epiScanpy

epiScanpy provides a computational framework to analyze single-cell DNA methylation and ATAC-seq data by adapting Scanpy RNA-seq workflows to extract regulatory information complementary to transcriptomics.

Key Features:

• Integration with Scanpy Workflows: Integrates Scanpy RNA-seq workflows to apply clustering, dimension reduction, and trajectory learning to single-cell DNA methylation and ATAC-seq data.
• Feature Space Construction: Implements multiple feature space constructions tailored for DNA methylation and ATAC-seq datasets to enable transcriptomics-style analyses on epigenetic data.
• Clustering and Dimension Reduction: Provides clustering and dimension reduction methods to identify cell subtypes and states from single-cell epigenomic profiles.
• Trajectory Learning: Supports trajectory learning on epigenetic data to investigate developmental processes and lineage relationships.

Scientific Applications:

• Benchmarking on mouse brain atlases: Benchmarked on single-cell mouse brain atlases containing DNA methylation, ATAC-seq, and transcriptomics datasets.
• Cell-type classification and marker discovery: Enhances cell type classification by incorporating orthogonal epigenetic information alongside transcriptome-based labels and identifies differentially methylated and differentially open markers between cell clusters.

Methodology:

Adapts Scanpy single-cell RNA-seq analysis techniques to epigenetic data, develops feature space constructions for DNA methylation and ATAC-seq, and applies clustering, dimension reduction, and trajectory learning methods.