Related Experiment Video
Updated: Oct 17, 2025
10:15
Capturing Chromosome Conformation Across Length Scales
Published on: January 20, 2023
3.7K
Multiscale and integrative single-cell Hi-C analysis with Higashi
Ruochi Zhang1, Tianming Zhou1, Jian Ma2
1Computational Biology Department, School of Computer Science, Carnegie Mellon University, Pittsburgh, PA, USA.
Nature Biotechnology
|October 12, 2021
Summary
Higashi, a new algorithm, enhances the analysis of single-cell Hi-C data by improving 3D chromatin organization imputation. It reveals cell-to-cell variability in genome structure and its links to gene regulation.
Area of Science:
- Genomics
- Computational Biology
- Epigenetics
Background:
- Single-cell Hi-C (scHi-C) offers insights into 3D chromatin organization variability.
- Data sparseness in scHi-C presents significant analytical challenges.
- Existing methods struggle with accurate imputation and feature identification.
Purpose of the Study:
- To develop an advanced algorithm for analyzing scHi-C data.
- To improve the imputation of contact maps and identify 3D genome features.
- To integrate multimodal single-cell omics data for enhanced analysis.
Main Methods:
- Developed Higashi, a hypergraph representation learning algorithm.
- Applied Higashi to scHi-C data for contact map imputation.
- Integrated epigenomic signals into the hypergraph framework for multimodal analysis.
Main Results:
- Higashi outperforms existing methods in embedding and imputation of scHi-C data.
- Identified multiscale 3D genome features (compartments, TAD boundaries) and their cell-to-cell variability.
- Demonstrated improved embeddings for single-nucleus methyl-3C data by integrating epigenomic signals.
- Revealed connections between 3D genome features and cell-type-specific gene regulation in human prefrontal cortex.
Conclusions:
- Higashi effectively addresses scHi-C data sparseness and enhances 3D genome organization analysis.
- The algorithm enables refined delineation of cell-to-cell variability in chromatin structure.
- Higashi's framework shows potential for multimodal single-cell omics data analysis, including multiway interactions.
