mahmoudibrahim/genesorteR: Feature Ranking in Clustered Single Cell Data

The main purpose of this R extension is to select features in (possibly very large) single cell data including scRNA-Seq and scATAC-Seq. The main idea is that the dropout rate of a gene is a good measure of its expression, and that empirical statistics calculated based on binarized expression matrices are sufficient to select marker genes in a way that is consistent with the expected definition of "marker gene" in experimental biology research. It can provide a ranking of genes specificity in each cell cluster, as well as select large or small sets of marker genes by a permutation test or using entropy-based feature selection. To assess cell clustering quality, some functions can also compute cell cluster quality metrics.

install.packages("remotes")
remotes::install_github("mahmoudibrahim/genesorteR")