Package 'RTN'

Description

A transcriptional regulatory network (TRN) consists of a collection of transcription factors (TFs) and the regulated target genes. TFs are regulators that recognize specific DNA sequences and guide the expression of the genome, either activating or repressing the expression the target genes. The set of genes controlled by the same TF forms a regulon. This package provides classes and methods for the reconstruction of TRNs and analysis of regulons.

Index

Further information is available in the vignettes by typing vignette("RTN"). Documented topics are also available by typing help.start() and selecting the RTN package from the menu.

Author(s)

Maintainer: Mauro Castro <[email protected]>

References

Fletcher M.N.C. et al., Master regulators of FGFR2 signalling and breast cancer risk. Nature Communications, 4:2464, 2013.

Castro M.A.A. et al., Regulators of genetic risk of breast cancer identified by integrative network analysis. Nature Genetics, 48:12-21, 2016.

Description

This S4 class includes a series of methods to do enrichment analyses in Associated Variant Sets (AVSs).

Objects from the Class

Objects can be created by calls of the form new("AVS", markers).

Slots

markers:

Object of class "character", a data frame, a 'BED file' format with rs# markers mapped to the same genome build of the LD source in the RTNdata package.

validatedMarkers:

Object of class "data.frame", a data frame with genome positions of the validated markers.

variantSet:

Object of class "list", an associated variant set.

randomSet:

Object of class "list", a random associated variant set.

para:

Object of class "list", a list of parameters for variant set enrichment analysis.

results:

Object of class "list", a list of results (see return values in the AVS methods).

summary:

Object of class "list", a list of summary information for markers, para, and results.

status:

Object of class "character", a character value specifying the status of the AVS object based on the available methods.

Methods

avs.vse

signature(object = "AVS"): see avs.vse

avs.evse

signature(object = "AVS"): see avs.evse

avs.rvse

signature(object = "AVS"): see avs.rvse

avs.pevse

signature(object = "AVS"): see avs.pevse

avs.get

signature(object = "AVS"): see avs.get

Author(s)

Mauro Castro

See Also

TNA-class

Examples

## Not run: 
#This example requires the RTNdata package! (currently available under request)
library(RTNdata.LDHapMapRel27.hg18)
data(bcarisk, package = "RTNdata.LDHapMapRel27.hg18")
avs <- avs.preprocess.LDHapMapRel27.hg18(bcarisk, nrand=100)

## End(Not run)

Description

The VSE method (avs.vse) provides a robust framework to cope with the heterogeneous structure of haplotype blocks, and has been designed to test enrichment in cistromes and epigenomes. In order to extend the variant set enrichment to genes this pipeline implements an additional step using expression quantitative trait loci (eQTLs).

Usage

avs.evse(object, annotation, gxdata, snpdata, glist=NULL, maxgap=250, minSize=100, 
pValueCutoff=0.05, pAdjustMethod="bonferroni", boxcox=TRUE, 
fineMapping=TRUE, verbose=TRUE)

Arguments

Value

a data frame in the slot "results", see 'what' options in avs.get.

Author(s)

Mauro Castro

See Also

AVS-class

Examples

## Not run: 
# This example requires the RTNdata package! (currently available under request)
library(RTNdata.LDHapMapRel27.hg18)
library(Fletcher2013b)
library(TxDb.Hsapiens.UCSC.hg18.knownGene)

##################################################
### Build AVS and random AVSs (mapped to hg18)
##################################################

#--- step 1: load 'risk SNPs' data (e.g. BCa risk SNPs from the GWAS catalog)
data(bcarisk, package = "RTNdata.LDHapMapRel27.hg18")

#--- step 2: build an AVS and 1000 matched random AVSs for the input 'risk SNPs'
bcavs <- avs.preprocess.LDHapMapRel27.hg18(bcarisk, nrand=1000)

##################################################
### Example of EVSE analysis for sets of genomic 
### annotations (e.g. regulons, gene sets, etc.)
##################################################

#--- step 1: load a precomputed AVS (same 'bcavs' object as above!)
data(bcavs, package="RTNdata.LDHapMapRel27.hg18")

#--- step 2: load genomic annotation for all genes
genemap <- as.data.frame(genes(TxDb.Hsapiens.UCSC.hg18.knownGene))
genemap <- genemap[,c("seqnames","start","end","gene_id")]
colnames(genemap) <- c("CHROM","START","END","ID")

#--- step 3: load a TNI object, or any other source of regulons (e.g. gene sets)
#--- and prepare a gene set list 
#--- (gene ids should be the same as in the 'genemap' object)
data("rtni1st")
glist <- tni.get(rtni1st,what="refregulons",idkey="ENTREZ")
glist <- glist[ c("FOXA1","GATA3","ESR1") ] #reduce the list for demonstration!

#--- step 4: input matched variation and gene expression datasets!
#--- here we use two "toy" datasets for demonstration purposes only.
data(toy_snpdata, package="RTNdata.LDHapMapRel27.hg18")
data(toy_gxdata, package="RTNdata.LDHapMapRel27.hg18")

#--- step 5: run the avs.evse pipeline
bcavs<-avs.evse(bcavs, annotation=genemap, gxdata=toy_gxdata, 
                snpdata=toy_snpdata, glist=glist, pValueCutoff=0.01)

#--- step 6: generate the EVSE plots
avs.plot2(bcavs,"evse", height=2.5, width.panels=c(1,2), rmargin=0)

### NOTE REGARDING THIS EXAMPLE ####
#- This example is for demonstration purposes only, using toy datasets. 
#- Any eventual positive/negative associations derived from these datasets 
#- are not comparable with the original studies that described the method 
#- (doi: 10.1038/ng.3458; 10.1038/ncomms3464).
####################################


## End(Not run)

Description

Get information from individual slots in an AVS object.

Usage

avs.get(object, what="summary", report=FALSE, pValueCutoff=NULL)

Arguments

Value

get the slot content from an object of class 'AVS' AVS-class.

Author(s)

Mauro Castro

Examples

## Not run: 
#This example requires the RTNdata package! (currently available under request)
library(RTNdata.LDHapMapRel27)
data(bcarisk, package="RTNdata.LDHapMapRel27")
bcavs <- avs.preprocess.LDHapMapRel27(bcarisk, nrand=1000)
avs.get(avs)

## End(Not run)

Description

The VSE method (avs.vse) provides a robust framework to cope with the heterogeneous structure of haplotype blocks, and has been designed to test enrichment in cistromes and epigenomes. In order to extend the variant set enrichment to genes this pipeline implements an additional step using precomputed expression quantitative trait loci (eQTLs).

Usage

avs.pevse(object, annotation, eqtls, glist, maxgap=250, minSize=100, 
pValueCutoff=0.05, pAdjustMethod="bonferroni", boxcox=TRUE,
verbose=TRUE)

Arguments

Value

a data frame in the slot "results", see 'what' options in avs.get.

