mendelian randomization analysis	Twelve blood-based measures of hemostasis	Bidirectional	Yes	(111)
Habitual sleep disturbances and migraine: a Mendelian randomization study	All sleep traits ascertained in UKB: sleep duration, morning diurnal preference (also referred to as “chronotype”), daytime napping frequency, snoring, insomnia symptoms, difficulty awakening, and daytime sleepiness	Bidirectional	Yes	(82)
Irritable bowel syndrome and migraine: evidence from Mendelian randomization analysis in the UK Biobank	Irritable bowel syndrome	One-directional	Yes	(112)
Association Between Insomnia and Migraine Risk: A Case-Control and Bidirectional Mendelian Randomization Study.	Insomnia	Bidirectional	Yes	(100)
No causal association Between coffee consumption and risk of migraine: A Mendelian randomization Study	Coffee consumption	One-directional	No	(113)
Alcohol, coffee consumption, and smoking in relation to migraine: a bidirectional Mendelian randomization study	Alcohol, coffee consumption, and smoking	Bidirectional	Yes	(101)
Serum calcium and risk of migraine: a Mendelian randomization study.	Serum calcium	One directional	Yes	(114)
Phenotypic and genotypic associations between migraine and lipoprotein subfractions	Nineteen lipoprotein subfractions	Bidirectional	No	(115)
Effect of genetic liability to migraine on cognition and brain volume: A Mendelian randomization study	Cognition and brain volume	One-directional	No	(104)
Effect of genetic liability to migraine on coronary artery disease and atrial fibrillation: a Mendelian randomization study	Coronary artery disease and atrial fibrillation	One-directional	Yes	(99)
Genetic overlap and causality between blood metabolites and migraine	Seventy-nine unique metabolic traits	Bidirectional	Yes	(90)
Association and genetic overlap between clinical chemistry tests and migraine	Clinical chemistry tests (HDL-C, LDL-C, TG, heart/cardiac problems, IRON, TC, iron deficiency anaemia, TB, GGT, Disease of liver)	One-directional	No	(89)
Exploring the causal inference of migraine on stroke: A Mendelian randomization study	Stroke, ischemic stroke, and haemorrhagic stroke	One-directional	No	(95)
Shared molecular genetic mechanisms underlie endometriosis and migraine comorbidity.	Endometriosis	One-directional	No	(80)
Migraine, stroke, and cervical arterial dissection: Shared genetics for a triad of brain disorders with vascular involvement	Stroke and cervical arterial dissection	One-directional	Yes	(97)
Common health conditions in childhood and adolescence, school absence, and educational attainment: Mendelian randomization study	Educational attainment (years of schooling)	One-directional	No	(116)
The causal effects of health conditions and risk factors on social and socioeconomic outcomes: Mendelian randomization in UK Biobank.	Twenty-six social and socioeconomic outcomes	One-directional	Yes	(117)
Cross-trait analyses with migraine reveal widespread pleiotropy and suggest a vascular component to migraine headache	Fourthy-seven traits from the UK Biobank (including cardiovascular disease, blood pressure, cholesterol, blood pressure, neuroticism, asthma, autoimmune disease, education, white blood cell count, platelet count and smoking status)	One-directional	Yes	(84)
The causal role of smoking on the risk of headache. A Mendelian randomization analysis in the HUNT study	Smoking	One-directional (one sample)	No	(102)
A genome-wide cross-phenotype meta-analysis of the association of blood pressure with migraine	Blood pressure	Bidirectional	Yes	(98)
Pleiotropic effects of trait-associated genetic variation on DNA methylation: utility for refining GWAS loci	DNA Methylation	One-directional	Yes	(118)

