The distributions of protein coding genes within chromatin domains in relation to human disease

Abstract Background Our understanding of the nuclear chromatin structure has increased hugely during the last years mainly as a consequence of the advances in chromatin conformation capture methods like Hi-C. The unprecedented resolution of genome-wide interaction maps shows functional consequences that extend the initial thought of an efficient DNA packaging mechanism: gene regulation, DNA repair, chromosomal translocations and evolutionary rearrangements seem to be only the peak of the iceberg. One key concept emerging from this research is the topologically associating domains (TADs) whose functional role in gene regulation and their association with disease is not fully untangled. Results We report that the lower the number of protein coding genes inside TADs, the higher the tendency of those genes to be associated with disease (p-value = 4 × $$10^{-54}$$10-54). Moreover, housekeeping genes are less associated with disease than other genes. Accordingly, they are depleted in TADs containing less than three protein coding genes (p-value = 3.9 × $$10^{-34}$$10-34). We observed that TADs with higher ratios of enhancers versus genes contained higher numbers of disease-associated genes. We interpret these results as an indication that sharing enhancers among genes reduces their involvement in disease. Larger TADs would have more chances to accommodate many genes and select for enhancer sharing along evolution. Conclusions Genes associated with human disease do not distribute randomly over the TADs. Our observations suggest general rules that confer functional stability to TADs, adding more evidence to the role of TADs as regulatory units.