An oligogenic risk-model for Gilles de la Tourette syndrome based on whole-genome sequencing data

Abstract

AbstractGilles de la Tourette syndrome (GTS) is a neurodevelopmental disorder from the spectrum of tic disorders (TDs). GTS and other TDs have a substantial genetic component with the heritability estimated at between 60 and 80%. Here we propose an oligogenic risk model of GTS and other TDs using whole-genome sequencing (WGS) data from a group of Polish GTS patients and their families (n=185). The model is based on the overrepresentation of putatively pathogenic coding and non-coding genetic variants in genes selected from a set of 86 genes previously suggested to be associated with GTS. Based on the variant overrepresentation (SKAT test results) between unrelated GTS patients and controls based on gnomAD database allele frequencies five genes (HDC, CHADL, MAOA, NAA11, and PCDH10) were selected for the risk model. Putatively pathogenic variants (n = 98) with the median allele frequency of ∼0.04 in and near these genes were used to build an additive classifier which was then validated on the GTS patients and their families. This risk model successfully assigned individuals from 22 families to either healthy or GTS groups (AUC-ROC = 0.6, p < 0.00001). These results were additionally validated using the GTS GWAS data from the Psychiatric Genomic Consortium. To investigate the GTS genetics further we identified 32 genes from the list of 86 genes as candidate genes in 14 multiplex families, including NEGR1 and NRXN with variants overrepresented in multiple families. WGS data allowed the construction of an oligogenic risk model of GTS based on possibly pathogenic variants likely contributing to the risk of GTS and TDs. The model includes putatively deleterious rare and non-coding variants in and near GTS candidate genes that may cooperatively contribute to GTS etiology and provides a novel approach to the analysis of clinical WGS data.