Multi-scale systems genomics analysis predicts pathways, cell types and drug targets involved in normative human cognition variation

Abstract

Identifying genes and cellular pathways associated with normative brain physiology and behavior could help discover molecular therapies that target specific psychiatric symptoms with minimal side effects. Linking genotype-phenotype associations from population-scale datasets to brain function is challenging because of the multi-level, heterogeneous nature of brain organization. To address this challenge, we developed a novel brain-focused gene and pathway prioritization workflow, which maps variants to genes based on knowledge of brain genome regulation, and subsequently to pathways, cells, diseases and drugs (21 resources). We applied this workflow to nine cognitive tasks from the Philadelphia Neurodevelopmental Cohort (subset of 3,319 individuals aged 8-21 years). We report genome-wide significance of variants associated with nonverbal reasoning within the 3’ end of the FBLN1 gene (p=4.6×10-8), itself linked to fetal neurodevelopment and psychotic disorders. These findings suggest that nonverbal reasoning and FBLN1 variation warrant further investigation in studies of psychosis. Multiple cognitive tasks demonstrated significant enrichment of variants in cellular pathways and brain-related gene sets, such as organ development, cell proliferation and nervous system dysfunction. Top-ranking genes in working memory associated pathways are genetically associated with multiple diseases with working memory deficits, including schizophrenia and Parkinson’s disease, and with multiple drugs, suggesting that choice of therapy for memory deficits should consider disease context. Given the large amount of additional biological insight derived from our pathway analysis, versus a standard gene-based approach, we propose that “genes to behaviour” frameworks for modeling brain-related phenotypes, like RDoC, should include pathway information to create a “genes to pathways to behaviour” approach. Our workflow is broadly useful to put genotype-phenotype associations of brain-related phenotypes into the context of brain organization, function, disease and known molecular therapies.