Causality between migraine and different traits

Of the many published MR studies that examine the relationship between migraine and a given disorder/trait, only some are discussed below. First, the causality between migraine and stroke has been investigated in three studies, two of which do not support a causal relationship between migraine and stroke (95 –97), whereas one found a protective influence of large artery stroke (OR [95% CI] = 0.86 [0.76–0.96], P = 0.0067) on migraine, but no relation with stroke subtypes (97). Second, several other cardiovascular traits/risk factors and their possible causal relationship with migraine have been investigated, one showing that an increase in blood pressure directly contributes to migraine risk (84). A similar result was observed by Guo et al. (98) where an increase in both diastolic blood pressure (OR [95% CI] = 1.20 [1.15–1.25]/10 mmHg; P = 5.57 × 10⁻²⁵) and systolic blood pressure (1.05 [1.03–1.07]/10 mmHg; P = 2.60 × 10⁻⁷) increased migraine susceptibility. In addition, liability to migraine leads to an increased coronary artery disease risk (99). Third, two studies explored the relationship with sleep-related comorbidities and provided evidence for a significant causal effect of difficulty awakening (OR [95% CI] = 1.37 [1.12–1.68], P = 0.002) (81) and insomnia (OR [95% CI] = 1.24 [1.11–1.38], P = 1.01 × 10⁻⁴) (100) on the risk of migraine. Fourth, various MRs exploring behavioural traits, such as smoking, alcohol and coffee consumption, have shown that a higher alcohol or coffee consumption leads to a reduced migraine risk (101). Although initially no causal relationship was found between smoking and migraine (102), the bidirectional MR study by Yuan et al. (101) showed that a decreased prevalence of smoking initiation led to reduced migraine risk. However, the reverse hypothesis that migraine leads to a reduced smoking initiation, was not found in this study (101). Although from a clinical perspective that might have been expected. Finally, despite the above-mentioned genetic correlation between migraine and endometriosis, there was no evidence for a causal relationship between both disorders (80).

In addition to hypothesis-driven MR studies, a recent hypothesis-generating phenome-wide analysis examined 1504 phenotypes and identified 17 potential causal relationships with self-reported migraine and traits related to e.g. stressful life events, vasoconstriction, platelet clumping, and occupational exposure to paints, glues or thinner, and diesel exhaust (103). Studies are also starting to emerge that examine other health-related markers such as brain volume (104), blood metabolites (90), and blood proteins (105). In addition to the caution that replication of positive MR findings is urgently needed, it is also important to note, that the power of a MR study is dependent on the size of the original GWAS. This means that when no causality is found at this moment, it can be causal in the future when there are more loci found.

Polygenic risk score in migraine

One way of making better use of the large number of small effect variants identified in migraine GWAS to have clinical benefit is the calculation of polygenic risk scores (PRSs). A PRS is the combined effect of many common risk variants of genetic load for the discovery trait that can be used to estimate risk for a certain trait/phenotype in individuals in a target sample (106). This is done by testing whether a higher PRS based on the discovery sample is associated with case status or a specific trait in the target sample via regression models. A PRS provides a promising possibility to investigate the shared genetic architecture between migraine with known and hitherto unknown co-morbidities or traits.

The aggregation of migraine in families and the earlier age of onset of migraine can to some extent be contributed to common polygenic variations, where the PRS explained a larger part of the phenotype variance in familial cases, especially those with migraine with aura and hemiplegic migraine compared to population cases (12). Other PRS research shows a correlation between the current ICHD-3 criteria and the PRS of migraine, where the migraine phenotype progressively follows the polygenic burden (107). A migraine PRS is also able to identify patient groups more likely to respond to triptans (108). On the other hand, no association between the migraine PRS and individual cerebral blood flow (CBF) measurements was found, although one individual SNP (rs67338227) was associated with CBF measures (109). An inverse association was seen between the coronary artery PRS and migraine headaches (OR [95% CI]: 0.94 [0.93–0.96]), where the signal was due to a large number of genetic variants with opposing effects on the two diseases (110). The continued discovery of new risk loci for common migraine based on ever-growing size GWAS will only further improve risk prediction by PRS.

