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

TL;DR

This paper introduces scGeom, a novel computational tool that uses geometric and topological analysis of high-dimensional single-cell RNA sequencing data to better understand cell states, transitions, and classifications.

Contribution

scGeom leverages graph curvature and persistent homology to analyze multiscale structures in scRNA-seq data, revealing biological insights beyond traditional methods.

Findings

Structural features reflect biological properties and functions.

Topological signatures help identify transition cells.

Improves accuracy of cell type classification.

Abstract

Single-cell RNA sequencing (scRNA-seq) enables dissecting cellular heterogeneity in tissues, resulting in numerous biological discoveries. Various computational methods have been devised to delineate cell types by clustering scRNA-seq data where the clusters are often annotated using prior knowledge of marker genes. In addition to identifying pure cell types, several methods have been developed to identify cells undergoing state transitions which often rely on prior clustering results. Present computational approaches predominantly investigate the local and first-order structures of scRNA-seq data using graph representations, while scRNA-seq data frequently displays complex high-dimensional structures. Here, we present a tool, scGeom for exploiting the multiscale and multidimensional structures in scRNA-seq data by inspecting the geometry via graph curvature and topology via persistent homology of both cell networks and gene networks. We demonstrate the utility of these structural features for reflecting biological properties and functions in several applications where we show that curvatures and topological signatures of cell and gene networks can help indicate transition cells and developmental potency of cells. We additionally illustrate that the structural characteristics can improve the classification of cell types.