scaDA: A novel statistical method for differential analysis of single-cell chromatin accessibility sequencing data

Area of Science:

• Genomics
• Computational Biology
• Epigenetics

Background:

• Single-cell ATAC-seq (scATAC-seq) measures chromatin accessibility at a single-cell resolution.
• Differential chromatin accessibility (DA) analysis is crucial for scATAC-seq data interpretation.
• Existing methods struggle with scATAC-seq data's high zero counts and variability, often focusing only on mean differences.

Purpose of the Study:

• To develop a novel statistical method for differential distribution analysis of scATAC-seq data.
• To address the limitations of existing methods by considering abundance, prevalence, and dispersion.
• To improve the power and accuracy of DA analysis in scATAC-seq studies.

Main Methods:

• Introduced scaDA, a composite statistical test based on the zero-inflated negative binomial (ZINB) model.
• scaDA jointly tests for differences in chromatin accessibility abundance, prevalence, and dispersion.
• Employed dispersion shrinkage and iterative refinement for parameter estimation.

Main Results:

• scaDA demonstrated superior performance in comprehensive simulation studies, achieving higher power and better false discovery rate (FDR) control than existing methods.
• The method showed the highest power in three real sc-multiome datasets.
• scaDA identified key differentially accessible regions in microglia for an Alzheimer's disease study, enriched for neurogenesis and AD-related pathways.

Conclusions:

• scaDA provides a robust and powerful approach for differential distribution analysis of scATAC-seq data.
• The method effectively handles the unique characteristics of scATAC-seq data, including excess zeros and variability.
• scaDA facilitates the discovery of biologically relevant genomic regions in complex diseases like Alzheimer's.