Topological and geometric analysis of cell states in single-cell transcriptomic data

Area of Science:

• Computational Biology
• Genomics
• Bioinformatics

Background:

• Single-cell RNA sequencing (scRNA-seq) is crucial for understanding cellular heterogeneity.
• Current computational methods often rely on clustering and marker genes, primarily analyzing local data structures.
• Identifying cells in transition states typically depends on pre-existing clustering outcomes.

Purpose of the Study:

• To introduce scGeom, a novel computational tool for scRNA-seq data analysis.
• To leverage geometric and topological features for a deeper understanding of cellular states and heterogeneity.
• To enhance cell type classification and identification of transitional cell populations.

Main Methods:

• scGeom analyzes multiscale and multidimensional structures within scRNA-seq data.
• It employs curvature and persistent homology on both cell and gene networks.
• The method investigates the inherent geometry and topology of the data.

Main Results:

• Geometric and topological features effectively reflect biological properties and functions.
• Curvatures and topological signatures can identify transition cells and their differentiation potential.
• Structural characteristics derived from scGeom improve cell type classification accuracy.

Conclusions:

• scGeom offers a new perspective on scRNA-seq data analysis by exploring complex high-dimensional structures.
• The tool's ability to analyze network geometry and topology enhances biological insights.
• scGeom provides a valuable method for dissecting cellular heterogeneity and dynamics.