Abstract
The significant contribution of structural variants to function, disease, and evolution is widely reported. However, in many cases, the mechanism by which these variants contribute to the phenotype is not well understood. Recent studies reported structural variants that disrupted the three-dimensional genome structure by fusing two topologically associating domains (TADs), such that enhancers from one TAD interacted with genes from the other TAD, and could cause severe developmental disorders. However, no computational method exists for directly scoring and ranking structural variations based on their effect on the three-dimensional structure such as the TAD disruption to guide further studies of their biological function. In this paper, we formally define TAD fusion and provide a combinatorial approach for assigning a score to quantify the level of TAD fusion for each deletion denoted as TAD fusion score. We also show that our method outperforms the approaches which use predicted TADs and overlay the deletion on them to predict TAD fusion. Furthermore, we show that deletions that cause TAD fusion are rare and under negative selection in general population. Finally, we show that our method correctly gives higher scores to deletions reported to cause various disorders (developmental disorder and cancer) in comparison to the deletions reported in the 1000 genomes project.