nem.discretize {nem}R Documentation

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

Details

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

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]

Note

preliminary! will be developed to be more generally applicable

Author(s)

Florian Markowetz <URL: http://genomics.princeton.edu/~florian>

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])   

[Package nem version 2.2.1 Index]