Calculate the contrastive PCA — rs_contrastive

This function calculates the contrastive PCA given a target covariance matrix and the background covariance matrix you wish to subtract. The alpha parameter controls how much of the background covariance you wish to remove. You have the options to return the feature loadings and you can specify the number of cPCAs to return.

Usage

rs_contrastive_pca(
  target_covar,
  background_covar,
  target_mat,
  alpha,
  n_pcs,
  return_loadings
)

Arguments

target_covar: The co-variance matrix of the target data set.
background_covar: The co-variance matrix of the background data set.
target_mat: The original values of the target matrix. Rows = samples, columns = features.
alpha: How much of the background co-variance should be removed.
n_pcs: How many contrastive PCs to return
return_loadings: Shall the loadings be returned from the contrastive PCA

Value

A list containing:

factors - The factors of the contrastive PCA, i.e. target_mat multiplied by the loadings. Samples x n_pcs.
loadings - The loadings (top eigenvectors) of the contrastive PCA. Features x n_pcs. Will be NULL if return_loadings = FALSE.