nem (version 2.46.0)

nem.discretize: Discretize perturbation data according to control experiments

Description

discretizes raw data to define effects of interventions with respect to wildtype/control measurements

Usage

nem.discretize(D,neg.control=NULL,pos.control=NULL,nfold=2,cutoff=0:10/10, pCounts=20, 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

cutoff

a (vector of) cutoff value(s) weighting the pos. controls versus the neg. controls. Default: 0:10/10

pCounts

pseudo-counts to guard against unreasonable low error estimates

empPval

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

verbose

Default: TRUE

Value

dat

discretized data matrix

pos

discretized positive controls [in the two-controls setting]

neg

discretized negative controls [in the two-controls setting]

sel

effect reporters selected [in the two-controls setting]

cutoff

error rates for different cutoff values [in the two-controls setting]

para

estimated error rates [in the two-controls setting]

Details

Chooses cutoff such that separation between negative and positive controls becomes optimal.

References

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

See Also

BoutrosRNAi2002

Examples

   # discretize Boutros data as in
   # Markowetz et al, 2005
   data("BoutrosRNAi2002")
   disc <- nem.discretize(BoutrosRNAiExpression,neg.control=1:4,pos.control=5:8,cutoff=.7)
   stopifnot(disc$dat==BoutrosRNAiDiscrete[,9:16])