W110. COMPARATIVE AND INTEGRATIVE ANALYSIS OF PSYCHOSIS GENOMICS IN NIGERIA AND TRINIDAD

Abstract

Background Psychotic disorders are a major source of disability and premature mortality worldwide, yet there are remarkably few robust, directly comparable epidemiological studies of untreated psychosis in the Global South, and the genomic architecture of psychosis has been defined overwhelmingly in European-ancestry populations. Methods INTREPID-II is therefore distinctive: it the first multi-country epidemiological programme of untreated psychosis in the Global South in more than four decades, combining incidence, case-control and cohort designs with rigorous population-based case ascertainment, matched controls and harmonised cross-setting assessment. We present ongoing genomic analyses from the Nigerian and Trinidadian arms of the International Research Programme on Psychoses in Diverse Settings (INTREPID-II) [N=243 cases and 235 controls] and Psychosis in Ancestrally Diverse Settings (PADS) [N=2,940 cases and 2,580 controls]. Results We compare diagnostic distributions across INTREPID/PADS settings and test genomic influences on clinical presentation, symptom dimensions, cognition, physical comorbidity and treatment response. Whole-genome sequencing data, generated through the Blended Genome Exome sequencing platform, enables joint analysis of common, low-frequency and structural variation, alongside fine-scale population structure and local ancestry in Nigerian and admixed Trinidadian genomes. To determine how far established psychosis findings extend across populations, we evaluate the generalisability of established common-variant associations and construct ancestry-aware polygenic scores using TrANCES-PRS, a new benchmarking framework, to quantify portability, calibration and gains from trans-ancestry modelling. We also interrogate rare-variant architecture through the burden of disruptive coding mutations and pathogenic copy number variants. Discussion A major strength of the INTREPID/PADS programme is the depth of harmonised assessment across genetic and environmental domains. The platform combines population-based ascertainment with structured phenotyping of symptoms, cognition, disability, physical health, substance use, childhood trauma, lifetime adverse events, support networks, family burden, help-seeking and impact, alongside biological sampling and geospatial data. Using Bayesian software developed by PGC Africa, we model how cumulative exposures (including trauma, adverse life events, urbanicity, substance exposure, and broader economic and social conditions) combine with genomic liability to shape risk, help-seeking and illness course across settings.