Inference of marker genes of subtle cell state changes via iterative logistic regression

We present iterative logistic regression (iLR) for the identification of small sets of informative marker genes. Differential expression and marker gene selection methods for single-cell RNA sequencing (scRNA-seq) data can struggle to identify small sets of informative genes, especially for subtle differences between cell states, as can be induced by disease or treatment. iLR applied logistic regression iteratively with a Pareto front optimization to balance gene set size with classification performance. We benchmark iLR on in silico datasets demonstrating comparable performance to the state-of-the-art at single-cell classification using only a fraction of the genes. We test iLR on its ability to distinguish neuronal cell subtypes in healthy vs. autism spectrum disorder patients and find it achieves high accuracy with small sets of disease-relevant genes. We apply iLR to investigate immunotherapeutic effects in cell types from different tumor microenvironments and find that iLR infers informative genes that translate across organs and even species (mouse-to-human) comparison. We predicted via iLR that entinostat acts in part through the modulation of myeloid cell differentiation routes in the lung microenvironment. Overall, iLR provides means to infer interpretable transcriptional signatures from complex datasets with prognostic or therapeutic potential.