Future of omics

GWAS is considered part of the most mature omics field, namely genomics and as presented in this review its data can be utilised to assess the correlation between different traits, assess causality and calculate PRS. However, multiple additional omics fields have arisen (e.g. epigenomics, transcriptomics, metabolomics, and proteomics). Integration of all different types of omics data will further our understanding of the underlying biology of migraine.

Key findings

There is an overlap of 68 genetic loci in between the two latest GWAS studies in migraine.

Based on these studies there are in total 181 independent SNPs associated with migraine, of which 3 are female specific.

There are multiple genetic methods to characterise genetic risk factors and architecture of migraine.

Supplemental Material

sj-xlsx-1-cep-10.1177_03331024221145962 - Supplemental material for Migraine genetics: Status and road forward

Supplemental material, sj-xlsx-1-cep-10.1177_03331024221145962 for Migraine genetics: Status and road forward by Aster VE Harder, Gisela M. Terwindt, Dale R Nyholt and Arn MJM van den Maagdenberg in Cephalalgia

Supplemental Material

sj-xlsx-2-cep-10.1177_03331024221145962 - Supplemental material for Migraine genetics: Status and road forward

Supplemental material, sj-xlsx-2-cep-10.1177_03331024221145962 for Migraine genetics: Status and road forward by Aster VE Harder, Gisela M. Terwindt, Dale R Nyholt and Arn MJM van den Maagdenberg in Cephalalgia

Supplemental Material

sj-xlsx-3-cep-10.1177_03331024221145962 - Supplemental material for Migraine genetics: Status and road forward

Supplemental material, sj-xlsx-3-cep-10.1177_03331024221145962 for Migraine genetics: Status and road forward by Aster VE Harder, Gisela M. Terwindt, Dale R Nyholt and Arn MJM van den Maagdenberg in Cephalalgia

Footnotes

Declaration of conflicting interests

The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: GMT reports consultancy or industry support from Novartis, Lilly and Teva, Abbvie, and Lundbeck and independent support from the Dutch Heart and Brain Foundations, IRRF, Dioraphte and the Dutch Research Council. AMJMvdM. reports research support from Praxis Precision Medicine and Schedule 1 Therapeutics and consultancy support from AbbVie. All other authors declare no conflict of interests.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Aster VE Harder

References

Russell

Olesen

Increased familial risk and evidence of genetic factor in migraine. BMJ 1995; 311: 541–544.

Stewart

Bigal

Kolodner

, et al. Familial risk of migraine: variation by proband age at onset and headache severity. Neurology 2006; 66: 344–348.

Ferrari

Goadsby

Burstein

, et al. Migraine. Nat Rev Dis Primers 2022; 8: 2.

Ophoff

Terwindt

Vergouwe

, et al. Familial hemiplegic migraine and episodic ataxia type-2 are caused by mutations in the Ca2+ channel gene CACNL1A4. Cell 1996; 87: 543–552.

Dichgans

Freilinger

Eckstein

, et al. Mutation in the neuronal voltage-gated sodium channel SCN1A in familial hemiplegic migraine. Lancet 2005; 366: 371–377.

De Fusco

Marconi

Silvestri

, et al. Haploinsufficiency of ATP1A2 encoding the Na+/K+ pump alpha2 subunit associated with familial hemiplegic migraine type 2. Nat Genet 2003; 33: 192–196.

Ferrari

Klever

Terwindt

, et al. Migraine pathophysiology: lessons from mouse models and human genetics. Lancet Neurol 2015; 14: 65–80.

Pietrobon

Brennan

KC.

Genetic mouse models of migraine. J Headache Pain 2019; 20: 79.

Hiekkala

Vuola

Artto

, et al. The contribution of CACNA1A, ATP1A2 and SCN1A mutations in hemiplegic migraine: A clinical and genetic study in Finnish migraine families. Cephalalgia 2018; 38: 1849–1863.

10.

Sutherland

Maksemous

Albury

, et al. Comprehensive exonic sequencing of hemiplegic migraine-related genes in a cohort of suspected probands identifies known and potential pathogenic variants. Cells 2020; 9.

