BMix provides univariate Binomial and Beta-Binomial mixture models. Count-based mixtures can be used in a variety of settings, for instance to model genome sequencing data of somatic mutations in cancer. BMix fits these mixtures by maximum likelihood exploiting the Expectation Maximization algorithm. Model selection for the number of mixture components is by the Integrated Classification Likelihood, an extension of the Bayesian Information Criterion that includes the entropy of the latent variables.

If you use BMix, please cite:

• G. Caravagna, T. Heide, M.J. Williams, L. Zapata, D. Nichol, K. Chkhaidze, W. Cross, G.D. Cresswell, B. Werner, A. Acar, L. Chesler, C.P. Barnes, G. Sanguinetti, T.A. Graham, A. Sottoriva. Subclonal reconstruction of tumors by using machine learning and population genetics. Nature Genetics 52, 898–907 (2020).

You can install the released version of BMix from GitHub with:

# install.packages("devtools")
devtools::install_github("caravagnalab/BMix")

Giulio Caravagna. Cancer Data Science (CDS) Laboratory.