JOnTAD: Joint Hierarchical TAD Caller for High Resolution, Single Cell and Bulk Cell Hi-C Data

• Activity Number: 70 - Novel Approaches for Omics and Multi-Omics Analysis
• Type: Contributed
• Date/Time: Sunday, August 7, 2022 : 4:00 PM to 5:50 PM
• Sponsor: Section on Statistics in Genomics and Genetics
• Author(s): Qiuhai Zeng* and Qunhua Li
• Companies: Pennsylvania State Univeristy and Pennsylvania State University
• Keywords: TAD; single cell Hi-C; dynamic programming; MicroC

Abstract

The three-dimensional (3D) organization of the chromatin is critical for gene expression and genome replication. We propose Joint Optimized nested TAD (JOnTAD) model, based on dynamic programming and quantile regression, to identify hierarchical topologically associating domains (TADs) or TAD like structures for high resolution, single cell and bulk cell Hi-C data. It is able to explore TAD dynamics or similarity by borrowing strength across multiple contact matrices. We verified the model’s performance on deep sequencing data, ESC lineage data, mitosis-to-G1 phase transition data, single cell Hi-C data, and MicroC data.