Functional Impact of Copy Number Variants in Autism Probands

Rohan Alibutud, Vaidhyanathan Mahaganapathy, Xiaolong Cao, Marco Azaro, Christine Gwin, Sherri Wilson, Steven Buyske, Christopher W. Bartlett, Judy F. Flax, Linda M. Brzustowicz, Jinchuan Xing

Autism spectrum disorder (ASD) is a neurodevelopmental disorder with a complex polygenic genetic architecture. Large structural variants (SVs), and in particular rare copy number variants (CNVs), have been previously linked to idiopathic ASD. This project is a part of the New Jersey Language and Autism Genetics Study (NJLAGS), a collaborative effort to increase the searching power behind ASD gene identification by studying potential overlap with genes causal for language impairment (LI) and reading impairment (RI) disorders. The NJLAGS cohort provided genotyping microarray data from which we called CNVs using multiple calling algorithms for increased coverage. From 522 individuals across 115 families, we filtered and prioritized a set of 186 de novo CNVs (ranging from ~15 kbp to ~5.2 Mbp in size) that are most likely to be contributing to idiopathic ASD in probands. In order to assemble a ranking of the best candidate variants, we employed the Structural Variant Caller Trained on Variants Rare and Exonic (StrVCTVRE), a random forest classifier that predicts pathogenicity of SVs based on an array of traits. We combined StrVCTVRE score with tissue expression, prior annotation, segregation pattern, and overrepresentation analysis to accumulate evidence for ASD causation. The results implicate genes in the mTOR signaling and clathrin-mediated endocytosis pathways, such as RPTOR, CLTC, and PICALM as potential ASD contributors and as targets for future research.