scINSIGHT for interpreting single-cell gene expression from biologically heterogeneous data

The study introduces scINSIGHT, a novel method for analyzing single-cell RNA sequencing data from biologically heterogeneous samples. It uses non-negative matrix factorization to identify both common and condition-specific gene expression patterns, aiding in the clustering of cells and functional annotation. scINSIGHT was tested on simulated and real datasets, including immune cells from melanoma patients and B cells from COVID-19 patients, demonstrating high accuracy in identifying cell clusters and capturing condition-specific gene modules. It outperformed other methods in sensitivity and consistency, revealing insights into cellular identities and functional differences across biological conditions. Despite being slower, scINSIGHT offers robust integration and interpretability, making it a valuable tool in various biomedical and clinical contexts.