11.

Pelzer

Haan

Stam

, et al. Clinical spectrum of hemiplegic migraine and chances of finding a pathogenic mutation. Neurology 2018; 90: e575–e582.

12.

Gormley

Kurki

Hiekkala

, et al. Common variant burden contributes to the familial aggregation of migraine in 1,589 families. Neuron 2018; 99: 1098.

13.

Pelzer

de Vries

Kamphorst

, et al. PRRT2 and hemiplegic migraine: a complex association. Neurology 2014; 83: 288–290.

14.

Rasmussen

Olofsson

Chalmer

, et al. Higher burden of rare frameshift indels in genes related to synaptic transmission separate familial hemiplegic migraine from common types of migraine. J Med Genet 2020; 57: 610–616.

15.

Polderman

Benyamin

de Leeuw

, et al. Meta-analysis of the heritability of human traits based on fifty years of twin studies. Nat Genet 2015; 47: 702–709.

16.

Pelzer

Louter

van Zwet

, et al. Linking migraine frequency with family history of migraine. Cephalalgia 2018: 333102418783295.

17.

Ulrich

Gervil

Kyvik

, et al. Evidence of a genetic factor in migraine with aura: a population-based Danish twin study. Ann Neurol 1999; 45: 242–246.

18.

Gormley

Anttila

Winsvold

, et al. Meta-analysis of 375,000 individuals identifies 38 susceptibility loci for migraine. Nat Genet 2016; 48: 856-+.

19.

Winsvold

Nelson

Malik

, et al. Genetic analysis for a shared biological basis between migraine and coronary artery disease. Neurol Genet 2015; 1: e10.

20.

Malik

Freilinger

Winsvold

, et al. Shared genetic basis for migraine and ischemic stroke: A genome-wide analysis of common variants. Neurology 2015; 84: 2132–2145.

21.

Finucane

Reshef

Anttila

, et al. Heritability enrichment of specifically expressed genes identifies disease-relevant tissues and cell types. Nat Genet 2018; 50: 621–629.

22.

Nyholt

Borsook

Griffiths

LR.

Migrainomics – identifying brain and genetic markers of migraine. Nat Rev Neurol 2017; 13: 725–741.

23.

Hautakangas

Winsvold

Ruotsalainen

, et al. Genome-wide analysis of 102,084 migraine cases identifies 123 risk loci and subtype-specific risk alleles. Nat Genet 2022.

24.

van Dongen

Zielman

Noga

, et al. Migraine biomarkers in cerebrospinal fluid: A systematic review and meta-analysis. Cephalalgia 2017; 37: 49–63.

25.

Juhasz

Zsombok

Modos

, et al. NO-induced migraine attack: strong increase in plasma calcitonin gene-related peptide (CGRP) concentration and negative correlation with platelet serotonin release. Pain 2003; 106: 461–470.

26.

Diener

H-C.

CGRP antibodies for migraine prevention — new kids on the block. Nature Reviews Neurology 2019; 15: 129–130.

27.

de Vries

Villalón

MaassenVanDenBrink

Pharmacological treatment of migraine: CGRP and 5-HT beyond the triptans. Pharmacol Ther 2020; 211: 107528.

28.

Nyholt

Gillespie

Heath

, et al. Latent class and genetic analysis does not support migraine with aura and migraine without aura as separate entities. Genet Epidemiol 2004; 26: 231–244.

29.

Ligthart

Boomsma

Martin

, et al. Migraine with aura and migraine without aura are not distinct entities: further evidence from a large Dutch population study. Twin Res Human Gen 2006; 9: 54–63.

30.

Choquet

Yin

Jacobson

, et al. New and sex-specific migraine susceptibility loci identified from a multiethnic genome-wide meta-analysis. Commun Biol 2021; 4: 864.

31.

Meng

Adams

Hebert

, et al. A genome-wide association study finds genetic associations with broadly-defined headache in UK biobank (N = 223,773). EBioMedicine 2018; 28: 180–186.

