Clique-Based Topological Characterization of Chromatin Interaction Hubs

Chromatin conformation capture technologies are a vital source of information about the spatial organization of chromatin in eukaryotic cells. Of these technologies, Hi-C and related methods have been widely used to obtain reasonably complete contact maps in many cell lines and tissues under a wide variety of conditions. This data allows for the creation of chromatin interaction graphs from which topological generalizations about the structure of chromatin may be drawn. Here we outline and utilize a clique-based approach to analyzing chromatin interaction graphs which allows for both detailed analysis of strongly interconnected regions of chromatin and the unraveling of complex relationships between genomic loci in these regions. We find that clique-rich regions are significantly enriched in distinct gene ontologies as well as regions of transcriptional activity compared to the entire set of links in the respective datasets, and that these cliques are also not entirely preserved in randomized Hi-C data. We conclude that cliques and the denser regions of connectivity in which they are common appear to indicate a consistent pattern of chromatin spatial organization that resembles transcription factories, and that cliques can be used to identify functional modules in Hi-C data.