A major challenge to understanding the causes of autism spectrum disorder (ASD) is the high and ever-growing number of genes linked to the disorder. Focusing on one gene at a time using a limited set of very low-throughput assays has yielded potential cellular phenotypes but for only a handful of the hundreds of genes linked to ASD. The complexity of this problem is further confounded by the identification of multiple mutations within the same ASD-associated genes, culminating in many thousands of gene variants currently without functional phenotyping.

In order to make significant headway into understanding the functional impact that large numbers of variants from many genes has in ASD, Kurt Haas and his colleagues have developed a multiple-model systems approach. Their approach takes advantage of high-throughput assays to quickly identify likely disease-causing missense mutations. These mutations are then prioritized for testing in low-throughput but high-resolution assays that assess the effects of these variants on vertebrate neural development likely central to the pathophysiology underlying ASD. In earlier efforts, supported by a previous SFARI award, Haas and colleagues have already begun to demonstrate the power of this approach^{1, 2}.

The current project will expand upon these efforts, taking advantage of a centralized molecular biology pipeline, along with expertise in bioinformatics and with various organisms and bioassays. These include high-throughput screens of genetic and protein interactions, and protein stability and function in human cell lines, Saccharomyces, Drosophila and C. elegans; high-resolution assays of neuronal growth and synaptogenesis in rat primary hippocampal and dorsal root ganglion neuronal cultures; and in vivo studies in Drosophila and Xenopus tadpoles.

This platform that Haas and colleagues have put in place allows a seamless transition between model systems, producing deep functional profiling of individual gene variants in a wide assortment of cellular environments, identifying mechanisms of protein dysfunction for each variant, advancing structure-function models, and providing a rich data set for the optimization of algorithms for predicting the impact of missense mutations on protein function. This platform will also aid in validating low/medium confidence ASD genes by testing whether such missense variants produce protein dysfunction.

Combined, the results of this large-scale variant screening and assessment approach will provide critical data on a large number of genes and gene variants that are likely causal or increase the risk for ASD. The findings are also expected to provide insight into the mechanisms by which these genes and gene variants cause protein dysfunction and neuropathophysiology.