Author(s)

Mauro Castro, Steve Booth

See Also

AVS-class

Examples

## Not run: 

# This example requires the RTNdata package! (currently available under request)
library(RTNdata.LDHapMapRel27.hg18)
library(Fletcher2013b)
library(TxDb.Hsapiens.UCSC.hg18.knownGene)

##################################################
### Example of EVSE analysis for sets of genomic 
### annotations (e.g. regulons, gene sets, etc.)
##################################################

#--- step 1: load a precomputed AVS
data(bcavs, package="RTNdata.LDHapMapRel27.hg18")

#--- step 2: load genomic annotation for all genes
genemap <- as.data.frame(genes(TxDb.Hsapiens.UCSC.hg18.knownGene))
genemap <- genemap[,c("seqnames","start","end","gene_id")]
colnames(genemap) <- c("CHROM","START","END","ID")

#--- step 3: load a TNI object (or any other source of regulons)
#--- and prepare a gene set list.
#--- Note: gene ids should be the same as in the 'genemap' object.
data("rtni1st")
glist <- tni.get(rtni1st,what="refregulons",idkey="ENTREZ")
glist <- glist[ c("FOXA1","GATA3","ESR1") ] #reduce the list for demonstration!

#--- step 4: load precomputed eQTLs
#--- Please note that the input data should represent eQTLs from genome-wide
#--- calls, that is, the universe size should cover 'all SNPs' vs. 'all genes'.
#--- The correct represetation of universe size is essential to build the
#--- null distributions. Or, to put it another way, the eQTL analysis 
#--- should unbiasedly test linked and random markers from the AVS.
#--- In this example it represents 2029 + 966240 SNPs:

lkMarkers <- avs.get(bcavs,what="linkedMarkers")
length(lkMarkers) # i.e. 2029 risk associated and linked SNPs 

rdMarkers <- avs.get(bcavs,what="randomMarkers")
length(rdMarkers) # i.e. 966240 random SNPs

#--- Now we prepare a 'toy' dataset for demonstration purposes only
#--- by picking (naively) SNPs within 250 kb window around the 
#--- genomic annotation.

## load HapMap SNPs (release 27) mapped to hg18
data("popsnp2")

## map SNPs to the genomic annotation
query <- with(popsnp2, 
              GRanges(chrom, IRanges(position, position), 
              id=rsid)
              )
subject <- with(genemap, 
                GRanges(CHROM, IRanges(START, END), 
                id=ID)
                )
hits <- findOverlaps(query,subject, maxgap = 250000)

## reduce 'hits' just for demonstration
hits <- hits[sort(sample(length(hits), 50000))]

## build a 'toy_eqtls' data frame
toy_eqtls <- data.frame(rsid = popsnp2$rsid[from(hits)], 
                        geneid = genemap$ID[to(hits)])

#--- step 5: run the 'avs.pevse' pipeline
#--- important: set 'maxgap' to the same searching window used 
#--- in the dQTL analysis (e.g. 250kb)
bcavs <- avs.pevse(bcavs, annotation=genemap, glist=glist, 
                   eqtls=toy_eqtls, maxgap = 250)

#--- step 6: generate the pEVSE plots
avs.plot2(bcavs,"pevse",height=2.5)


#####################################################
#--- parallel version for 'step 5' with SNOW package!
# library(snow)
# options(cluster=snow::makeCluster(3, "SOCK"))
# bcavs <- avs.pevse(bcavs, annotation=genemap, glist=glist, 
#          eqtls=toy_eqtls, maxgap = 250)
# stopCluster(getOption("cluster"))

## ps. as a technical note, the parallel version uses a
## slightly different overlap-based operation, which might
## bring slightly different counts depending on the data 
## input organization


## End(Not run)

Description

This function takes an AVS object and plots results from the VSE and EVSE methods.

Usage

avs.plot1(object, what="vse", fname=what, ylab="genomic annotation", 
xlab="Number of clusters mapping to genomic annotation", breaks="Sturges", 
maxy=200, pValueCutoff=1e-2, width=8, height=3)

Arguments

Value

A plot showing results from the VSE and EVSE methods.

Author(s)

Mauro Castro

Examples

# see 'avs.vse' and 'avs.evse' methods.

Description

This function takes an AVS object and plots results from the VSE and EVSE methods.

Usage

avs.plot2(object, what="evse", fname=what, width=14, height=2.5, 
width.panels=c(1,3), rmargin=1, at.x=seq(-4,8,2), decreasing=TRUE, 
ylab="Annotation", xlab="Clusters of risk-associated and linked SNPs", 
tfs=NULL)

Arguments

Value

A plot showing results from the VSE and EVSE methods.

Author(s)

Mauro Castro

Examples

# see 'avs.vse' and 'avs.evse' methods.

Description

The VSE method (avs.vse) provides a robust framework to cope with the heterogeneous structure of haplotype blocks, and has been designed to test enrichment in cistromes and epigenomes. In order to extend the variant set enrichment to genes this pipeline implements an additional step using regulon quantitative trait loci (rQTLs).

Usage

avs.rvse(object, annotation, regdata, snpdata, glist, 
maxgap=250, minSize=100, pValueCutoff=0.05, 
pAdjustMethod="bonferroni", boxcox=TRUE, verbose=TRUE)

Arguments

Value

a data frame in the slot "results", see 'what' options in avs.get.

Author(s)

Mauro Castro

See Also

AVS-class

Examples

## Not run: 
# This example requires the RTNdata package! (currently available under request)
library(RTNdata.LDHapMapRel27.hg18)
library(Fletcher2013b)
library(TxDb.Hsapiens.UCSC.hg18.knownGene)

##################################################
### Build AVS and random AVSs (mapped to hg18)
##################################################

#--- step 1: load 'risk SNPs' data (e.g. BCa risk SNPs from the GWAS catalog)
data(bcarisk, package = "RTNdata.LDHapMapRel27.hg18")

#--- step 2: build an AVS and 1000 matched random AVSs for the input 'risk SNPs'
bcavs <- avs.preprocess.LDHapMapRel27.hg18(bcarisk, nrand=1000)

##################################################
### Example of RVSE analysis
##################################################

#--- step 1: load a precomputed AVS (same 'bcavs' object as above!)
data(bcavs, package="RTNdata.LDHapMapRel27.hg18")

#--- step 2: load genomic annotation for all genes
genemap <- as.data.frame(genes(TxDb.Hsapiens.UCSC.hg18.knownGene))
genemap <- genemap[,c("seqnames","start","end","gene_id")]
colnames(genemap) <- c("CHROM","START","END","ID")

#--- step 3: load a TNI object and get a list with regulons
#--- (ids should be the same as in 'genemap')
data("rtni1st")
glist <- tni.get(rtni1st,what="refregulons",idkey="ENTREZ")
glist <- glist[ c("FOXA1","GATA3","ESR1") ] #reduce the list for demonstration!