32.

Lin

, et al. PRDM16 rs2651899 variant is a risk factor for Chinese common migraine patients. Headache 2013; 53: 1595–1601.

33.

Fan

Wang

Fan

, et al. Replication of migraine GWAS susceptibility loci in Chinese Han population. Headache 2014; 54: 709–715.

34.

Lin

Yao

, et al. Association of genetic loci for migraine susceptibility in the she people of China. J Headache Pain 2015; 16: 553.

35.

Fang

, et al. Multilocus analysis reveals three candidate genes for Chinese migraine susceptibility. Clin Genet 2017; 92: 143–149.

36.

Ghosh

Pradhan

Mittal

Genome-wide-associated variants in migraine susceptibility: a replication study from North India. Headache 2013; 53: 1583–1594.

37.

Kaur

Ali

Ahmad

, et al. rs2651899 variant is associated with risk for migraine without aura from North Indian population. Mol Biol Rep 2019; 46: 1247–1255.

38.

Jiang

Zhao

Zhang

, et al. Common variants in KCNK5 and FHL5 genes contributed to the susceptibility of migraine without aura in Han Chinese population. Sci Rep 2021; 11: 6807.

39.

Tsai

Liang

Lin

, et al. Identifying genetic variants for age of migraine onset in a Han Chinese population in Taiwan. J Headache Pain 2021; 22: 89.

40.

Tsao

Wang

Hsu

, et al. Genome-wide association study reveals susceptibility loci for self-reported headache in a large community-based Asian population. Cephalalgia 2021: 3331024211037269.

41.

Gusev

Shi

, et al. Integrative approaches for large-scale transcriptome-wide association studies. Nat Genet 2016; 48: 245–252.

42.

Zhu

Zhang

, et al. Integration of summary data from GWAS and eQTL studies predicts complex trait gene targets. Nat Genet 2016; 48: 481–487.

43.

Visscher

Wray

Zhang

, et al. 10 Years of GWAS discovery: biology, function, and translation. Am J Hum Genet 2017; 101: 5–22.

44.

Gupta

Hadaya

Trehan

, et al. A genetic variant associated with five vascular diseases is a distal regulator of endothelin-1 gene expression. Cell 2017; 170: 522–533.e515.

45.

Wood

Antonopoulos

Chuaiphichai

, et al. PHACTR1 modulates vascular compliance but not endothelial function: a translational study. Cardiovasc Res 2022. doi: 10.1093/cvr/cvac092.

46.

Lappalainen

MacArthur

DG.

From variant to function in human disease genetics. Science 2021; 373: 1464–1468.

47.

Dourlen

Chapuis

Lambert

JC.

Using high-throughput animal or cell-based models to functionally characterize GWAS signals. Curr Genet Med Rep 2018; 6: 107–115.

48.

Gasparini

Sutherland

Griffiths

LR.

Studies on the pathophysiology and genetic basis of migraine. Curr Genomics 2013; 14: 300–315.

49.

Sutherland

Albury

Griffiths

LR.

Advances in genetics of migraine. J Headache Pain 2019; 20: 72.

50.

de Vries

Anttila

Freilinger

, et al. Systematic re-evaluation of genes from candidate gene association studies in migraine using a large genome-wide association data set. Cephalalgia 2016; 36: 604–614.

51.

Ibrahim

Sutherland

Maksemous

, et al. Exploring neuronal vulnerability to head trauma using a whole exome approach. J Neurotrauma 2020; 37: 1870–1879.

52.

Gil-Perotín

Jaijo

Verdú

, et al. Epilepsy, status epilepticus, and hemiplegic migraine coexisting with a novel SLC4A4 mutation. Neurol Sci 2021; 42: 3647–3654.

53.

Nuottamo

Häppölä

Artto

, et al. NCOR2 is a novel candidate gene for migraine-epilepsy phenotype. Cephalalgia 2022; 42: 631–644.

54.

