AESPCA_pVals: Test pathway association with AES-PCA

In pathwayPCA: Integrative Pathway Analysis with Modern PCA Methodology and Gene Selection

Description

Given a supervised OmicsPath object (one of OmicsSurv, OmicsReg, or OmicsCateg), extract the first k adaptive, elastic-net, sparse principal components (PCs) from each pathway-subset of the features in the -Omics assay design matrix, test their association with the response matrix, and return a data frame of the adjusted p-values for each pathway.

Usage

AESPCA_pVals(
  object,
  numPCs = 1,
  numReps = 0L,
  parallel = FALSE,
  numCores = NULL,
  asPCA = FALSE,
  adjustpValues = TRUE,
  adjustment = c("Bonferroni", "Holm", "Hochberg", "SidakSS", "SidakSD", "BH", "BY",
    "ABH", "TSBH"),
  ...
)

## S4 method for signature 'OmicsPathway'
AESPCA_pVals(
  object,
  numPCs = 1,
  numReps = 1000,
  parallel = FALSE,
  numCores = NULL,
  asPCA = FALSE,
  adjustpValues = TRUE,
  adjustment = c("Bonferroni", "Holm", "Hochberg", "SidakSS", "SidakSD", "BH", "BY",
    "ABH", "TSBH"),
  ...
)

Arguments

object	An object of class OmicsPathway with a response matrix or vector.
numPCs	The number of PCs to extract from each pathway. Defaults to 1.
numReps	How many permutations to estimate the p-value? Defaults to 0 (that is, to estimate the p-value parametrically). If numReps > 0, then the non-parametric, permutation p-value will be returned based on the number of random samples specified.
parallel	Should the computation be completed in parallel? Defaults to FALSE.
numCores	If parallel = TRUE, how many cores should be used for computation? Internally defaults to the number of available cores minus 1.
asPCA	Should the computation return the eigenvectors and eigenvalues instead of the adaptive, elastic-net, sparse principal components and their corresponding loadings. Defaults to FALSE; this should be used for diagnostic or comparative purposes only.
adjustpValues	Should you adjust the p-values for multiple comparisons? Defaults to TRUE.
adjustment	Character vector of procedures. The returned data frame will be sorted in ascending order by the first procedure in this vector, with ties broken by the unadjusted p-value. If only one procedure is selected, then it is necessarily the first procedure. See the documentation for the ControlFDR function for the adjustment procedure definitions and citations.
...	Dots for additional internal arguments.

Details

This is a wrapper function for the ExtractAESPCs, PermTestSurv, PermTestReg, and PermTestCateg functions.

Please see our Quickstart Guide for this package: https://gabrielodom.github.io/pathwayPCA/articles/Supplement1-Quickstart_Guide.html

Value

A results list with class aespcOut. This list has three components: a data frame of pathway details, pathway p-values, and potential adjustments to those values (pVals_df); a list of the first numPCs score vectors for each pathway (PCs_ls); and a list of the first numPCs feature loading vectors for each pathway (loadings_ls). The p-value data frame has columns:

• pathways : The names of the pathways in the Omics* object (given in object@trimPathwayCollection$pathways.)

• setsize : The number of genes in each of the original pathways (given in the object@trimPathwayCollection$setsize object).

• n_tested : The number of genes in each of the trimmed pathways (given in the object@trimPathwayCollection$n_tested object).

• terms : The pathway description, as given in the object@trimPathwayCollection$TERMS object.

• rawp : The unadjusted p-values of each pathway.

• ... : Additional columns of adjusted p-values as specified through the adjustment argument.

The data frame will be sorted in ascending order by the method specified first in the adjustment argument. If adjustpValues = FALSE, then the data frame will be sorted by the raw p-values. If you have the suggested tidyverse package suite loaded, then this data frame will print as a tibble. Otherwise, it will print as a data frame.

See Also

CreateOmics; ExtractAESPCs; PermTestSurv; PermTestReg; PermTestCateg; TabulatepValues; clusterApply

Examples

  ###  Load the Example Data  ###
  data("colonSurv_df")
  data("colon_pathwayCollection")

  ###  Create an OmicsSurv Object  ###
  colon_Omics <- CreateOmics(
    assayData_df = colonSurv_df[, -(2:3)],
    pathwayCollection_ls = colon_pathwayCollection,
    response = colonSurv_df[, 1:3],
    respType = "surv"
  )

  ###  Calculate Pathway p-Values  ###
  colonSurv_aespc <- AESPCA_pVals(
    object = colon_Omics,
    numReps = 0,
    parallel = TRUE,
    numCores = 2,
    adjustpValues = TRUE,
    adjustment = c("Hoch", "SidakSD")
  )

pathwayPCA documentation built on Dec. 15, 2020, 6:14 p.m.