cit_gsa: gene set analysis for scRNA-seq with covariate adjustment

Abstract

Single-cell RNA-seq enables gene expression analysis at single-cell resolution. Differential Expression Analysis (DEA) performed at the gene set level for better power and interpretation, identifies gene sets associated with a factor of interest. Existing single-cell DEA methods rely on two-step approaches, a gene-level analysis followed by an enrichment step, are often limited to two-group comparisons and do not support covariate adjustment. We introduce cit_gsa, a novel one-step self-contained gene set DEA method for Single-cell RNA-seq that handles more than two conditions and allows covariate adjustment. A benchmark of cit_gsa against state-of-the-art methods shows that our method has better performance in complex scenarios, and real-data analysis confirms that covariate adjustment improves biological interpretation.