SelGenes: a tool for selecting marker genes in heterogeneous samples

Abstract

SelGenes is a tool for selecting marker genes for the dominating cell type in heterogeneous samples. Based on a framework from an existing algorithm, SelGenes selects cell-type specific marker genes for the dominating cell-type and uses these marker genes to estimate cell-type proportions in the sample and cell-type specific expression proles. We test the performance of SelGenes on a benchmark set and validate the marker genes against an external database and further apply SelGenes to a real data set containing gene expression data from cancer samples. Compared to an existing method the results from the test were consistently better for SelGenes.