#--- step 4: compute single-sample regulon activity
rtni1st <- tni.gsea2(rtni1st, regulatoryElements = c("FOXA1","GATA3","ESR1"))
regdata <- tni.get(rtni1st, what = "regulonActivity")$diff

#--- step 5: load a variation dataset matched with the regulon activity dataset
#--- here we use a "toy" dataset for demonstration purposes only.
data(toy_snpdata, package="RTNdata.LDHapMapRel27")

#--- step 6: check "snpdata"" and "regdata"" alignment (samples on cols)
regdata <- t(regdata)
all(colnames(toy_snpdata)==colnames(regdata))
#TRUE

#--- step 7: run the avs.rvse pipeline
bcavs <- avs.rvse(bcavs, 
                  annotation=genemap,
                  regdata=regdata, 
                  snpdata=toy_snpdata, 
                  glist=glist,
                  pValueCutoff=0.01)

#--- step 8: generate the RVSE plot
avs.plot2(bcavs, "rvse", height=2.5)


## End(Not run)

Description

The VSE method tests the enrichment of an AVS for a particular trait in a genomic annotation.

Usage

avs.vse(object, annotation, glist=NULL, maxgap=0, minSize=100,
pValueCutoff=0.05, pAdjustMethod="bonferroni", boxcox=TRUE,
verbose=TRUE)

Arguments

Value

a data frame in the slot "results", see 'what' options in avs.get.

Author(s)

Mauro Castro

See Also

AVS-class

Examples

## Not run: 
# This example requires the RTNdata package! (currently available under request)
library(RTNdata.LDHapMapRel27.hg18)
library(Fletcher2013b)
library(TxDb.Hsapiens.UCSC.hg18.knownGene)

##################################################
### Build AVS and random AVSs (mapped to hg18)
##################################################

#--- step 1: load 'risk SNPs' data (e.g. BCa risk SNPs from the GWAS catalog)
data(bcarisk, package = "RTNdata.LDHapMapRel27.hg18")

#--- step 2: build an AVS and 1000 matched random AVSs for the input 'risk SNPs'
bcavs <- avs.preprocess.LDHapMapRel27.hg18(bcarisk, nrand=1000)

##################################################
### Example of VSE analysis for ERa and FOXA1 
### cistromes (one genomic annotation each time)
##################################################

#--- step 1: load a precomputed AVS (same 'bcavs' object as above!)
data(bcavs, package="RTNdata.LDHapMapRel27.hg18")

#--- step 2: load cistrome data from the Fletcher2013b package
#NOTE: Fletcher2013b is a large data package, but only two 'bed files' 
#are used to illustrate this analysis (ESR1bdsites and FOXA1bdsites). 
#these bed files provide ERa and FOXA1 binding sites mapped by 
#ChIP-seq experiments
data(miscellaneous)

#--- step 3: run the avs.vse pipeline
bcavs <- avs.vse(bcavs, annotation=ESR1bdsites$bdsites, 
                 pValueCutoff=0.001)
bcavs <- avs.vse(bcavs, annotation=FOXA1bdsites$bdsites, 
                 pValueCutoff=0.001)

#--- step 4: generate the VSE plots
avs.plot2(bcavs,"vse",height=2.2, width.panels=c(1,2), rmargin=0)

##################################################
### Example of VSE analysis for sets of genomic 
### annotations (e.g. regulons, gene sets, etc.)
##################################################

#--- step 1: load the precomputed AVS (same 'bcavs' object as above!)
data(bcavs, package="RTNdata.LDHapMapRel27.hg18")

#--- step 2: load genomic annotation for all genes
genemap <- as.data.frame(genes(TxDb.Hsapiens.UCSC.hg18.knownGene))
genemap <- genemap[,c("seqnames","start","end","gene_id")]
colnames(genemap) <- c("CHROM","START","END","ID")

#--- step 3: load a TNI object, or any other source of regulons (e.g. gene sets)
#--- and prepare a gene set list 
#--- (gene ids should be the same as in the 'genemap' object)
data("rtni1st")
glist <- tni.get(rtni1st,what="refregulons",idkey="ENTREZ")
glist <- glist[ c("FOXA1","GATA3","ESR1") ] #reduce the list for demonstration!

#--- step 4: run the avs.vse pipeline
bcavs<-avs.vse(bcavs, annotation=genemap, glist=glist, pValueCutoff=0.05)

#--- step 5: generate the VSE plots
avs.plot2(bcavs,"vse", height=2.5, width.panels=c(1,2), rmargin=0)

### NOTE REGARDING THIS EXAMPLE ####
#- This example is for demonstration purposes only;
#- we recommend using the EVSE/eQTL approach when analysing genes/regulons.
#- Also, the AVS object here is not the same as the one used in the study that  
#- extended the method (doi:10.1038/ng.3458), so the results are not comparable;
#- (here fewer risk SNPs are considered, and without an eQTL step).
####################################


## End(Not run)

Description

Datasets used to demonstrate RTN main functions.

Format

tniData, tnaData, and tfsData

Details

The tniData and tnaData datasets were extracted, pre-processed and size-reduced from Fletcher et al. (2013) and Curtis et al. (2012). They consist of two lists used in the RTN vignettes for demonstration purposes only. The tniData list contains the 'expData', 'rowAnnotation' and 'colAnnotation' R objects, while the tnaData list contains the 'phenotype', 'phenoIDs' and 'hits' R objects.

The tfsData consists of a list with gene annotation for human transcription factors (TFs), compiled from 6 resources (Lambert et. al 2018; Carro et al. 2010; Vaquerizas et al. 2009; D. L. Fulton et al. 2009; Yusuf et al. 2012; and Ravasi et al. 2010), and provided here in the form of 'data frame' objects that include ENTREZ and HGNC gene symbol annotation.

The pksData consists of a list with gene annotation for human protein kinases, retrived from the Swiss-Prot Protein Knowledgebase (https://www.uniprot.org/docs/pkinfam), release 2020_02 of 22-Apr-2020.

tniData$expData

a named gene expression matrix with 120 samples (a subset from the Fletcher2013b).

tniData$rowAnnotation

a data.frame of characters with gene annotation (a subset from the Fletcher2013b).

tniData$colAnnotation

a data.frame of characters with sample annotation (a subset from the Fletcher2013b).

tnaData$phenotype

a named numeric vector with differential gene expression data.

tnaData$phenoIDs

a data.frame of characters with probe ids matching a secundary annotation source (e.g. Probe-to-ENTREZ).

tnaData$hits

a character vector with genes differentially expressed.

tfsData$Lambert2018

a data.frame listing TFs from Lambert et. al (2018).

tfsData$Yusuf2012

a data.frame listing TFs from Yusuf et al. (2012).

tfsData$Carro2010

a data.frame listing TFs from Carro et al. (2010).

tfsData$Ravasi2010

a data.frame listing TFs from Ravasi et al. (2010).

tfsData$Fulton2009

a data.frame listing TFs from Fulton et al. (2009).

tfsData$Vaquerizas2009

a data.frame listing TFs from Vaquerizas et al. (2009).

tfsData$all

a data.frame listing all TFs.

pksData$pkinfam2020

a data.frame listing human protein kinases.

