Singular Value Decomposition wrapper for supervised PCA

Center and compute the SVD of a matrix

mysvd(mat, method = svd, n.components = NULL)

Arguments

mat	A matrix of data frame in "tall" format (\(p \times n\)).
method	What function should be used to extract the left- and right- singular vectors and singular values? Any function that returns the values as a list with components u, v, and d is appropriate. Defaults to svd.
n.components	How many singular values / vectors to return? Must be an integer less than \(min(p, n)\). Best performance increase is for values much less than \(min(p, n)\). Defaults to NULL.

Value

A list containing:

• u : The first n.components left singular vectors of mat.

• d : The largest n.component singular values of mat.

• v : The first n.components right singular vectors of mat.

• feature.means : A named vector of the feature means of mat.

Details

The mysvd function takes in a tall -Omics data matrix, extracts the feature means, centers the matrix on this mean vector, and calculates the Singular Value Decomposition (SVD) of the centered data matrix. Currently, the SVD is calculated via the fast.svd function from corpcor package. However, this function calculates all the singular vectors, even when n.components is non-NULL. We should experiment with other SVD functions, such as the rsvd function from the rsvd package. ENHANCEMENT.

Examples

  # DO NOT CALL THIS FUNCTION DIRECTLY.
  # Use SuperPCA_pVals() instead

if (FALSE) {
  data("colon_pathwayCollection")
  data("colonSurv_df")

  colon_OmicsSurv <- CreateOmics(
    assayData_df = colonSurv_df[,-(2:3)],
    pathwayCollection_ls = colon_pathwayCollection,
    response = colonSurv_df[, 1:3],
    respType = "surv"
  )

  asthmaGenes_char <-
    getTrimPathwayCollection(colon_OmicsSurv)[["KEGG_ASTHMA"]]$IDs

  mysvd(t(getAssay(colon_OmicsSurv))[asthmaGenes_char, ])
}