2.1. Preprocessing of Single-Cell Datasets

Datasets were preprocessed with QuanTC to select a number of top expressed genes (n = 3000) and discard cells that express less than a minimal percentage of genes (5%). QuanTC preprocessing selected n = 361 cells for the SCC dataset, n = 665, 1772, 2794 cells for OVCA420 under TGF-beta, EGF, and TNF induction, respectively, n = 2034 and 2239 cells in the inner and outer MCF10A datasets, respectively, and n = 2215 cells in the HNSCC dataset. For all datasets, a pseudo count was added before log-transforming the count matrix.