References

Carro, M.S. et al., The transcriptional network for mesenchymal transformation of brain tumors. Nature, 463(7279):318-325, 2010.

Curtis C. et al., The genomic and transcriptomic architecture of 2,000 breast tumours reveals novel subgroups. Nature, 486(7403):346-352, 2012.

Fletcher, M.N.C. et al., Master regulators of FGFR2 signalling and breast cancer risk. Nature Communications, 4:2464, 2013.

Fulton, D.L. et al., TFCat: the curated catalog of mouse and human transcription factors. Genome Biology, 10(3):R29, 2009.

Lambert, S.A. et al., The Human Transcription Factors. Cell, 172(4):650-665, 2018.

Ravasi, T. et al., An atlas of combinatorial transcriptional regulation in mouse and man. Cell, 140(5):744-752, 2010.

Vaquerizas, J.M. et al., A census of human transcription factors: function, expression and evolution. Nature Reviews Genetics, 10(4):252-263, 2009.

Yusuf, D. et al., The Transcription Factor Encyclopedia. Genome Biology, 13(3):R24, 2012.

Examples

data(tniData)
    data(tnaData)
    data(tfsData)
    data(pksData)

Description

This S4 class includes a series of methods to do enrichment analyses in transcriptional networks.

Objects from the Class

Objects can be created by calls of the form new("TNA", referenceNetwork, transcriptionalNetwork, regulatoryElements, phenotype, hits).

Slots

referenceNetwork:

Object of class "matrix", an optional partial co-expression matrix.

transcriptionalNetwork:

Object of class "matrix", a partial co-expression matrix.

regulatoryElements:

Object of class "char_Or_NULL", a vector of regulatory elements (e.g. transcription factors).

phenotype:

Object of class "num_Or_int", a numeric or integer vector of phenotypes named by gene identifiers.

hits:

Object of class "character", a character vector of gene identifiers for those considered as hits.

gexp:

Object of class "matrix", a gene expression matrix.

rowAnnotation:

Object of class "data.frame", a data frame with row annotation (e.g. probe-to-gene information).

colAnnotation:

Object of class "data.frame", a data frame with column annotation (e.g. sample information).

listOfReferenceRegulons:

Object of class "list", a list of regulons derived from the referenceNetwork.

listOfRegulons:

Object of class "list", a list of regulons derived from the transcriptionalNetwork (a 'regulon' is a vector of genes or potential transcription factor targets).

listOfModulators:

Object of class "list", a list of modulators derived from the tni.conditional analysis.

para:

Object of class "list", a list of parameters for transcriptional network analysis. These parameters are those listed in the functions tna.mra, tna.gsea1, and tna.gsea2.

results:

Object of class "list", a list of results (see return values in the functions tna.mra,tna.gsea1, and tna.gsea2)

summary:

Object of class "list", a list of summary information for transcriptionalNetwork, regulatoryElements, phenotype,listOfRegulons, para, and results.

status:

Object of class "character", a character value specifying the status of the TNI object based on the available methods.

Methods

tna.mra

signature(object = "TNA"): see tna.mra

tna.gsea1

signature(object = "TNA"): see tna.gsea1

tna.gsea2

signature(object = "TNA"): see tna.gsea2

tna.get

signature(object = "TNA"): see tna.get

Author(s)

Mauro Castro

See Also

TNI-class. tni2tna.preprocess.

Examples

## see 'tni2tna.preprocess' method!

Description

Get information from individual slots in a TNA object. Available results from a previous analysis can be selected either by pvalue cutoff (default) or top significance.

Usage

tna.get(object, what="summary", order=TRUE, ntop=NULL, reportNames=TRUE, 
idkey=NULL)

Arguments

Details

Options for the 'what' argument retrieve the following types of information:

summary

A list summarizing parameters and results available in the TNA object.

status

A vector indicating the status of each available method in the pipeline.

para

A list with the parameters used by each available method in the pipeline.

pheno

A numeric vector of phenotypes named by gene identifiers (see tni2tna.preprocess).

hits

A character vector of gene identifiers for those considered as hits (see tni2tna.preprocess).

regulatoryElements

A vector of regulatory elements (e.g. transcription factors).

tnet

A data matrix with MI values, evaluated by the DPI filter. MI values are computed between regulators and targets, with regulators on cols and targets on rows. Note that signals (+/-) are assigned to the inferred associations in order to represent the 'mode of action', which is derived from Pearson's correlation between regulators and targets.

refnet

A data matrix with MI values (not evaluated by the DPI filter). MI values are computed between regulators and targets, with regulators on cols and targets on rows. Note that signals (+/-) are assigned to the inferred associations in order to represent the 'mode of action', which is derived from Pearson's correlation between regulators and targets.

regulons

A list with regulons extracted from the 'tnet' data matrix.

refregulons

A list with regulons extracted from the 'refnet' data matrix.

regulons.and.mode

A list with regulons extracted from the 'tnet' data matrix, including the assiged 'mode of action'.

refregulons.and.mode

A list with regulons extracted from the 'refnet' data matrix, including the assiged 'mode of action'.

rowAnnotation

A data frame with probe-to-gene annotation.

colAnnotation

A data frame with sample annotation.

mra

A data frame with results from the tna.mra analysis pipeline.

gsea1

A data frame with results from the tna.gsea1 analysis pipeline.

gsea2

A data frame with results from the tna.gsea2 analysis pipeline.

Value

Get the slot content from a TNA-class object.

Author(s)

Mauro Castro

Examples

data(tniData)
data(tnaData)

## Not run: 

rtni <- tni.constructor(expData=tniData$expData, 
        regulatoryElements=c("PTTG1","E2F2","FOXM1","E2F3","RUNX2"), 
        rowAnnotation=tniData$rowAnnotation)
rtni <- tni.permutation(rtni)
rtni <- tni.bootstrap(rtni)
rtni <- tni.dpi.filter(rtni)
rtna <- tni2tna.preprocess(rtni, phenotype=tnaData$phenotype, 
        hits=tnaData$hits, phenoIDs=tnaData$phenoIDs)

# run MRA analysis pipeline
rtna <- tna.mra(rtna)

# check summary
tna.get(rtna,what="summary")

# get results, e.g., from the MRA analysis
tna.get(rtna,what="mra")


## End(Not run)

Description

This function takes a TNA object and returns the results of the GSEA analysis over a list of regulons in a transcriptional network (with multiple hypothesis testing corrections).

