SEX-STRATIFIED GWAS META-ANALYSES OF SUICIDAL IDEATION AND SUICIDAL BEHAVIORS

Background Genetic research on suicidality has established significant heritability of suicidal ideation (SI), suicide attempt (SA), and suicide death (SD), and yet epidemiological findings also point to significant sex differences across these risk phenotypes: rates of death by suicide are approximately 3:1 male to female, despite a higher population prevalence of suicide attempts in females. Until now, there have been inadequate genetic resources to examine genetic risks for these suicide phenotypes in a sex-specific manner, and to test whether sex-specific analysis can reduce the genetic heterogeneity of SI, SA and SD. In this presentation, the Suicide Working Group of the Psychiatric Genomics Consortium will report updated results from sex-stratified GWAS meta-analyses of these risk phenotypes. Methods 120 analytic cohorts (61 female, 59 male) contributed to sex-specific GWAS meta-analyses of suicidal ideation, suicide attempt, and suicide death. A suicidal behaviors (SB) GWAS meta-analysis comprised the aggregation of all SA and SD cohorts. Cohorts included African/African American (SI, SA), European (SI, SA, SD), Hispanic/Latinx (SI, SA, SD) and East Asian ancestral admixtures (SI, SA, SD). Standardized phenotyping and analytic protocols were employed by PGC SUI to ensure exceptional rigor and comparability across cohorts. GWAS meta-analyses were conducted within sex via inverse variance-weighted fixed effects models, and post-GWAS analyses estimated SNP-heritabilities (h2SNP), gene/pathway/tissue enrichment, sex-specific and female-male shared genetic covariances of SI, SA, and SD, and sex-stratified genetic correlations of each suicide phenotype with psychiatric and medical risk phenotypes. SMR and drug enrichment analyses were also conducted. Results Sex-stratified, multi-ancestry GWAS of SB yielded 9 GWS loci in females (N = 32,340/291,861 cases/controls) and 8 GWS loci in males (N = 36,135/740,708 cases/controls). Several mapped genes were also GWS in sex-stratified MAGMA gene-based tests. h2SNP estimates were significant at 10.2% for females and 6.4% for males. Genetic correlations within sex and across SI and SA were statistically significant and were also significantly < 1.0, indicating impartial overlap of genetic signal across suicidality phenotypes. Genetic correlations across sex and within the suicide phenotypes were also statistically significant and significantly < 1.0, indicating sex-specific genetic signal. SD genetic correlations evidenced the same pattern of results, but SD correlational analyses in females remain underpowered. Genetic correlations of suicide phenotypes with psychiatric disorders pointed to sex-specific risk factors driving genetic heterogeneity across SI and SA. Several additional post-GWAS results will be presented. Discussion Results underscore the importance of considering sex-specific genetic architectures in research on suicide risk and point to unique cross-phenotype relationships in females and males. Suicide phenotypes in females and males showed substantial genetic overlap, but also indicated a substantial burden of genetic risk which could be due to sex-specific risk variants. Lack of power for SD genetic analyses in females and non-European ancestral populations remains a concern, and results underscore the significant potential scientific and public health impact of increasing these data resources.