Cutrer

Moyer

Atkinson

, et al. Genetic variants related to successful migraine prophylaxis with verapamil. Mol Genet Genomic Med 2021; 9: e1680.

55.

Techlo

Rasmussen

Møller

, et al. Familial analysis reveals rare risk variants for migraine in regulatory regions. Neurogenetics 2020; 21: 149–157.

56.

Rasmussen

Kogelman

LJA

Kristensen

, et al. Functional gene networks reveal distinct mechanisms segregating in migraine families. Brain 2020; 143: 2945–2956.

57.

Bar-Joseph

Gitter

Simon

Studying and modelling dynamic biological processes using time-series gene expression data. Nat Rev Gen 2012; 13: 552–564.

58.

Gerring

Rodriguez-Acevedo

Powell

, et al. Blood gene expression studies in migraine: Potential and caveats. Cephalalgia 2016; 36: 669–678.

59.

Gerring

Powell

Montgomery

, et al. Genome-wide analysis of blood gene expression in migraine implicates immune-inflammatory pathways. Cephalalgia 2018; 38: 292–303.

60.

Aczél

Körtési

Kun

, et al. Identification of disease- and headache-specific mediators and pathways in migraine using blood transcriptomic and metabolomic analysis. J Headache Pain 2021; 22: 117.

61.

Kogelman

Falkenberg

Halldorsson

, et al. Comparing migraine with and without aura to healthy controls using RNA sequencing. Cephalalgia 2019; 39: 1435–1444.

62.

Kogelman

LJA

Falkenberg

Buil

, et al. Changes in the gene expression profile during spontaneous migraine attacks. Sci Rep 2021; 11: 8294.

63.

LaPaglia

Sapio

Burbelo

, et al. RNA-Seq investigations of human post-mortem trigeminal ganglia. Cephalalgia 2018; 38: 912–932.

64.

Edvinsson

Haanes

Warfvinge

, et al. CGRP as the target of new migraine therapies – successful translation from bench to clinic. Nat Rev Neurol 2018; 14: 338–350.

65.

Renthal

Localization of migraine susceptibility genes in human brain by single-cell RNA sequencing. Cephalalgia 2018; 38: 1976–1983.

66.

Vgontzas

Renthal

Migraine-associated gene expression in cell types of the central and peripheral nervous system. Cephalalgia 2020; 40: 517–523.

67.

Yang

Bhuiyan

, et al. Human and mouse trigeminal ganglia cell atlas implicates multiple cell types in migraine. Neuron 2022.

68.

Gallardo

Vila-Pueyo

Pozo-Rosich

The impact of epigenetic mechanisms in migraine: Current knowledge and future directions. Cephalalgia (forthcoming).

69.

Breslau

Lipton

Stewart

, et al. Comorbidity of migraine and depression: investigating potential etiology and prognosis. Neurology 2003; 60: 1308–1312.

70.

Louter

Pelzer

de Boer

, et al. Prevalence of lifetime depression in a large hemiplegic migraine cohort. Neurology 2016; 87: 2370–2374.

71.

Minen

Begasse De Dhaem

Kroon Van Diest

, et al. Migraine and its psychiatric comorbidities. J Neurol Neurosurg Psychiatry 2016; 87: 741–749.

72.

Lantz

Sieurin

Sjölander

, et al. Migraine and risk of stroke: a national population-based twin study. Brain 2017; 140: 2653–2662.

73.

Anttila

Bulik-Sullivan

Finucane

, et al. Analysis of shared heritability in common disorders of the brain. Science 2018; 360.

74.

Yang

Zhao

Boomsma

, et al. Molecular genetic overlap between migraine and major depressive disorder. Eur J Hum Genet 2018; 26: 1202–1216.

75.

Bahrami

Hindley

Winsvold

, et al. Dissecting the shared genetic basis of migraine and mental disorders using novel statistical tools. Brain 2021.

76.

Schurks

Rist

Bigal

, et al. Migraine and cardiovascular disease: systematic review and meta-analysis. BMJ 2009; 339: b3914.

