From gene to cognition: mapping the effects of genomic deletions and duplications on cognitive ability

Copy Number Variants (CNVs) are DNA segments’ deletions or duplications with potential consequences for Cognitive Ability (CA) and many psychiatric disorders. Recent research suggested that scoring genes encompassed in a CNV based on their intolerance to loss of function (LOEUF) could predict the effects of CNVs on CA. However, these models remain inaccurate due to the fact that they carry no information on protein function coded by the gene. In the past decade, functional genomic resources characterizing spatial and cell-type and tissue expressions have greatly developed. These resources have been essential for annotating genes linked to psychiatric conditions. However, connecting postmortem-derived data to clinical phenotypes has proven challenging. The one-by-one analysis of rare CNVs is limited to a small number of variants. To accurately assess the risk associated with rare CNVs, it is necessary to investigate them in the aggregate and consider the information on the nature and function of the disrupted genes. The functional information of genes, including the information of where (space) genes are expressed in the brain and in what cell types and tissues, and Gene Ontology terms (GO-terms) can explain the association between CNVs and CA. To estimate the effect of CNVs on CA using functional information. We analyzed a dataset, including CNVs > 50kb and CA assessment of 258k individuals from 6 general population cohorts. Our analytical approach comprised two steps: Partitioning the genome into functional gene sets by using i) spatial gene expression data across the cortex (space) and, ii) pre and post-natal cell types using single-cell transcriptomic data, bulk using GTEx, and GO-terms. Estimating the mean effect size on CA per category using a linear regression model, for deletion and duplication separately. In this model, CA is considered as a function of the number of genes deleted or duplicated, inside and outside of each category. Preliminary results using space and LOEUF information indicate that gene with deletion has a greater effect on sensorimotor regions and a smaller effect on association regions, with a 7-fold difference in the most intolerant genes. Gene with duplication, on the other hand, has a larger impact on association regions and a smaller impact on sensorimotor regions, showing a 4-fold difference in the most intolerant genes. Moreover, using information of cell types, tissues, GO-terms, and LOEUF, we found that deleted genes decrease CA more than duplicated genes; the direction of the significant effect sizes on CA is highly dependent on how much genes are mutation-intolerant: 91% and 85% of intolerant genes decrease the CA when deleted and duplicated respectively (the remaining increase the CA). Moreover, 33% of the tolerant genes increase CA when deleted or duplicated. We discovered that there is a mirror effect size on cognition of deletion and duplication across the cortical brain gradient. For deletions, the largest effects on CA are observed in genes highly expressed in the unimodal regions of the brain involved in action and perception. Duplications affect CA most through integrative regions on the other end of the cortical organization; key brain structures in human evolution that are involved in perceptually decoupled cognition and internal thought. Integrating different data sources adds complexity to the analysis, necessitating additional investigations that combine mathematical and biological validation methods.