Poster #64: Lionel Sequeira

Identification of Candidate Causal Genes for Tourette Disorder from a Quartet Family

Lionel Sequeira, BSc (1, 2) , Robert A. King, MD (3) , Jay A. Tischfield, PhD (1, 2), Gary A. Heiman, PhD (1, 2) , and Jinchuan Xing, PhD (1, 2) 1. Department of Genetics, Rutgers, The State University of New Jersey, Piscataway, NJ 08854, USA 2. Human Genetics Institute of New Jersey, Rutgers, The State University of New Jersey, Piscataway, NJ 08854, USA 3.Yale Child Study Center, Yale School of Medicine, New Haven, CT 06510, USA

Neurodevelopmental disorders (NDDs) are a group of heritable conditions that share a high degree of comorbidity. However, extensive phenotypic diversity across different individuals obscures simple genotype-phenotype correlations. CELSR3 (Cadherin EGF LAG Seven-pass G-type receptor 3) has recently been identified as a high-confidence risk gene across several NDD phenotypes, including Tourette Disorder. Prior trio analyses of a pathologically complex family identified two high-impact CELSR3 variants: a de novo splice-site variant in the son and a stop-gain variant in the father. We extend these analyses by sequencing the daughter, who was previously unstudied, and re-diagnosing the neurological conditions of all four family members, enabling a quartet-based assessment of CELSR3 to identify de novo, co-segregating, and potentially damaging variants as candidate causal contributors to Tourette Disorder and NDDs. Whole exome sequencing data were mapped and aligned using BWA and GATK, and variants were annotated using ANNOVAR and slivar, supplemented by gnomAD API queries. Filters were applied for mapping quality, read depth, genotype quality, rarity (alternative allele frequency <= 0.01), and predicted variant damage consequence. This was followed with segregation filters retaining variants that were de novo in either child, or shared by both parents and inherited by at least one child. Remaining variants of non-PASS quality were manually curated in IGV. Candidate genes were annotated using HGNC, IMPC, SFARI, a curated NDD gene list, brain expression datasets (GTEx, BrainSpan, HDBR), and the GWAS Catalog. This pipeline reduced 181,964 called variants to 35 candidate variants across 30 genes. A more stringent subset consisting of rare, non-dispensable variants free of gnomAD region flags yielded ten variants in ten genes. The de novo CELSR3 splice variant in the son was recovered, reaffirming CELSR3 as a high-confidence NDD risk gene, while the remaining candidates provide novel targets for functional validation.