sva.id {eigenR2}R Documentation

Estimate the number of significant eigenvectors to include in analysis

Description

Estimate the number of significant eigenvectors from the residuals of a high-dimensional data set given the model fit.

Usage

  sva.id(dat, mod, B=20, eigen.sig=0.10, seed=NULL)

Arguments

dat A m response variables by n samples matrix of data
mod A n by k model matrix corresponding to the primary model fit, if mod is NULL, we estimate the number of significant eigenvectors in the expression data. (see model.matrix)
B The number of null iterations to perform.
eigen.sig The significance cutoff for eigenvectors.
seed A numeric seed for reproducible results. Optional.

Details

Note that this function is a modified function from the package sva by Leek JT and Storey JD. The model matrix should include a column for an intercept. sva.id estimates the number of surrogate variables to include in the analysis as described in Leek and Storey (2007).

Value

A list containing:

n.sv The number of significant surrogate variables
p.sv The p-values for each eigenvector

Author(s)

Jeff T. Leek jtleek@gmail.com and John D. Storey jstorey@princeton.edu

References

Leek JT and Storey JD (2007) Capturing heterogeneity in gene expression studies by "surrogate variable analysis." PLoS Genetics, 3:e161.

See Also

eigenR2


[Package eigenR2 version 1.0 Index]