UCSF

Autism Spectrum Disorder (ASD) is a genetically complex and heterogenous neurodevelopmental disorder. As such, much effort has been put into uncovering the risk genes underlying ASD. A recent large-scale whole exome sequencing study focusing on de novo and case-control rare variants has identified 102 high-confidence ASD (hcASD) risk genes with a False Discovery Rate (FDR) ≤ 0.1(Satterstrom et al., 2020). Despite the advances in the discovery of ASD risk genes, we have yet to understand the molecular underpinnings of ASD pathobiology. To understand how hcASD risk genes contribute to ASD phenotypes, it is imperative to utilize integrative networks and systems biology approaches to unravel the molecular pathways connecting these hcASD risk genes. In this dissertation, I show how we used quantitative proteomics to systematically define the physical interaction landscape of proteins encoded by hcASD genes, and how these interactions are disrupted when we introduced de novo missense mutations as observed in the patients. The ASD protein-protein interaction network identifies 1024 unique proteins that interact with at least one of the 102 hcASD risk genes; of note, 82% of the interactions are novel. When we introduce patient-derived missense mutations in 35 out of 102 hcASD risk genes, we observed 133 protein interactions that are more specific to the mutants and 152 proteins that are more specific to the wild-type (WT) proteins. These differential protein interactions can be used to generate hypothesis regarding the molecular underpinnings of ASD etiology. Additionally, I present how we can elucidate biological processes, molecular pathways, and protein complexes from the generated protein-protein interaction network using network biology approaches and functional enrichment analyses, highlighting the convergent pathways that these high confident ASD risk genes may be involved in.