HiCRep: assessing the reproducibility of Hi-C data using a stratum-adjusted correlation coefficient

Area of Science:

• Genomics
• Molecular Biology
• Bioinformatics

Background:

• Hi-C technology enables genome-wide chromatin interaction studies.
• Existing reproducibility assessments for Hi-C data are limited by ignoring spatial features like domain structure and distance dependence.
• These limitations can lead to misleading evaluations of Hi-C data quality.

Purpose of the Study:

• To introduce HiCRep, a novel framework for assessing Hi-C data reproducibility.
• To develop a new similarity measure, the stratum adjusted correlation coefficient (SCC), that accounts for spatial features.
• To provide a statistically sound, interpretable, and scalable quality control method for Hi-C data.

Main Methods:

• Development of the stratum adjusted correlation coefficient (SCC) for quantifying similarity between Hi-C interaction matrices.
• Implementation of the HiCRep framework in an R package.
• Systematic accounting for spatial features (domain structure, distance dependence) in Hi-C data.

Main Results:

• HiCRep and SCC provide a statistically sound and reliable evaluation of Hi-C data reproducibility.
• SCC demonstrates higher accuracy than existing methods in distinguishing subtle reproducibility differences and depicting cell lineage interrelationships.
• The framework facilitates quantification of differences between Hi-C contact matrices and determination of optimal sequencing depth.

Conclusions:

• HiCRep offers a superior approach to Hi-C data quality control by incorporating spatial features.
• The SCC measure provides a standardized, interpretable, automatable, and scalable method for assessing reproducibility.
• The freely available HiCRep R package enables widespread adoption of this improved methodology.