De novo missense mutations in neurodevelopmental disorders

Abstract

Autism spectrum disorder (ASD) is a pervasive neurodevelopmental disorder (NDD) with a high prevalence in the US (1 in 59 children). It is commonly comorbid with other NDDs such as developmental delay (DD), intellectual disability (ID), and epilepsy (EPI). In this thesis, I examine the role of de novo missense mutations in NDDs with a goal of identifying genes and specific mutations that are candidates for pathogenicity. I characterize the aggregate signal for de novo missense mutations in 8,477 NDD cases, finding both quantitative and qualitative differences between mutations in cases and controls. I also find 40 amino acids that bear de novo substitutions in two or more unrelated individuals and develop a tool to assess the likelihood of these observations in the context of stochastic de novo events. I then use targeted sequencing to further establish the association of these recurrent mutations with disease. Upon finding the same p.Ala646Thr substitution in five cases in glutamate receptor subunit GRIA1, I carry out functional experiments that show alterations in ion flux. I also assessed clustering of de novo missense mutations as this pattern is associated with NDDs, such as Schinzel-Giedion syndrome. I used an unsupervised clustering algorithm, CLUMP, to compare the distribution of de novo missense mutations in NDD cases with private missense events in controls and found 200 genes that were significantly more clustered (p < 0.05). As this set of genes is enriched for neuronal functions, a known association of NDD risk genes, it is likely that clustering is a valid feature for identification of disease genes. With increased exome sequencing on NDD cases, I was able to assess de novo mutation burden in 10,927 cases with ASD, DD, or ID. With two different models, I found 253 total genes with more de novo mutations than expected, 123 of which have a burden of missense mutations. Protein-protein interaction and enrichment analyses of genes with a burden of mutation finds that those with a burden of truncating mutations have roles in transcription regulation while those with missense burden have roles in synaptic signaling. This same neuronal enrichment, including in the amygdala and cortex during fetal development, is seen in genes with clustered de novo missense mutations. Interestingly, the phenotypes of patients with missense mutations in a novel gene, TRRAP, segregate with mutation clustering, suggesting the biological relevance of this pattern of mutation. As burden analysis only identified some of the expected pathogenic NDD genes, I included mutations from patients with EPI to my discovery set. Novel genes identified with this addition are enriched for expression in the striatum. Targeted sequencing of these hotspots of mutation identified additional substitutions at 20 recurrent sites and established 28 new recurrent sites. Eighteen of the sites are known to be pathogenic, and some evidence supports the disease association of the remaining 30 sites. Continued assessment of genes with these patterns of mutation, as well as expansion into gene families, will help to characterize the genetic architecture of NDDs, specifically missense mutations, and provide increased understanding of brain development and pathogenesis.