Comparison of computational methods for Hi-C data analysis.

Abstract

Hi-C is a genome-wide sequencing technique to investigate the 3D chromatin conformation inside the nucleus. The most studied structures that can be identified from Hi-C - chromatin interactions and topologically associating domains (TADs) - require computational methods to analyze genome-wide contact probability maps. We quantitatively compared the performances of 13 algorithms for the analysis of Hi-C data from 6 landmark studies and simulations. The comparison revealed clear differences in the performances of methods to identify chromatin interactions and more comparable results of algorithms for TAD detection.