Related Experiment Video
Updated: Dec 15, 2025
10:15
Capturing Chromosome Conformation Across Length Scales
Published on: January 20, 2023
4.0K
FreeHi-C spike-in simulations for benchmarking differential chromatin interaction detection
Ye Zheng1, Peigen Zhou2, Sündüz Keleş3
1Biostatistics, Bioinformatics and Epidemiology Program, Vaccine and Infectious Disease Division, Fred Hutchinson Cancer Research Center, Seattle, WA 98109, USA; Department of Statistics, University of Wisconsin - Madison, Madison, WI 53706, USA.
Methods (San Diego, Calif.)
|July 15, 2020
Summary
We developed FreeHi-C with a spike-in module to benchmark differential chromatin interaction detection methods. This benchmarking revealed varying performances and highlighted a lack of power in small replication settings for all tested methods.
Area of Science:
- Genomics
- Bioinformatics
- Computational Biology
Background:
- High-throughput genome-wide chromatin conformation capture (Hi-C) assays are vital for studying genome 3D organization and long-range interactions.
- Comparative Hi-C analysis is crucial for understanding genomic changes across conditions, but methods for detecting differential interactions are less understood.
Purpose of the Study:
- To extend FreeHi-C with a spike-in module for benchmarking differential chromatin interaction detection methods.
- To provide a robust framework for evaluating method performance when ground truth is known.
Main Methods:
- Developed a user/data-driven spike-in module for FreeHi-C.
- Benchmarked four differential chromatin interaction detection methods (HiCcompare, multiHiCcompare, diffHic, Selfish) using FreeHi-C simulations.
- Evaluated methods across varying sequencing depths and spike-in proportions.
Main Results:
- Distinguished performances were observed among the four methods based on standard metrics like FDR control, detection power, and ROC/PR curves.
- The study identified specific genomic properties of detectable differential interactions for each method.
- All methods demonstrated limited power in small replication settings.
Conclusions:
- FreeHi-C with its spike-in module offers a valuable tool for benchmarking differential interaction detection methods.
- The benchmarking highlights the need for improved methods, especially for studies with limited replicates.
- Understanding method-specific performance is critical for accurate comparative Hi-C analyses.