77.

Kurth

Chabriat

Bousser

MG.

Migraine and stroke: a complex association with clinical implications. Lancet Neurol 2012; 11: 92–100.

78.

Tietjen

Conway

Utley

, et al. Migraine is associated with menorrhagia and endometriosis. Headache 2006; 46: 422–428.

79.

Nyholt

Gillespie

Merikangas

, et al. Common genetic influences underlie comorbidity of migraine and endometriosis. Genet Epidemiol 2009; 33: 105–113.

80.

Adewuyi

Sapkota

International Endogene Consortium

, et al. Shared molecular genetic mechanisms underlie endometriosis and migraine comorbidity. Genes (Basel) 2020; 11.

81.

Vgontzas

Pavlović

JM.

Sleep disorders and migraine: review of literature and potential pathophysiology mechanisms. Headache 2018; 58: 1030–1039.

82.

Daghlas

Vgontzas

Guo

, et al. Habitual sleep disturbances and migraine: a Mendelian randomization study. Ann Clin Translat Neurol 2020; 7: 2370–2380.

83.

Georges

Yang

M-L

Berrandou

T-E

, et al. Genetic investigation of fibromuscular dysplasia identifies risk loci and shared genetics with common cardiovascular diseases. Nature Comm 2021; 12: 6031–6031.

84.

Siewert

Klarin

Damrauer

, et al. Cross-trait analyses with migraine reveal widespread pleiotropy and suggest a vascular component to migraine headache. Int J Epidemiol 2020; 49: 1022–1031.

85.

Vollesen

Benemei

Cortese

, et al. Migraine and cluster headache - the common link. J Headache Pain 2018; 19: 89.

86.

Harder

AVE

Winsvold

Noordam

, et al. Genetic susceptibility loci in genomewide association study of cluster headache. Ann Neurol 2021; 90: 203–216.

87.

O'Connor

Fourier

Ran

, et al. Genome-wide association study identifies risk loci for cluster headache. Ann Neurol 2021; 90: 193–202.

88.

Kurth

Rist

Ridker

, et al. Association of migraine with aura and other risk factors with incident cardiovascular disease in women. JAMA 2020; 323: 2281–2289.

89.

Tanha

Martin

Whitfield

, et al. Association and genetic overlap between clinical chemistry tests and migraine. Cephalalgia 2021; 41: 1208–1221.

90.

Tanha

Sathyanarayanan

Nyholt

DR.

Genetic overlap and causality between blood metabolites and migraine. Am J Hum Genet 2021; 108: 2086–2098.

91.

Onderwater

GLJ

Ligthart

Bot

, et al. Large-scale plasma metabolome analysis reveals alterations in HDL metabolism in migraine. Neurology 2019; 92: e1899–e1911.

92.

Harder

AVE

Vijfhuizen

Henneman

, et al. Metabolic profile changes in serum of migraine patients detected using (1)H-NMR spectroscopy. J Headache Pain 2021; 22: 142.

93.

Zheng

Baird

Borges

, et al. Recent developments in mendelian randomization studies. Curr Epidemiol Rep 2017; 4: 330–345.

94.

Haycock

Burgess

Wade

, et al. Best (but oft-forgotten) practices: the design, analysis, and interpretation of Mendelian randomization studies. Am J Clin Nutr 2016; 103: 965–978.

95.

Lee

Bae

, et al. Exploring the causal inference of migraine on stroke: A Mendelian randomization study. Eur J Neurol 2022; 29: 335–338.

96.

Shu

Zhu

, et al. Migraine and ischemic stroke: a mendelian randomization study. Neurol Ther 2022; 11: 237–246.

97.

Daghals

Sargurupremraj

Danning

, et al. Migraine, stroke, and cervical arterial dissection: shared genetics for a triad of brain disorders with vascular involvement. Neurol Genet 2022; 8: e653.

98.

Guo

Rist

Daghlas

, et al. A genome-wide cross-phenotype meta-analysis of the association of blood pressure with migraine. Nat Commun 2020; 11: 3368.

