scHi-CSim: a flexible simulator that generates high-fidelity single-cell Hi-C data for benchmarking

Area of Science:

• Genomics
• Epigenetics
• Computational Biology

Background:

• Single-cell Hi-C technology offers insights into individual cell chromatin organization.
• High sequencing costs restrict data depth and replication, hindering comprehensive analysis.
• Development of robust simulation tools is crucial for advancing single-cell Hi-C research.

Purpose of the Study:

• To develop a high-fidelity single-cell Hi-C simulator (scHi-CSim) for benchmarking.
• To enable generation of simulated data with adjustable sequencing depth and replicates.
• To facilitate the evaluation and improvement of single-cell Hi-C analysis methods.

Main Methods:

• scHi-CSim merges neighboring cells to address data sparseness.
• It samples interactions across distance-stratified chromosomes to preserve cellular heterogeneity.
• The simulator estimates empirical restriction fragment distributions for realistic data generation.

Main Results:

• scHi-CSim generates high-fidelity data, validated by comparable single-cell clustering and high-order structure detection to raw data.
• Simulated data quality is tunable regarding sequencing depth and replicate number.
• Increased sequencing depth in simulations improves the accuracy of topologically associating domain detection.

Conclusions:

• scHi-CSim provides a valuable tool for benchmarking single-cell Hi-C data analysis methods.
• The simulator aids in understanding the impact of sequencing depth on chromatin structure detection.
• scHi-CSim facilitates the development and validation of computational approaches for single-cell epigenomics.