Characterization of cellular heterogeneity in the developing kidney and in pediatric tumors using unsupervised machine learning

Single-cell RNAseq measurements are useful for describing the cellular heterogeneity between different cells within the same tissue or tumor. Likewise, “bulk” RNAseq measurements are useful for describing the heterogeneity between tumors from different individuals. However, in many cases the cells or tumors differ from each other in a continuous manner that is difficult to quantify, characterize, and interpret. Here, I will demonstrate the use of un-supervised machine learning techniques for characterizing the continuous heterogeneity between individual cells in the developing fetal kidney and between pediatric kidney tumors from different patients. I will also demonstrate the use of these techniques for “deconvolving” hidden processes in a simple model system consisting of synchronized HeLa cells recovering from cell cycle arrest.