Investigating the Role of Noncoding De Novo Single-Nucleotide Variants in Autism Spectrum Disorder

Abstract

Autism spectrum disorder (ASD) is a complex and heterogeneous neurodevelopmental condition characterized by challenges with communication, social interaction, and behavior. With an estimated heritability between 50-90%, a strong genetic basis has been established ASD. While many ASD-associated genetic variants in coding regions have been identified, genome-wide association studies have shown that most trait-associated variants lie within noncoding regions. Therefore, here I have focused on characterizing the contribution of noncoding emph{de novo} SNVs (dnSNVs) to ASD risk.

To accomplish this, I leveraged whole-genome sequencing data from 1,917 families in the Simons Simplex Collection. In Chapter 2 describe the pipeline I have established to improve the accuracy of genotype and variant calls, in order to ensure a high quality list of dnSNVs. I then introduce the computational tools I used to prioritize variants and identify cis-regulatory elements. I show that there is a strong enrichment of high-impact coding dnSNVs in probands, but significance levels do not withstand multiple-testing correction in noncoding regions.

In Chapter 3 I present power analyses suggesting that a larger sample size may be necessary in order to detect association between ASD and noncoding dnSNVs in probands. I also show that certain annotation categories are better than others at capturing meaningful differences between probands and siblings. After discussing the challenges in screening for ASD-associated noncoding dnSNVs, I provide suggestions as to how those challenges can be addressed for future studies.

I developed a web application database, emph{De Novo} Browser, which I introduce in Chapter 4. In this work I have curated an annotated list of 267,000 dnSNVs. I have made this list publicly available on the emph{De Novo} Browser, where the variants can be explored in table form and sorted by a variety of features and annotations. Together, my dissertation enhances our understanding of the role of noncoding variation in ASD, while also providing a tool and recommendations to benefit future studies.