This R package implements the TreeMHN model for the joint inference of exclusivity patterns and recurrent trajectories from tumor mutation trees. (bioRxiv preprint)
Quick start
TreeMHN takes as input a set of independent tumor mutation trees containing a total number of n mutations. The format is a dataframe with five columns:
-
Patient_ID: IDs of patients, unique for each patient; -
Tree_ID: IDs of mutation trees, unique within each patient; -
Node_ID: IDs of each node in the tree, including the root node (with ID "1"), unique for each node; -
Mutation_ID: IDs of each mutational event. The root node has a mutation ID of "0", and other mutation IDs can be duplicated in the tree to allow for parallel mutations; -
Parent_ID: IDs of the parent node ID. The root node has itself as parent (ID "1").
The output is an n-by-n matrix representing the Mutual Hazard Network. The diagonal entries of this matrix indicate how often each mutation will occur and fixate independent of the other mutations. The off-diagonal entries encode the exclusivity and co-occurrence patterns of mutations. Conditioned on the estimated matrix, we can compute the probabilities of different evolutionary trajectories or evaluate the most likely next mutational events given a tumor tree.
Please see Demo.md for more details.
Installation
For Mac users, please compile the package with g++ instead of clang. To do this, you need to first install gcc using Homebrew:
brew install gcc
Then, create ~/.R/Makevars with entry
CXX=$(brew --prefix)/bin/g++-[INSTALLED VERSION]
For all users, install the devtools package in R and run
devtools::install_github("cbg-ethz/TreeMHN")
The installation typically takes around one minute to finish.