De novo structural variants in autism spectrum disorder disrupt distal regulatory interactions of neuronal genes [Hi-C]

Three-dimensional genome organization plays a critical role in gene regulation, and disruption of chromatin structure has been shown to lead to developmental disorders through the changed contact between key genes and their distal regulatory elements. Structural variants (SVs) have the ability to disrupt local genome organization, such as the joining of topologically associated domains upon deletion of a boundary. Unfortunately, testing large numbers of SVs for their effects on chromatin structure and gene expression is time and cost prohibitive. To overcome these experimental limitations, we previously developed SuPreMo, a convolutional neural network based method that accurately predicts how sequence variants change genome structure. Here, we extended the tool to specifically measure changed contact at regions of interest, such as cell type specific regulatory elements. Using this updated SuPreMo, we tested hundreds of de novo SVs (dnSVs) from autism spectrum disorder (ASD) individuals and their unaffected siblings and predicted how nearby genes' regulatory interactions were affected in neurons. We found that putative cis-regulatory element interactions (CREints) are more disrupted by dnSVs from ASD probands versus unaffected siblings, allowing us to prioritize proband CREints of genes involved in neuronal development. We experimentally validated our top locus using induced pluripotent stem cell-derived excitatory neurons with and without the dnSV and found that the variant resulted in mis-regulation of 1102 genes. This in vitro characterization of a candidate causal variant is important, because most ASD patients do not carry a damaging protein-coding variant that alters neurodevelopment or neuronal function. This study establishes disrupted genome folding as a new genetic mechanism for ASD and provides a general strategy for prioritizing variants predicted to disrupt regulatory interactions in any tissue. Overall design: To investigate the effect of the full and CTCF deletions on 3D genome contacts, HiC was performed on two WT cell lines, two CTCF deletion cell lines, and two full deletion cell lines.