svapls (version 1.0)

svapls-package: Surrogate variable analysis using Partial Least Squares in a gene expression data

Description

The package svapls contains functions that are intended for the identification, correction and visualization of the hidden variability owing to a variety of unknown subject/sample specific effects of residual heterogeneity in a gene expression data.

Details

ll{ Package: svapls Type: Package Version: 1.0 Date: 2012-07-23 License: GPL-3 } The package can be used to find the genes that are truly differentially expressed between two types of samples (tissue types, biological conditions like Cancer/Non- Cancer samples, etc.), after adjusting for the hidden factors of residual heterogeneity in the data. The function svpls detects the truly positive genes after correcting for the hidden variation and also provides a modified gene expression matrix which is free from the spurious effects of the residual expression heterogeneity. Another important function hfp produces a heat- map representing the intensity of latent variability due to the unknown sample- specific factors, for any specified set of genes and subjects. fitModel, svpls and hfp

References

Sutirtha Chakraborty, Somnath Datta and Susmita Datta. (2012) Surrogate Variable Analysis Using Partial Least Squares in Gene Expression Studies. Bioinformatics.

Examples

data(batch.dat)
fit <- svpls(10,10,batch.dat,pmax = 5)
fit$genes
Y.corrected <- fit$Y.corr

data(hidden_fac.dat)
gen <- c(1:15,50:65)
sub <- c(1:5,11:17)

hfp(fit,gen,sub,hidden_fac.dat)