Rust version of the GSVA algorithm — rs

Rust-based implementation of the popular GSVA algorithm. Has further performance optimisations compared to the original implementation.

Usage

rs_gsva(exp, gs_list, tau, kernel, max_diff, abs_rank, timings)

exp: Numerical matrix. The expression matrix with rows = genes, and columns = samples
gs_list: List. A list containing the 0-based indices of the pathway genes. See rs_prepare_gsva_gs().
tau: Float. Tau parameter, usual recommendation is to use 1.0 here. Larger values emphasise the tails more.
kernel: String. One of c("gaussian", "poisson", "none"). The kernel function to use. Unknown strings default to "gaussian".
max_diff: Boolean. Scoring mode: TRUE = difference, FALSE = larger absolute value
abs_rank: Boolean. If TRUE = pos-neg, FALSE = pos+neg
timings: Boolean. Prints timings from the algorithm.

Returns a matrix of gene set ES scores x samples.