The goal of mSigHdp is mutational signature discovery using hierarchical Dirichlet process (HDP) mixture models. mSigHdp is only supported on Linux systems. Most users will use the function RunHdpxParallel.
This package uses https://github.com/steverozen/hdpx for the hierarchical Dirichlet process implementation.
Please also see our paper: Mo Liu, Yang Wu, Nanhai Jiang, Arnoud Boot, Steven G. Rozen, mSigHdp: hierarchical Dirichlet process mixture modeling for mutational signature discovery, https://doi.org/10.1093/nargab/lqad005.
singularity pull library://rozen-lab/msighdp/msighdp:2.1.2
For more details, see https://github.com/steverozen/mSigHdp/blob/master/data-raw/container_scripts/mSigHdp_container_installation.pdf
A toy-example R script for using this container is available at https://github.com/steverozen/mSigHdp/blob/v2.1.2-branch/data-raw/container_scripts/test_mSigHdp.R.
install.packages("remotes")
remotes::install_github(repo = "steverozen/mSigHdp", ref = "v2.1.2-branch", build_vignettes = TRUE)To use new features in the development version, you can install mSigHdp from the master branch on GitHub, which may not be stable:
install.packages("remotes")
remotes::install_github(repo = "steverozen/mSigHdp", ref = "master")https://github.com/steverozen/mSigHdp/blob/v2.1.2-branch/mSigHdp_2.1.2.pdf