Deciphering the shared genetic architecture between schizophrenia and suicide attempt: a statistical genomic study

Evidence for reciprocal comorbidity of schizophrenia (SCZ) and suicide attempt (SA) has been discussed in recent years. However, little evidence is known regarding the shared genetic architecture between SCZ and SA. This study investigated the genetic correlation and causality between them.

Leveraging the genome-wide association summary statistics from European ancestry, we qualified local and global genetic correlations between the two traits through Heritability Estimation from Summary Statistics (ρ-HESS) and Linkage Disequilibrium Score Regression (LDSC). Cross-trait meta-analyses, cross-phenotype association, and colocalization analyses were utilized to identify and validate novel risk loci. Bidirectional Mendelian randomization (MR) was used to investigate causality. Tissue-level SNP heritability enrichment was assessed through Stratified LDSC and Multi-marker Analysis of GenoMic Annotation (MAGMA).

LDSC identified significant positive genetic correlation between SCZ and SA (r_g = 0.48, p = 1.23 × 10^- 89), and ρ-HESS observed 13 distinct regions associated with both traits. Cross-trait meta-analysis identified 4,025 significant SNPs across the two phenotypes, corresponding to 79 independent loci after LD-based clumping and removing those in LD region. Then, variant functional annotation identified 44 SNPs as novel risk loci, and colocalization analysis identified five novel shared loci among these SNPs between SCZ and SA, including rs10456045 (PP.H4 97.98%), rs13195401 (PP.H4 93.65%), rs13198474 (PP.H4 95.91%), rs60135207 (PP.H4 89.89%), and rs3132450 (PP.H4 88.68%). In MR analysis, genetically predicted SCZ was positively associated with SA (OR 1.17, 95%CI 1.37 to 1.21, p = 7.39 × 10^- 22), with no reverse causality. Tissue-specific analysis showed that associated genes were predominantly enriched in brain tissues, particularly cortical, limbic, and basal ganglia-related regions.

Our study suggests a shared genetic architecture between SCZ and SA, characterized by overlapping pleiotropic variants, tissue-specific expression, and functionally shared genes. These findings delineate previously underappreciated genetic overlaps between them and provide a framework for future mechanistic exploration and therapeutic target discovery.

Not applicable.