Whole-Genome Sequencing in ASD Quartets and Integration with Regulatory Elements Active during Human Brain Development

Background: Large-scale genotyping and whole-exome sequencing studies have strongly established the contribution of de novo copy number variants (CNVs) and rare de novo coding mutations, respectively, to autism spectrum disorder (ASD) risk. However, the contribution of rare de novo non-coding mutations, such as those identified by whole-genome sequencing (WGS) has not been well characterized. This is partly due to the difficulty in interpreting the biological effects of non-coding variation, especially since the complement of regulatory regions is temporally and spatially specific. Several recent studies have identified strong convergence of ASD-associated mutations during midfetal development, indicating that regulatory elements active during this critical point of brain development may be particularly relevant to ASD. Thus, we report preliminary results from WGS of 40 quartets with simplex ASD, and interpret them in the context of regulatory elements active during midfetal human brain development. 

Objectives: (1) To utilize WGS to identify de novo non-coding mutations in probands and matched sibling controls from 40 quartets in the Simons Simplex Collection (SSC); (2) to annotate these mutations based on non-coding regulatory elements active in developing cortex; and (3) to associate regulatory elements and their genic regulatory targets, with ASD based on recurrence of mutations within these regions.

Methods: WGS was conducted on the Illumina HiSeq X Ten to a minimum mean coverage of 30x. Data were aligned using BWA-mem and duplicate reads were removed with Picard. Variants were called with the Genome Analysis Toolkit. Published and in-house maps of regulatory elements active during human brain development were integrated with identified variants in order to determine the burden of these mutations within active regulatory elements in probands versus matched sibling controls. Based on the rate in unaffected siblings, statistical models were developed to identify particular regulatory elements with a statistically significant overrepresentation of mutations. These regulatory regions were then associated with nearby genes using computational methods and existing databases of genome topology domains.

Results: Non-coding de novo variants identified by WGS are present within regulatory elements active during midfetal brain development, as identified by ChIP-seq for active chromatin marks in human cortical tissue. Additionally, regulatory targets of specific chromatin modifiers and transcription factors associated with ASD are also impacted by de novo variation.

Conclusions: WGS in 40 quartet families with simplex ASD has identified a large number of non-coding de novo mutations. Integration of this data with chromatin state and targeted ChIP-seq data from the developing human brain helps prioritize which of these variants are functional, and indicates which regulatory elements and their corresponding genes may be relevant to ASD risk.