80. IMPLICATION OF COMPLEX STRUCTURAL GENOME VARIATION IN THE GENETIC ARCHITECTURE OF NEUROPSYCHIATRIC DISORDERS: INSIGHTS FROM HUMAN POPULATION ANALYSIS AND FROM POSTMORTEM BRAINS OF INDIVIDUALS WITH PSYCHIATRIC DISORDERS

• Published in: European neuropsychopharmacology Vol. 87; p. 93
• Main Authors: Zhou, Bo, Arthur, Joseph, Guo, Hanmin, Kim, Taeyoung, Huang, Yiling, Pattni, Reenal, Song, Giltae, Palejev, Dean, Dohna, Heinrich, Roussos, Panos, Kundaje, Anshul, Hallmayer, Joachim, Snyder, Michael, Wong, Wing, Urban, Alexander
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01-10-2024
• ISSN: 0924-977X
• DOI: 10.1016/j.euroneuro.2024.08.194

Psychiatric disorders such as schizophrenia and bipolar disorder have a strong but complex genetic component to their etiology. Multiple candidate loci have been identified by large GWAS, but individual variants are expected to have small effect sizes and to act in combination with other variants across individual genomes. Oftentimes the functional variant will not be the GWAS marker SNP, but a nearby variant of a different nature, such as a structural variant in the genome. All human genomes contain complex structural variations (cxSVs), but their functions in genome biology are mostly unknown as they are effectively excluded from standard genome analyses, due to the severe technical limitations on accurately detecting these variants.
We developed Automated Reconstruction of Complex Structural Variations (ARC-SV), a probabilistic and machine-learning method (which leverages the new Human Pangenome Reference) that permits the characterization of cxSVs on a population scale with unprecedented accuracy from standard short-read whole-genome sequencing (WGS). We applied ARC-SV to 4,262 genomes spanning all continental populations to analyze the patterns of rare and common cxSVs in comparison to other variant types. We then performed linkage analysis between cxSVs and the psychiatric GWAS risk alleles found in 119 PsychENCODE post-mortem brains. By developing a statistically rigorous bioinformatic method that overcome sample size limitations, we integrated multi-modal data dimensions (genotype, phenotype, single-nuclei RNA-seq/ATAC-Seq) from these 119 brains.
From 4,262 genomes spanning all continental populations, we identified 8,493 cxSVs belonging to more than 12 subclasses. We found cxSVs that are at very low allele frequencies in the human population (i.e. those with signatures of negative selection) to be especially enriched, more than other variant types, for neural genes and for loci undergoing rapid evolution, including those that regulate human-specific corticogenesis. By leveraging single-nuclei multiomics across 119 PsychENCODE brains, where cxSVs are enriched for linkage with psychiatric GWAS risk alleles, we find, in multiple brain regions, significant associations between cxSVs and differentially expressed genes and accessible chromatin. Integration of genotype, single-nuclei RNA-seq, and phenotype reveals significantly decreased expression of cxSV-associated genes among psychiatric cases, implicating cxSVs as a contributing factor in neuropsychiatric disorders.
ARC-SV is an efficient tool for SV analysis, especially for population-scale studies of the human genome, in general, and across large disease cohorts, and for the analysis of the large-scale WGS paired with functional genomics data for the understanding of the molecular etiology of major human diseases. We show for the first time that cxSVs in the human genome are integral components of the complex genetic architecture of neuropsychiatric disorders and that they are enriched for linkage with psychiatric GWAS risk alleles directly in the post-mortem brains of individuals with psychiatric disorders. This connection between GWAS loci, the actual functional DNA sequence variants that are tagged by the marker SNPs used in the GWAS, and the molecular effects in the actual disease tissue, is one of the major and unresolved challenges in psychiatric genetics (and for complex diseases in general).
Nothing to disclose.