Algorithmic considerations when analysing capture Hi-C data

Linden Disney-Hogg; Ben Kinnersley; Richard Houlston

doi:10.12688/wellcomeopenres.16394.2

Algorithmic considerations when analysing capture Hi-C data

Wellcome Open Res. 2022 Nov 11:5:289. doi: 10.12688/wellcomeopenres.16394.2. eCollection 2020.

Authors

Linden Disney-Hogg^{1

2}, Ben Kinnersley¹, Richard Houlston¹

Affiliations

¹ Division of Genetics and Epidemiology, The Institute of Cancer Research and The Royal Marsden NHS Foundation Trust, Sutton, Surrey, SM2 5NG, UK.
² Present address: School of Mathematics, University of Edinburgh, Edinburgh, EH9 3FD, UK.

Abstract

Chromosome conformation capture methodologies have provided insight into the effect of 3D genomic architecture on gene regulation. Capture Hi-C (CHi-C) is a recent extension of Hi-C that improves the effective resolution of chromatin interactions by enriching for defined regions of biological relevance. The varying targeting efficiency between capture regions, however, introduces bias not present in conventional Hi-C, making analysis more complicated. Here we consider salient features of an algorithm that should be considered in evaluating the performance of a program used to analyse CHi-C data in order to infer meaningful interactions. We use the program CHICAGO to analyse promoter capture Hi-C data generated on 28 different cell lines as a case study.

Keywords: Cancer; Capture Hi-C; Model assessment.