Shared genetic architecture between depression and migraine: a large-scale genome-wide cross-trait analysis

Depression and migraine are highly prevalent neuropsychiatric disorders that frequently co-occur. However, the extent to which they share genetic determinants, particularly for postpartum depression (PPD), an underexplored subtype, remains insufficiently understood.

We performed a large-scale genome-wide cross-trait analysis using summary statistics for depression phenotypes (overall depression, major depressive disorder [MDD], PPD) and migraine subtypes (migraine, migraine with aura [MA], migraine without aura [MO]). Genetic correlations were estimated using linkage disequilibrium score regression (LDSC) and GNOVA. Pleiotropically associated loci were identified through cross-trait meta-analysis, followed by fine-mapping and colocalization analyses to prioritize potential loci with stronger evidence for shared causal signals. Functional annotation and gene-based analyses were conducted to explore biological relevance. Bidirectional and multivariable Mendelian randomization (MVMR) were applied to assess potential causal relationships.

We observed significant global and local genetic correlations between depression and migraine (LDSC r_g=0.289, P = 3.36 × 10^- 33; GNOVA r_g=0.266, P = 1.32 × 10^- 47), with consistent findings across MDD, PPD, and migraine subtypes. Cross-trait meta-analysis identified 39 genome-wide pleiotropically associated loci, of which two (rs11877758 and rs138709872) showed strong evidence of colocalization (PPH4 > 0.8). Functional annotation demonstrated enrichment in brain regulatory elements and neural pathways. MR analysis provided evidence consistent with a potential causal effect of genetically predicted depression liability on the risk of migraine (odds ratio [OR] = 1.439, P = 6.09 × 10^- 20) and MO (OR = 1.502, P = 4.73 × 10^- 5), whereas reverse-direction associations were also observed but were smaller in magnitude (OR = 1.037, P = 2.56 × 10^- 4). In MVMR, the depression-to-migraine associations remained consistent after adjustment (OR range: 1.405-1.585), while migraine-to-depression associations were attenuated. PPD showed significant genetic correlation with migraine but limited evidence for a potential causal relationship.

Depression and migraine show substantial genetic overlap and exhibit an asymmetric pattern of potential causal relationships, with stronger MR evidence for the depression-to-migraine direction than for the reverse direction. The pleiotropically associated loci provide genomic evidence for future mechanistic and translational investigation across depression and migraine subtypes. The findings also highlight PPD as a genetically related but potentially distinct subtype warranting further investigation.