PANR (version 1.18.0)

BetaMixture-class: An S4 class for beta mixture modelling of functional gene associations

Description

This S4 class includes methods to do beta-mixture modelling of functional gene associations given rich phenotyping screens.

Objects from the Class

Objects of class BetaMixture can be created from new("BetaMixture", metric, order, association, model, pheno, partition) (see the example below for details).

Slots

pheno:

a numeric matrix of rich phenotypes with rows and columns specifying genes and samples, respectively.

metric:

a character value specifying the metric to compute similarity scores. Currently, 'cosine' and 'correlation' are supported (see assoScore for more details).

order:

a numeric value specifying the order of the similarity score to be computed. Only 1 and 2 is supported for the current version. The first order (when order=1) similarities are used for quatification of the strength of functional associations between genes, whilst the second order (when code=2) ones are employed to compute the strength of modularity between genes.

association:

a numeric vector providing all association scores between genes. This can be useful when pheno is not available or the user has a different way to compute functional associations.

model:

a character value specifying whether the original (if global) or extended (if stratified) model is used.

partition:

a numeric of gene partition labels (e.g. c(rep(1, 100), rep(2, 20), rep(3, 80)) is a valid vector of partition labels for a vector of associations falling into three categories of interaction types 1, 2 and 3).

result:

a list storing results from S4 methods of this class.

summary:

a list of summary information for available results.

Methods

An overview of methods (More detailed introduction can be found in help for each specific function.):

permNULL

do permutations for input rich phenotyping screens ('pheno').

fitNULL

fit the permuted association scores to a beta distribution.

fitBM

fit the functional association scores computed from input screens to a three-beta mixture model.

p2SNR

Translate p-values to Signal-to-Noise Ratios.

SNR2p

Translate Signal-to-Noise Ratios to p-values.

view

view the fitting results (a histogram of the original data and fitted probability density curves) for NULL and real data.

summarize

summarize results including input data and parameters, NULL fitting and beta mixture fitting.

References

Xin Wang, Mauro Castro, Klaas W. Mulder and Florian Markowetz, Posterior association networks and enriched functional gene modules inferred from rich phenotypic perturbation screens, in preparation.

See Also

permNULL fitNULL fitBM view summarize

Examples

## Not run: 
# 	data(Bakal2007)
# 	bm1<-new("BetaMixture", pheno=Bakal2007, metric="cosine",
# 		model="global", order=1)
# 	bm1<-fitNULL(bm1, nPerm=10, thetaNULL=c(alphaNULL=4, betaNULL=4),
# 		sumMethod="median", permMethod="all", verbose=TRUE)
# 	bm1<-fitBM(bm1, para=list(zInit=NULL, thetaInit=c(alphaNeg=2, betaNeg=4, 
# 		alphaNULL=bm1@result$fitNULL$thetaNULL[["alphaNULL"]], 
# 		betaNULL=bm1@result$fitNULL$thetaNULL[["betaNULL"]], 
# 		alphaPos=4, betaPos=2), gamma=NULL), 
# 		ctrl=list(fitNULL=FALSE, tol=1e-1), verbose=TRUE, gradtol=1e-3)
# 	view(bm1, "fitNULL")
# 	view(bm1, "fitBM")
# 	bm1
# ## End(Not run)