scEvoNet

scEvoNet predicts cell type evolution from single-cell RNA sequencing (scRNA-seq) datasets to identify gene sets and relationships underlying changes in cell states across species and cancer.

Key Features:

• Cross-species and cancer-related analysis: Handles cross-species comparisons and cancer-related scRNA-seq datasets for comparative cellular analyses.
• Confusion matrix construction: Constructs a confusion matrix to map relationships among cell states derived from scRNA-seq data.
• Bipartite gene–cell state network: Builds a bipartite network linking genes to cell states to visualize and analyze gene–state associations.
• Shared gene-set identification: Identifies sets of genes shared by characteristic signatures of two distinct cell states, including across distantly related datasets, to indicate divergence or co-option.
• Initial screening and similarity measurement: Performs initial screening for genes involved in cell state transitions and quantifies similarity between different cell states.
• Implementation: Implemented as a Python package.

Scientific Applications:

• Evolutionary biology: Explores continua of transcriptome states across developmental stages and species to study molecular mechanisms of cell-type evolution.
• Cancer research: Analyzes cancer-related scRNA-seq datasets to identify genetic changes and reused programs associated with tumor evolution.

Methodology:

Uses single-cell RNA sequencing (scRNA-seq) data to compute a confusion matrix of cell states, construct a bipartite network linking genes to cell states, identify shared gene signatures between cell states across datasets, and measure state similarity to detect signatures indicative of evolutionary divergence or co-option.