Local assortativity as a chromatin network analysis tool

Abstract

Using network analysis in the study of chromatin 3D structure has proved to be a powerful and useful approach. In order to determine which factors bound to DNA might play a role in establishing these 3D chromatin structure, the concept of chromatin assortativity (ChAs) was introduced. ChAs measures to what extent one specific feature of a chromatin fragment (node in the network) is shared by the nodes that interact preferentially with it. If one feature is localized in some specific region, it suggests that this feature might be playing a role in mediating genomic contacts in that region. Here we propose a Local assortativity approach to study a Mouse Embryonic Stem Cell (mESC) Promoter Capture Hi-C (PCHi-C) chromatin network, studying 78 different chromatin features to get a deeper biological insight compared to the global assortativity approach already described. We propose a different method of computing global assortativity by defining subnetworks, and two methods for computing local assortativity: based on genomic distances (1D) and based on shortest path distance between nodes (3D).