Usage

tna.gsea1(object, pValueCutoff=0.05, pAdjustMethod="BH",  minRegulonSize=15, 
sizeFilterMethod="posORneg", nPermutations=1000, exponent=1, tnet="dpi", 
signature=c("phenotype","hits"), orderAbsValue=TRUE, tfs=NULL, verbose=TRUE)

Arguments

Value

a data frame in the slot "results", see 'gsea1' option in tna.get.

Author(s)

Mauro Castro, Xin Wang

See Also

TNA-class tna.plot.gsea1

Examples

data(tniData)
data(tnaData)

## Not run: 

rtni <- tni.constructor(expData=tniData$expData, 
        regulatoryElements=c("PTTG1","E2F2","FOXM1","E2F3","RUNX2"), 
        rowAnnotation=tniData$rowAnnotation)
rtni <- tni.permutation(rtni)
rtni <- tni.bootstrap(rtni)
rtni <- tni.dpi.filter(rtni)
rtna <- tni2tna.preprocess(rtni, phenotype=tnaData$phenotype, 
        hits=tnaData$hits, phenoIDs=tnaData$phenoIDs)

#run GSEA1 analysis pipeline
rtna <- tna.gsea1(rtna)

#get results
tna.get(rtna, what="gsea1")

# run parallel version with SNOW package!
library(snow)
options(cluster=snow::makeCluster(3, "SOCK"))
rtna <- tna.gsea1(rtna)
stopCluster(getOption("cluster"))

## End(Not run)

Description

This function takes a TNA object and returns a CMAP-like analysis obtained by two-tailed GSEA over a list of regulons in a transcriptional network (with multiple hypothesis testing corrections).

Usage

tna.gsea2(object, pValueCutoff=0.05, pAdjustMethod="BH", minRegulonSize=15, 
sizeFilterMethod="posORneg", nPermutations=1000, exponent=1, tnet="dpi", 
signature=c("phenotype","hits"), tfs=NULL, verbose=TRUE, doSizeFilter=NULL)

Arguments

Value

a data frame in the slot "results", see 'gsea2' option in tna.get.

Author(s)

Mauro Castro

See Also

TNA-class tna.plot.gsea2

Examples

data(tniData)
data(tnaData)

## Not run: 

rtni <- tni.constructor(expData=tniData$expData, 
        regulatoryElements=c("PTTG1","E2F2","FOXM1","E2F3","RUNX2"), 
        rowAnnotation=tniData$rowAnnotation)
rtni <- tni.permutation(rtni)
rtni <- tni.bootstrap(rtni)
rtni <- tni.dpi.filter(rtni)
rtna <- tni2tna.preprocess(rtni, phenotype=tnaData$phenotype, 
        hits=tnaData$hits, phenoIDs=tnaData$phenoIDs)

#run GSEA2 analysis pipeline
rtna <- tna.gsea2(rtna)

#get results
tna.get(rtna, what="gsea2")

# run parallel version with SNOW package!
library(snow)
options(cluster=snow::makeCluster(3, "SOCK"))
rtna <- tna.gsea2(rtna)
stopCluster(getOption("cluster"))

## End(Not run)

Description

This function takes a TNA object and returns the results of the RMA analysis over a list of regulons from a transcriptional network (with multiple hypothesis testing corrections).

Usage

tna.mra(object, pValueCutoff=0.05, pAdjustMethod="BH", minRegulonSize=15, 
tnet="dpi", tfs=NULL, verbose=TRUE)

Arguments

Value

a data frame in the slot "results", see 'rma' option in tna.get.

Author(s)

Mauro Castro

See Also

TNA-class

Examples

data(tniData)
data(tnaData)

## Not run: 

rtni <- tni.constructor(expData=tniData$expData, 
        regulatoryElements=c("PTTG1","E2F2","FOXM1","E2F3","RUNX2"), 
        rowAnnotation=tniData$rowAnnotation)
rtni <- tni.permutation(rtni)
rtni <- tni.bootstrap(rtni)
rtni <- tni.dpi.filter(rtni)
rtna <- tni2tna.preprocess(rtni, phenotype=tnaData$phenotype, 
        hits=tnaData$hits, phenoIDs=tnaData$phenoIDs)

#run MRA analysis pipeline
rtna <- tna.mra(rtna)

#get results
tna.get(rtna,what="mra")


## End(Not run)

Description

This function takes a TNA object and plots the one-tailed GSEA results for individual regulons.

Usage

tna.plot.gsea1(object, labPheno="", file="tna_gsea1", 
filepath=".", regulon.order="size", ntop=NULL, tfs=NULL, 
ylimPanels=c(0.0,3.5,0.0,0.8), heightPanels=c(1,1,3), width=4.4, 
height=4, ylabPanels=c("Phenotype","Regulon","Enrichment score"), 
xlab="Position in the ranked list of genes", alpha=0.5, 
sparsity=10, autoformat=TRUE, plotpdf=TRUE, ...)

Arguments

Value

A plot showing results from the 'tna.gsea1' method.

Author(s)

Mauro Castro

See Also

tna.gsea1

Examples

data(tniData)
data(tnaData)

## Not run: 

rtni <- tni.constructor(expData=tniData$expData, 
        regulatoryElements=c("PTTG1","E2F2","FOXM1","E2F3","RUNX2"), 
        rowAnnotation=tniData$rowAnnotation)
rtni <- tni.permutation(rtni)
rtni <- tni.bootstrap(rtni)
rtni <- tni.dpi.filter(rtni)
rtna <- tni2tna.preprocess(rtni, phenotype=tnaData$phenotype, 
        hits=tnaData$hits, phenoIDs=tnaData$phenoIDs)

# run GSEA analysis pipeline
rtna <- tna.gsea1(rtna)

# plot available GSEA results
tna.plot.gsea1(rtna, labPheno="test")

## End(Not run)

Description

This function takes a TNA object and plots the two-tailed GSEA results for individual regulons.

Usage

tna.plot.gsea2(object, labPheno="", file="tna_gsea2", filepath=".", 
regulon.order="size", ntop=NULL, tfs=NULL, ylimPanels=c(-3.0,3.0,-0.5,0.5), 
heightPanels=c(2.0,0.8,5.0), width=2.8, height=3.0, 
ylabPanels=c("Phenotype","Regulon","Enrichment score"), 
xlab="Position in the ranked list of genes", alpha=1.0, 
sparsity=10, autoformat=TRUE, plotpdf=TRUE, ...)

Arguments

Value

A plot showing results from the 'tna.gsea2' method.

Author(s)

Mauro Castro

See Also

tna.gsea2

Examples

data(tniData)
data(tnaData)

## Not run: 

rtni <- tni.constructor(expData=tniData$expData, 
        regulatoryElements=c("PTTG1","E2F2","FOXM1","E2F3","RUNX2"), 
        rowAnnotation=tniData$rowAnnotation)
rtni <- tni.permutation(rtni)
rtni <- tni.bootstrap(rtni)
rtni <- tni.dpi.filter(rtni)
rtna <- tni2tna.preprocess(rtni, phenotype=tnaData$phenotype, 
        hits=tnaData$hits, phenoIDs=tnaData$phenoIDs)

# run GSEA2 analysis pipeline
rtna <- tna.gsea2(rtna)

# plot available GSEA2 results
tna.plot.gsea2(rtna, labPheno="test")

## End(Not run)

Description

This S4 class includes a series of methods to do transcriptional network inference for high-throughput gene expression.

