In Alexandrov et al. develop a framework for learning the mutational processes operative in cancer genomes. This framework builds on a method called non-negative matrix factorization. In this talk, I will discuss the developed framework and some of the issues (not treated by Alexandrov et al.) that arise from the use of this method. These issues include the formulation of a statistical model for the setup, model checking and non-uniques of the model. I will also give a short introduction to mutational processes in cancer to demonstrate the application of the model.

Reference:

L.B. Alexandrov , S. Nik-Zainal , D.C. Wedge, P.J. Campbell, and M.R. Stratton: Deciphering signatures of mutational processes operative in human cancer, Cell Reports 3, 246-259 (2013).