¹Centro de Biologia Molecular Severo Ochoa (CBMSO), Molecular Neuropathology, Madrid, Spain, ²University of New South Wales (UNSW), School of Clinical Medicine, Discipline of Psychiatry and Mental Health, Sydney, Australia, ³Autonomous University of Madrid, Institute of Molecular Biology, Madrid, Spain, ⁴Autonomous University of Madrid, Department of Biology, Madrid, Spain, ⁵University Hospital Jiménez Díaz Foundation, Department of Psychiatry, Madrid, Spain, ⁶University Hospital Infanta Elena, Department of Psychiatry, Valdemoro, Madrid, Spain, ⁷University Hospital Rey Juan Carlos, Department of Psychiatry, Mostoles, Madrid, Spain, ⁸General Hospital of Villalba, Department of Psychiatry, Villalba, Madrid, Spain, ⁹University Hospital Jimenez Díaz Foundation, Department of Psychiatry, Madrid, Spain, ¹⁰Autonomous University of Madrid, Department of Psychiatry, Madrid, Spain

Background: Suicide represents a major public health challenge and remains the leading cause of unnatural death in Spain, with more than 4,000 deaths annually. Suicidal behaviour shows heritability, partly shared with psychiatric disorders but also driven by distinct biological mechanisms. In parallel, pharmacological treatments commonly used in patients with suicidal ideation, such as antidepressants, mood stabilizers, and antipsychotics, exhibit marked interindividual variability in efficacy and adverse effects, partially explained by genetic variation in drug-metabolizing enzymes. Integration of genetic, epigenetic, transcriptomic, and pharmacogenetic information offers a promising avenue to advance personalized suicide prevention.

Aims & Objectives: This study aims to characterize the genetic and pharmacogenetic architecture of suicidal behaviour in Spanish populations by integrating multiple omics layers. The specific objectives are: i) to identify genomic risk loci and sex-specific genetic effects associated with suicide attempts (SA), and generate polygenic risk scores (PRS) for suicide; ii) to uncover transcriptomic and epigenomic signatures linked to SA, including brain-imputed regulatory effects; and iii) to develop and validate computational models integrating common and rare alleles (SNPs; SNVs) with copy number variants (CNVs) in key pharmacogenes to accurately infer drug metabolizer phenotypes, and to investigate biological age acceleration derived from epigenomic and transcriptomic data in relation to psychotropic medication exposure and treatment duration.

Method: We conducted a multi-omics investigation in a nationwide Spanish cohort. Genome-wide association studies (GWAS) were performed in 812 individuals with SA and 4,446 controls, followed by sex-stratified analyses in 303 male SA versus 2,303 male controls and 509 female SA versus 2,143 female controls. Genomic data were also used to generate PRS for suicide behaviour and related psychiatric phenotypes. Transcriptomic profiling using whole-blood RNA sequencing was conducted in 153 SA and 567 controls to identify differentially expressed genes (DEGs) and assess brain-imputed expression patterns. Epigenome-wide association studies (EWAS) were performed in 43 SA and 195 controls to identify differentially methylated positions (DMPs) and regions (DMRs). Pharmacogenetic analyses of allelic variants and CNVs in key psychotropic drug-metabolizing genes were conducted using GenoStaR.

Results: GWAS analyses did not identify genome-wide significant loci for SA; however, 19 suggestive association signals were observed, supporting a highly polygenic architecture. Sex-stratified analyses revealed partially distinct association patterns between males and females. PRS showed modest but consistent associations with SA. Transcriptomic analyses identified DEGs involved in immune, inflammatory, and stress-related pathways. EWAS did not identify significant DMPs and DMRs. Pharmacogenetic analyses revealed substantial interindividual variability in metabolizer status driven by SNPs and CNVs in CYP genes, highlighting potential clinical relevance for psychotropic treatment optimization.

Discussion & Conclusions: This integrative multi-omics approach provides a comprehensive view of the biological architecture underlying suicidal behaviour, emphasizing polygenic risk, sex-specific effects, and regulatory mechanisms beyond single-locus associations. The convergence of genomic, transcriptomic, and epigenomic findings supports biological pathways partially independent of psychiatric diagnoses. Incorporation of pharmacogenetic variation in CYP enzymes addresses a major source of variability in treatment response and safety, advancing the translational potential of precision approaches to suicide prevention.