PAN-class: An S4 class for inferring a posterior association network

In PANR: Posterior association networks and functional modules inferred from rich phenotypes of gene perturbations

Description

This S4 class includes methods to infer posterior association networks and enriched modules of functional gene interactions from rich phenotyping screens.

Objects from the Class

Objects of class PAN can be created from new("PAN", bm1, bm2) (see the example below for details).

Slots

bm1:

an object of S4 class BetaMixture, which models the first- order similarities between genes (see BetaMixture).

bm2:

an object of S4 class BetaMixture, which models the second- order similarities between genes (modularity).

edgeWt:

a weighted adjacency matrix computed from the posterior probabilities for gene associations to belong to mixture components (see edgeWeight).

engine:

the graphics visualization engine for PAN.

graph:

a weighted adjacency matrix with edge weights satisfying certain constraints specified by the user (see infer).

modules:

a list summarizing inferred enriched functional gene modules (see pvclustModule.

iPAN:

an igraph object for storing the inferred PAN.

legend:

a list of legends for built PAN graph.

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.):

edgeWeight

compute edge weights by signal-to-noise ratio, posterior odd or posterior probabilities (more details in edgeWeight).

infer

infer a posterior association network given the beta-mixture model(s) fitted to first- and/or second-order similarities (more details in infer).

pvclustModule

search significantly enriched functional gene modules by hierarchical clustering with bootstrap resampling based on the package pvclust (more details in pvclustModule).

exportPAN

export the inferred PAN or modules to file(s) in a variety of formats (more details in exportPAN).

sigModules

retrieve significant gene modules that satisfy the given p-value cutoff and module size range (more details in sigModules).

viewNestedModules

view a nested structure for gene modules searched by hierarchical clustering (more details in viewNestedModules).

viewPAN

view the inferred PAN or modules in igraph or RedeR (more details in viewPAN).

buildPAN

build a PAN graph for visualization in igraph or RedeR (more details in viewPAN).

viewLegend

View the legends for the graph built for PAN.

summarize

summarize results including input data and parameters, inferred graph and modules.

Author(s)

Xin Wang [e-mail]

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

edgeWeight infer pvclustModule exportPAN sigModules viewPAN viewNestedModules summarize

Examples

## Not run: 
data(bm, package="PANR")
##create an object of `PAN'
pan<-new("PAN", bm1=bm1)
##infer a PAN
pan<-infer(pan, para=list(type="SNR", log=TRUE, sign=TRUE, cutoff=log(5)),
	filter=FALSE, verbose=TRUE)
##build a PAN graph for RedeR, hide negative edges
##using colors scaled based on the clustering results from Bakal et al. 2007 
data(Bakal2007Cluster)
pan<-buildPAN(pan, engine="RedeR", para=list(nodeColor=nodeColor, hideNeg=TRUE))
##view PAN in RedeR
library(RedeR)
viewPAN(pan, what="graph")
##print a summary of results
summarize(pan, "ALL")

## End(Not run)

PANR documentation built on Nov. 8, 2020, 8:15 p.m.