99.

Daghlas

Guo

Chasman

DI.

Effect of genetic liability to migraine on coronary artery disease and atrial fibrillation: a Mendelian randomization study. Eur J Neurol 2020; 27: 550–556.

100.

Chu

, et al. Association between insomnia and migraine risk: a case-control and bidirectional mendelian randomization study. Pharmgenomics Pers Med 2021; 14: 971–976.

101.

Yuan

Daghlas

Larsson

SC.

Alcohol, coffee consumption, and smoking in relation to migraine: a bidirectional Mendelian randomization study. Pain 2022; 163: e342–e348.

102.

Johnsen

Winsvold

Børte

, et al. The causal role of smoking on the risk of headache. A Mendelian randomization analysis in the HUNT study. Eur J Neurol 2018; 25: 1148–e1102.

103.

García-Marín

Campos

Martin

, et al. Phenome-wide analysis highlights putative causal relationships between self-reported migraine and other complex traits. J Headache Pain 2021; 22: 66.

104.

Daghlas

Rist

Chasman

DI.

Effect of genetic liability to migraine on cognition and brain volume: A Mendelian randomization study. Cephalalgia 2020; 40: 998–1002.

105.

Tanha

Nyholt

DR.

Genetic analyses identify pleiotropy and causality for blood proteins and highlight Wnt/β-catenin signalling in migraine. Nat Commun 2022; 13: 2593.

106.

Purcell

Wray

Stone

, et al. Common polygenic variation contributes to risk of schizophrenia and bipolar disorder. Nature 2009; 460: 748–752.

107.

Häppölä

Gormley

Nuottamo

, et al. Polygenic risk provides biological validity for the ICHD-3 criteria among Finnish migraine families. Cephalalgia 2022; 42: 345–356.

108.

Kogelman

LJA

Esserlind

A-L

Francke Christensen

, et al. Migraine polygenic risk score associates with efficacy of migraine-specific drugs. Neurology Genetics 2019; 5: e364.

109.

Knol

Loehrer

Wen

, et al. Migraine genetic variants influence cerebral blood flow. Headache 2020; 60: 90–100.

110.

Ntalla

Kanoni

Zeng

, et al. Genetic risk score for coronary disease identifies predispositions to cardiovascular and noncardiovascular diseases. J Am Coll Cardiol 2019; 73: 2932–2942.

111.

Guo

Rist

Sabater-Lleal

, et al. Association between hemostatic profile and migraine: a mendelian randomization analysis. Neurology 2021; 96: e2481–e2487.

112.

Chen

Xie

, et al. Irritable bowel syndrome and migraine: evidence from Mendelian randomization analysis in the UK Biobank. Expert Rev Gastroenterol Hepatol 2021; 15: 1233–1239.

113.

Chen

Zhang

Zheng

No causal association between coffee consumption and risk of migraine: a mendelian randomization study. Front Genet 2022; 13: 792313.

114.

Yin

Anttila

Siewert

, et al. Serum calcium and risk of migraine: a Mendelian randomization study. Hum Mol Genet 2017; 26: 820–828.

115.

Guo

Daghlas

Gormley

, et al. Phenotypic and genotypic associations between migraine and lipoprotein subfractions. Neurology 2021; 97: e2223–e2235.

116.

Hughes

Wade

Dickson

, et al. Common health conditions in childhood and adolescence, school absence, and educational attainment: Mendelian randomization study. NPJ Sci Learn 2021; 6: 1.

117.

Harrison

Davies

Dickson

, et al. The causal effects of health conditions and risk factors on social and socioeconomic outcomes: Mendelian randomization in UK Biobank. Int J Epidemiol 2020; 49: 1661–1681.

118.

Hannon

Weedon

Bray

, et al. pleiotropic effects of trait-associated genetic variation on dna methylation: utility for refining GWAS loci. Am J Hum Genet 2017; 100: 954–959.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.02 MB

0.01 MB