Exome Sequencing Identifies Genes Where Rare Disruptive Variants Confer Risk for Autism

Panel Presentation
Friday, May 3, 2019: 10:55 AM
Room: 517A (Palais des congres de Montreal)
F. Satterstrom, Analytic and Translational Genetics Unit, Massachusetts General Hospital, Boston, MA; Stanley Center for Psychiatric Research, Broad Institute of MIT and Harvard, Cambridge, MA
Background: Rare genetic variants—particularly newly arising, or de novo, mutations—are major contributors to individual risk for autism spectrum disorder (ASD). As a class, the most impactful rare variants are those which disrupt gene function. These variants can be identified by exome sequencing, which captures the protein-coding portion of the genome. When individuals with ASD carry rare disruptive variants in a gene more often than expected by chance, it implicates that gene in risk for ASD. In this way, previous study (e.g. Sanders et al., 2015, Neuron 87:1215-1233) has identified upward of 65 genes associated with ASD at a false discovery rate (FDR) ≤ 0.1.

Objectives: We sought to 1) identify novel genes associated with ASD and 2) learn more about the neurobiology of ASD by analyzing the set of ASD-associated genes. To do this, we conducted the largest exome sequencing study to date focused on rare variation in ASD. We represent the Autism Sequencing Consortium, as well as the iPSYCH initiative for psychiatric research in Denmark, and our collaborative network enabled us to sequence samples from around the world.

Methods: Building upon previous work, our analysis included exome sequences from a total of 35,584 samples. We identified de novo variation in 6,430 ASD probands and 2,179 sibling controls for whom we had sequenced both parents, and we identified rare variation in an additional 5,556 ASD cases and 8,809 controls. We integrated the de novo and case-control variation using a Bayesian framework called TADA to identify ASD-associated genes. We improved the TADA model used in previous work by incorporating pLI score as a metric for weighting protein-truncating variants and MPC score as a metric for weighting missense variants. We then analyzed our list of ASD-associated genes in the light of multiple external datasets.

Results: We identify 102 genes associated with ASD at an FDR ≤ 0.1. Of these risk genes, 48 show higher frequencies of disruptive (i.e. protein-truncating or MPC ≥ 1 missense) de novo variants in individuals from a separate cohort ascertained for severe neurodevelopmental delay, while 51 show higher frequencies in individuals from our study ascertained for ASD, and comparing our ASD cases with disruptive mutations in the two groups of genes shows differences in phenotypes like IQ and age of walking. Expressed early in brain development, most of the risk genes have roles in neuronal communication or regulation of gene expression, and 12 fall within recurrent copy number variant loci. In human cortex single-cell gene expression data, expression of the 102 risk genes is also enriched in both excitatory and inhibitory neuronal lineages, implying that disruption of these genes alters the development of both neuron types.

Conclusions: Our identification of 102 ASD-associated genes represents a significant step forward from previous work. Next, we plan to jointly analyze our data with exome sequences from SPARK, the Simons Foundation study of ASD whose recent data release will nearly double our sample size. A few preliminary results may be ready to discuss at the meeting.