scCASE: accurate and interpretable enhancement for single-cell chromatin accessibility sequencing data

Area of Science:

• Genomics
• Epigenetics
• Computational Biology

Background:

• Single-cell chromatin accessibility sequencing (scCAS) is crucial for understanding epigenomic heterogeneity and gene regulation.
• scCAS data presents challenges like high sparsity and dimensionality, hindering downstream analysis.
• Existing scCAS data enhancement methods have limitations.

Purpose of the Study:

• To propose scCASE, an effective method for enhancing scCAS data using non-negative matrix factorization and an iterative cell-to-cell similarity matrix.
• To introduce scCASER, an extension of scCASE that incorporates external omics data for improved enhancement.
• To demonstrate the advantages of scCASE and scCASER over existing methods.

Main Methods:

• Developed scCASE, a novel enhancement method based on non-negative matrix factorization with an iteratively updating cell-to-cell similarity matrix.
• Expanded scCASE to scCASER to integrate external reference omics data.
• Conducted comprehensive experiments on multiple scCAS datasets.

Main Results:

• scCASE significantly outperforms existing methods in scCAS data enhancement.
• Identified interpretable cell type-specific peaks using scCASE, offering valuable biological insights into cell subpopulations.
• scCASER further improves enhancement performance by leveraging external reference data.

Conclusions:

• scCASE and scCASER provide effective solutions for addressing sparsity and dimensionality in scCAS data.
• These methods enhance the analysis of epigenomic heterogeneity and gene regulation.
• The identified cell type-specific peaks offer new avenues for biological discovery.