Slots

gexp:

Object of class "matrix", a gene expression matrix.

regulatoryElements:

Object of class "character", a vector of regulatory elements (e.g. transcription factors).

targetElements:

Object of class "character", a vector of target elements (e.g. target genes).

modulators:

Object of class "char_Or_NULL", a vector with modulator identifiers.

rowAnnotation:

Object of class "data.frame", a data frame with row annotation (e.g. probe-to-gene information).

colAnnotation:

Object of class "data.frame", a data frame with column annotation (e.g. sample information).

para:

Object of class "list", a list of parameters for transcriptional network inference. These parameters are those listed in the functions tni.permutation, tni.bootstrap and tni.dpi.filter.

results:

Object of class "list", a list of results (see the returned values in the functions tni.permutation).

summary:

Object of class "list", a list of summary information for gexp, regulatoryElements, para, and results.

status:

Object of class "character", a character value specifying the status of the TNI object based on the available methods.

Methods

tni.preprocess

signature(object = "TNI"): see tni.preprocess

tni.permutation

signature(object = "TNI"): see tni.permutation

tni.bootstrap

signature(object = "TNI"): see tni.bootstrap

tni.dpi.filter

signature(object = "TNI"): see tni.dpi.filter

tni.conditional

signature(object = "TNI"): see tni.conditional

tni.get

signature(object = "TNI"): see tni.get

tni.graph

signature(object = "TNI"): see tni.graph

tni.gsea2

signature(object = "TNI"): see tni.gsea2

tni.area3

signature(object = "TNI"): see tni.area3

tni.regulon.summary

signature(object = "TNI"): see tni.regulon.summary

tni.prune

signature(object = "TNI"): see tni.prune

tni.replace.samples

signature(object = "TNI"): see tni.replace.samples

tni.annotate.regulons

signature(object = "TNI"): see tni.annotate.regulons

tni.annotate.samples

signature(object = "TNI"): see tni.annotate.samples

tni.overlap.genesets

signature(object = "TNI"): see tni.overlap.genesets

tni2tna.preprocess

signature(object = "TNI"): see tni2tna.preprocess

tni.sre

signature(object = "TNI"): see tni.sre

Author(s)

Mauro Castro

See Also

TNA-class

Examples

## see 'tni.constructor'!

Description

When analyzing two datasets that have different numbers of samples, this function can be used to assist the choice of a p value threshold for the tni.permutation() function, in order that RTN will return regulon results that have been generated with similar tradeoffs between Type I and Type II errors for both datasets. Doing this should help ensure that it is reasonable to compare the regulons in the two datasets.

Usage

tni.alpha.adjust(nB, nA, alphaA, betaA = 0.2)

Arguments

Value

Significance level for 'nB', given 'nA', 'alphaA' and 'betaA'.

Note

The 'tni.alpha.adjust' function calls the pwr.r.test function, which uses the 'uniroot' function to solve a power equation. The 'uniroot' function aims to find a root in a given interval, searching from lower to upper end-points. As the upper end-point must be strictly larger than the lower end-point, in order to avoid an error when searching the root, 'nA' must be greater than or equal to 'nB' (i.e. 'nB' is expected to be the smallest data set). Also, please note that 'uniroot' eventually will not find the root for the input arguments, especially when searching thresholds of low stringency (we suggest to avoid setting 'alphaA' > 0.01).

Author(s)

Mauro Castro, Gordon Robertson

See Also

pwr.r.test

Examples

# estimate 'alphaB' for 'nB', given 'nA', 'alphaA' and 'betaA'
alphaB <- tni.alpha.adjust(nB = 100, nA = 300, alphaA = 1e-5, betaA = 0.2)

Description

This function calculates an enrichment score between gene sets and regulons.

Usage

tni.annotate.regulons(object, geneSetList, sampleSetList = NULL, 
regulatoryElements = NULL, minSetSize = 15, sizeFilterMethod="posORneg", 
exponent = 1, verbose = TRUE)

Arguments

Details

Using the samples available in the provided TNI object, the 'tni.annotate.regulons' calculates the enrichment of each regulon for each gene set. First, the samples are split into two groups, one with high average gene-set expression (GS_high) and the other with low average gene-set expression (GS_low). Then a gene-wise differential expression (DEG) signature is generated by comparing the GS_high vs. GS_low groups. The DEG signature is regarded as the gene-set phenotype in the cohort. A GSEA-2T approach is used to calculate the activity score (dES) of each regulon in the phenotype (for additional details on GSEA-2T, please see section 2.2 of the RTN's vignette).

Value

A numeric matrix with dES scores between gene sets vs. regulons.

Author(s)

Mauro Castro

See Also

TNI-class

Examples

data(tniData)

## Not run: 

#compute regulons
rtni <- tni.constructor(expData=tniData$expData, 
        regulatoryElements=c("PTTG1","E2F2","FOXM1","E2F3","RUNX2"), 
        rowAnnotation=tniData$rowAnnotation)
rtni <- tni.permutation(rtni)
rtni <- tni.bootstrap(rtni)
rtni <- tni.dpi.filter(rtni)

#load a gene set collection 
#here, we build three random gene sets for demonstration
geneset1 <- sample(tniData$rowAnnotation$SYMBOL,50)
geneset2 <- sample(tniData$rowAnnotation$SYMBOL,50)
geneset3 <- sample(tniData$rowAnnotation$SYMBOL,50)
geneSetList <- list(geneset1=geneset1,
                    geneset2=geneset2,
                    geneset3=geneset3)

#compute regulon activity
dES <- tni.annotate.regulons(rtni, geneSetList)


## End(Not run)

Description

This function calculates an enrichment score between gene sets and samples.

Usage

tni.annotate.samples(object, geneSetList, minSetSize = 15, 
    exponent = 1, samples=NULL, verbose = TRUE)

Arguments

Details

Using the samples available in the provided TNI object, the 'tni.annotate.samples' calculates the enrichment of each sample for each gene set. First, a gene-wise differential expression (DEG) signature is generated by comparing the expression of a given sample with the avarage expression of all samples. The DEG signature is regarded as a the sample phenotype, representing the relative expression of the sample's genes in the cohort. Then a single-sample Gene Set Enrichment Analysis (ssGSEA) is used to calculate the enrichment score (ES) of the sample for a given gene set.

Value

A numeric matrix with association statistics between gene sets vs. samples.

