One-way ANOVA test for multi-group comparisons

Performs one-way ANOVA to test for differences in scores across multiple levels of a trait. Tukey post-hoc tests are used to identify pairwise differences and significance is visualized as annotated boxplots.

Usage

scores.anova.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 including the grouping variable.

trait

Character. Name of the column in coldata used for the grouping variable.

pval

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

Value

A list of significant features or NULL if none are significant.