Genome wide clustering on integrated chromatin states and Micro-C contacts reveals chromatin interaction signatures

Abstract

Chromatin states based on various histone modifications are a common annotation for genomes, and have been shown to correspond to regulatory functions such as enhancers and transcription start sites. With the advent of Hi-C and other chromatin conformation capture technologies, we now have the ability to analyze 3-dimensional physical interactions of chromatin regions, in addition to the 1-dimensional regulatory annotation, but methods to integrate this information are lacking. We present a framework for integrating the chromatin state of interacting regions into a numeric vector through the contact-weighted sum of chromatin states. Unsupervised clustering on integrated chromatin states and Micro-C contacts reveals common patterns of chromatin interaction which we call chromatin interaction signatures. Our results indicate that most chromatin interaction signatures are found in all four cell types investigated here. Between 12-40% of the regions change chromatin interaction signatures between the cell types despite maintaining chromatin state, hinting at the dynamic nature of chromatin conformation. Although regions with similar chromatin states are often in contact as expected, subcategories of enhancers and transcription start sites have distinct chromatin interaction signatures that are associated with gene expression. The integrated information on the chromatin states that the region is in contact with adds another layer of annotation beyond chromatin state or Hi-C contact separately. In summary, we present the first set of chromatin interaction signatures for the human genome that provides an integrated view of the complex dynamics of concurrent change occurring in chromatin state and in chromatin interaction.