Author(s)

Mauro Castro

See Also

TNI-class

Examples

data(tniData)

## Not run: 

#generate a TNI object
rtni <- tni.constructor(expData=tniData$expData, 
        regulatoryElements=c("PTTG1","E2F2","FOXM1","E2F3","RUNX2"), 
        rowAnnotation=tniData$rowAnnotation)
rtni <- tni.permutation(rtni)
rtni <- tni.bootstrap(rtni)
rtni <- tni.dpi.filter(rtni)

#load a gene set collection
#here, we build three random gene sets for demonstration
geneset1 <- sample(tniData$rowAnnotation$SYMBOL,50)
geneset2 <- sample(tniData$rowAnnotation$SYMBOL,50)
geneset3 <- sample(tniData$rowAnnotation$SYMBOL,50)
geneSetList <- list(geneset1=geneset1,
                    geneset2=geneset2,
                    geneset3=geneset3)

#compute single-sample GSEA
#note: regulons are not required for this function, 
#as it will assess the samples in the TNI object
ES <- tni.annotate.samples(rtni, geneSetList)


## End(Not run)

Description

Uses aREA 3-tail algorithm to compute regulon activity for TNI-class objects.

Usage

tni.area3(object, minRegulonSize=15, sizeFilterMethod="posORneg", scale=FALSE, tnet="dpi", 
regulatoryElements=NULL, samples=NULL, features=NULL, refsamp=NULL, log=FALSE, verbose=TRUE, 
doSizeFilter=NULL)

Arguments

Value

a list with enrichment scores for all samples in the TNI.

References

Alvarez et al. Functional characterization of somatic mutations in cancer using network-based inference of protein activity. Nature Genetics, 48(8):838-847, 2016.

See Also

TNI-class aREA

Examples

data(tniData)

## Not run: 

rtni <- tni.constructor(expData=tniData$expData, 
        regulatoryElements=c("PTTG1","E2F2","FOXM1","E2F3","RUNX2"), 
        rowAnnotation=tniData$rowAnnotation)
rtni <- tni.permutation(rtni)
rtni <- tni.bootstrap(rtni)
rtni <- tni.dpi.filter(rtni)

#run aREA algorithm
rtni <- tni.area3(rtni)

#get results
regulonActivity <- tni.get(rtni, what = "regulonActivity")



## End(Not run)

Description

This function takes a TNI object and returns the consensus transcriptional network.

Usage

tni.bootstrap(object, nBootstraps=100, consensus=95, parChunks=NULL, verbose=TRUE)

Arguments

Value

a matrix in the slot "results" containing a reference transcriptional network, see 'tn.ref' option in tni.get.

Author(s)

Mauro Castro

See Also

TNI-class

Examples

data(tniData)

## Not run: 

# preprocessing
rtni <- tni.constructor(expData=tniData$expData, 
        regulatoryElements=c("PTTG1","E2F2","FOXM1","E2F3","RUNX2"), 
        rowAnnotation=tniData$rowAnnotation)

# linear version!
rtni <- tni.permutation(rtni)
rtni <- tni.bootstrap(rtni)

## parallel version with SNOW package!
#library(snow)
#options(cluster=snow::makeCluster(3, "SOCK"))
#rtni <- tni.permutation(rtni)
#rtni <- tni.bootstrap(rtni)
#stopCluster(getOption("cluster"))

## End(Not run)

Description

This function takes a TNI object and a list of candidate modulators, and computes the conditional mutual information over the TF-target interactions in a transcriptional network (with multiple hypothesis testing corrections). For each TF, the method measures the change in the mutual information between the TF and its targets conditioned to the gene expression of a modulator.

Usage

tni.conditional(object, modulators, tfs=NULL, sampling=35, pValueCutoff=0.01, 
pAdjustMethod="bonferroni", minRegulonSize=15, minIntersectSize=5, 
miThreshold="md", prob=0.99, medianEffect=FALSE, 
iConstraint=TRUE, verbose=TRUE, ...)

Arguments

Value

a data frame in the slot "results", see 'cdt' option in tni.get.

Author(s)

Mauro Castro

References

Wang, K. et al. Genome-wide identification of post-translational modulators of transcription factor activity in human B cells. Nat Biotechnol, 27(9):829-39, 2009.

Castro, M.A.A. et al. RTN: Reconstruction and Analysis of Transcriptional Networks. Journal Paper (in preparation), 2012.

See Also

TNI-class

Examples

data(tniData)

## Not run: 

# preprocessing
rtni <- tni.constructor(expData=tniData$expData, 
        regulatoryElements=c("PTTG1","E2F2","FOXM1","E2F3","RUNX2"), 
        rowAnnotation=tniData$rowAnnotation)

# permutation/bootstrap analysis (infers the reference/relevance network)
rtni <- tni.permutation(rtni)
rtni <- tni.bootstrap(rtni)

# dpi filter (infers the transcriptional network)
rtni <- tni.dpi.filter(rtni)

# get some candidate modulators for demonstration!
mod4test <- rownames(rtni@gexp)[sample(1:nrow(rtni@gexp),200)]

# conditional analysis
rtni <- tni.conditional(rtni, modulators=mod4test, pValueCutoff=1e-3)

# get results
cdt <- tni.get(rtni, what="cdt.table")

# get summary on a graph object
g <- tni.graph(rtni, gtype="mmap")

###---------------------------------------------
### optional: plot the igraph object using RedeR
library(RedeR)

#--load reder interface
rdp <- RedPort()
calld(rdp)

#---add graph and legends
addGraph(rdp,g)
addLegend.shape(rdp,g)
addLegend.size(rdp,g)
addLegend.color(rdp,g,type="edge")
relax(rdp,p1=50,p5=20)


## End(Not run)

Description

This function is the main entry point of the TNI pipeline.

Usage

tni.constructor(expData, regulatoryElements, rowAnnotation=NULL, 
colAnnotation=NULL, cvfilter=FALSE, verbose=TRUE)

Arguments

Value

A pre-processed TNI-class object.

Author(s)

Mauro Castro

See Also

TNI-class

Examples

data(tniData)

#--- run constructor
rtni <- tni.constructor(expData=tniData$expData, 
        regulatoryElements=c("PTTG1","E2F2","FOXM1","E2F3","RUNX2"), 
        rowAnnotation=tniData$rowAnnotation)

Description

A minimum TNI object that can be used to demonstrate RTN functionalities.

Usage

data(stni)

Format

stniA TNI-class with a subset of samples and genes from the Fletcher2013b package.

Details

The TNI consists of a TNI-class with a subsetted gene expression matrix and reduced list of transcription factors. It should be regarded as a toy example for demonstration purposes only, despite being extracted, pre-processed and size-reduced from Fletcher et al. (2013) and Curtis et al. (2012).

Value

a TNI-class.

References

Fletcher M.N.C. et al., Master regulators of FGFR2 signalling and breast cancer risk. Nature Communications, 4:2464, 2013.

Curtis C. et al., The genomic and transcriptomic architecture of 2,000 breast tumours reveals novel subgroups. Nature 486, 7403. 2012.

