Wilcoxon rank-sum test for binary traits

Performs a Wilcoxon rank-sum (Mann-Whitney U) test comparing scores between two levels of a binary clinical trait. Significant features are plotted as boxplots and saved to the "Results/" folder.

Usage

scores.wilcox.test(scores, coldata, trait, pval = 0.05)

Arguments

scores

A list, NMF output from compute.latent_factors(), or a score matrix. When a list, the first element must be a samples x features score matrix.

coldata

A data frame containing sample annotations and clinical traits.

trait

Character. Name of the column in coldata used as the binary grouping variable.

pval

Numeric. P-value threshold for significance (default = 0.05).

Value

A list containing significant features or NULL if none are found.