Identifying Topologically Associating Domain Structures from Single-Cell Hi-C Data

• Activity Number: 32 - Computational and Statistical Methods for Single-Cell Transcriptomics and Epigenomics Analyses
• Type: Topic Contributed
• Date/Time: Monday, August 3, 2020 : 10:00 AM to 11:50 AM
• Sponsor: Section on Statistics in Genomics and Genetics
• Author(s): Qunhua Li* and Qiuhai Zeng and Xi He and Yu Zhang and Tracey Zheng
• Companies: Penn State University and Wuhan University and Pennsylvania State University and Two sigmas and Pennsylvania State University
• Keywords: single cell; Hi-C; topologically association domain; TAD caller; 3D genome; chromatin organization

Abstract

Single-cell Hi-C technology makes it possible to probe cell-to-cell variation and understand the dynamics of 3D chromatin organization. One primary task in Hi-C analysis is to identify topologically associating domain (TAD) structures, as they usually represent biologically meaningful interactions. Though many bulk-cell TAD calling methods are available, they do not perform well when they are applied to single-cell Hi-C data, due to the extreme sparsity of the single-cell data. In this work, we developed a single-cell TAD caller (jOnTAD), extended from a bulk-cell caller we developed, that can identify hierarchical TAD structures for single cell Hi-C data and reduce spurious identifications by borrowing strength from different cells. Our model produces biologically much more meaningful results that existing TAD callers.