Lessons from the analysis of TAD boundary deletions in normal population

Abstract

Topologically Associating Domains (TAD)-boundaries induce spatial constraints, allowing interaction between regulatory elements and promoters only within their TAD. Their disruption could lead to disease, through gene-expression deregulation. This mechanism has been shown in only a relatively low number of diseases and a relatively low proportion of patients, raising the possibility of TAD boundary disruption without phenotypical consequence. We investigated, therefore, the occurrence of TAD boundaries disruption in the general population. Coordinates of 307,430 benign deletions from public databases were crossed with 36 Hi-C datasets. Differences in gene content and gene localization were compared in the TADs, according to the possible disruption of their boundaries by a deletion found in the general population. TADs with no deletion encompassing their boundaries (R-TAD) represented 38% of TADs. Enrichment in OMIM genes as well as in morbid genes was observed in R-TADs and genes in R-TADs were found to localize closer to the boundaries. Our results support recent publications tempering the impact of breaking TADs on gene expression with a majority of broken TADs in the general population. A subgroup of R-TAD emerges from this analysis with enrichment in disease genes and their coordinates could be used to annotate CNV from pangenomic approaches to enhance data interpretation.