This R package, called trac, performs tree-based aggregation of compositional data with a particular focus on microbiome data. Users are encouraged to visit the package's website and a vignette. It implements the method proposed in Tree-Aggregated Predictive Modeling of Microbiome Data. It uses internally the c-lasso solver in Python (described in this paper). The package also includes a fast solver for standard sparse log-contrast regression.

The easiest way to install trac is by using the devtools R package (if not already installed, open R and run install.packages("devtools")).

To install trac, run

devtools::install_github("jacobbien/trac")

in R.