nem (version 2.46.0)

nem.cont.preprocess: Calculate classification probabilities of perturbation data according to control experiments

Description

Calculates probabilities of data to define effects of interventions with respect to wildtype/control measurements

Usage

nem.cont.preprocess(D,neg.control=NULL,pos.control=NULL,nfold=2, influencefactor=NULL, empPval=.05, verbose=TRUE)

Arguments

D

matrix with experiments as columns and effect reporters as rows

neg.control

either indices of columns in D or a matrix with the same number of rows as D

pos.control

either indices of columns in D or a matrix with the same number of rows as D

nfold

fold-change between neg. and pos. controls for selecting effect reporters. Default: 2

influencefactor

factor multiplied onto the probabilities, so that all negative control genes are treated as influenced, usually automatically determined

empPval

empirical p-value cutoff for effects if only one control is available

verbose

Default: TRUE

Value

dat

data matrix

pos

positive controls [in the two-controls setting]

neg

negative controls [in the two-controls setting]

sel

effect reporters selected [in the two-controls setting]

prob.influenced

probability of a reporter to be influenced

influencefactor

factor multiplied onto the probabilities, so that all negative control genes are treated as influenced

Details

Determines the empirical distributions of the controls and calculates the probabilities of pertubartion data to belong to the control distribution(s).

References

Markowetz F, Bloch J, Spang R, Non-transcriptional pathway features reconstructed from secondary effects of RNA interference, Bioinformatics, 2005

See Also

BoutrosRNAi2002

Examples

   data("BoutrosRNAi2002")
   preprocessed <- nem.cont.preprocess(BoutrosRNAiExpression,neg.control=1:4,pos.control=5:8)