Examples

data(stni)

Description

This function takes a TNI object and returns the transcriptional network filtered by the data processing inequality algorithm.

Usage

tni.dpi.filter(object, eps = 0, sizeThreshold = TRUE, minRegulonSize = 15, verbose = TRUE)

Arguments

Value

a mutual information matrix in the slot "results" containing a dpi-filtered transcriptional network, see 'tn.dpi' option in tni.get.

Author(s)

Mauro Castro

See Also

TNI-class

Examples

data(tniData)

## Not run: 

# preprocessing
rtni <- tni.constructor(expData=tniData$expData, 
        regulatoryElements=c("PTTG1","E2F2","FOXM1","E2F3","RUNX2"), 
        rowAnnotation=tniData$rowAnnotation)
        
# permutation analysis (infers the reference/relevance network)
rtni <- tni.permutation(rtni)

# dpi filter (infers the transcriptional network)
rtni <- tni.dpi.filter(rtni)


## End(Not run)

Description

Get available results from individual slots in a TNI object.

Usage

tni.get(object, what="summary", order=TRUE, ntop=NULL, reportNames=TRUE, 
idkey=NULL)

Arguments

Details

Options for the 'what' argument:

summary

A list summarizing parameters and results available in the TNI object (see tni.regulon.summary for a summary of the network and regulons).

status

A vector indicating the status of each available method in the pipeline.

para

A list with the parameters used by each available method in the pipeline.

gexp

A gene expression matrix.

regulatoryElements

A vector of regulatory elements (e.g. transcription factors).

targetElements

A vector of target elements (e.g. TF targets).

modulators

A vector of modulators (e.g. TF modulators).

tnet

A data matrix with MI values, evaluated by the DPI filter. MI values are computed between regulators and targets, with regulators on cols and targets on rows. Note that signals (+/-) are assigned to the inferred associations in order to represent the 'mode of action', which is derived from Pearson's correlation between regulators and targets.

refnet

A data matrix with MI values (not evaluated by the DPI filter). MI values are computed between regulators and targets, with regulators on cols and targets on rows. Note that signals (+/-) are assigned to the inferred associations in order to represent the 'mode of action', which is derived from Pearson's correlation between regulators and targets.

regulons

A list with regulons extracted from the 'tnet' data matrix.

refregulons

A list with regulons extracted from the 'refnet' data matrix.

regulons.and.mode

A list with regulons extracted from the 'tnet' data matrix, including the assiged 'mode of action'.

refregulons.and.mode

A list with regulons extracted from the 'refnet' data matrix, including the assiged 'mode of action'.

rowAnnotation

A data frame with probe-to-gene annotation.

colAnnotation

A data frame with sample annotation.

cdt.table

A data frame with results from the tni.conditional analysis pipeline.

cdt.list

A list with results from the tni.conditional analysis pipeline.

regulonSize

A data frame with the number of targets annotated in each regulon.

regulonActivity

A list with results from the tni.gsea2 analysis pipeline.

Value

Get the slot content from a TNI-class object.

Author(s)

Mauro Castro

Examples

data(tniData)

## Not run: 

rtni <- tni.constructor(expData=tniData$expData, 
        regulatoryElements=c("PTTG1","E2F2","FOXM1","E2F3","RUNX2"), 
        rowAnnotation=tniData$rowAnnotation)
rtni <- tni.permutation(rtni)
rtni <- tni.bootstrap(rtni)
rtni <- tni.dpi.filter(rtni)

# check summary
tni.get(rtni, what="summary")

# get regulons
regulons <- tni.get(rtni, what = "regulons")

# get status of the pipeline
tni.get(rtni, what="status")

## End(Not run)

Description

Extract results from a TNI object and compute a graph (Deprecated).

Usage

tni.graph(x, ...)

Arguments

Value

—

Examples

# deprecated function

Description

Uses GSEA2 algorithm to compute regulon activity for TNI-class objects.

Usage

tni.gsea2(object, minRegulonSize=15, sizeFilterMethod="posORneg", scale=FALSE, 
exponent=1, tnet="dpi", regulatoryElements=NULL, features=NULL, samples=NULL, 
refsamp=samples, log=TRUE, alternative=c("two.sided", "less", "greater"), 
targetContribution=FALSE, additionalData=FALSE, verbose=TRUE, doSizeFilter=NULL)

Arguments

Value

a list with enrichment scores for all samples in the TNI. The list contains the following elements:

differential:

A numeric "matrix" with differential enrichment scores (dES).

positive:

A numeric "matrix" with enrichment scores from positive targets.

negative:

A numeric "matrix" with enrichment scores from negative targets.

status:

A numeric "matrix" with discretized scores derived from the dES values.

regulatoryElements:

A character vector listing the regulatory elements assessed by the GSEA-2T algorithm.

sections:

A single numeric value used in internal plots.

Author(s)

Mauro Castro

See Also

TNI-class tna.gsea2 tna.plot.gsea2

Examples

data(tniData)

## Not run: 

rtni <- tni.constructor(expData=tniData$expData, 
        regulatoryElements=c("PTTG1","E2F2","FOXM1","E2F3","RUNX2"), 
        rowAnnotation=tniData$rowAnnotation)
rtni <- tni.permutation(rtni)
rtni <- tni.bootstrap(rtni)
rtni <- tni.dpi.filter(rtni)

#run GSEA2 analysis pipeline
rtni <- tni.gsea2(rtni)

#get results
regulonActivity <- tni.get(rtni, what = "regulonActivity")

#parallel version with SNOW package!
library(snow)
options(cluster=snow::makeCluster(3, "SOCK"))
rtni <- tni.gsea2(rtni)
stopCluster(getOption("cluster"))


## End(Not run)

Description

This function tests the overlap between gene sets and regulons.

Usage

tni.overlap.genesets(object, geneSetList, regulatoryElements = NULL, 
    minSetSize = 15, sizeFilterMethod="posORneg",
    method = c("HT","JC"), pValueCutoff = 0.05, 
    pAdjustMethod = "BH", verbose = TRUE)

Arguments

Details

The 'HT' option assesses the overlap between gene sets and regulons using a hypergeometric test, and returns a data frame with the overlap statistics. The 'JC' option assesses the overlap between gene sets and regulons using the Jaccard Coefficient (JC), and retuns a matrix with JC values.

Value

Either a data frame or a numeric matrix with association statistics between gene sets vs. regulons.

Author(s)

Mauro Castro

See Also

TNI-class

Examples

data(tniData)

## Not run: 

#compute regulons
rtni <- tni.constructor(expData=tniData$expData, 
        regulatoryElements=c("PTTG1","E2F2","FOXM1","E2F3","RUNX2"), 
        rowAnnotation=tniData$rowAnnotation)
rtni <- tni.permutation(rtni)
rtni <- tni.bootstrap(rtni)
rtni <- tni.dpi.filter(rtni)

#load a gene set collection 
#here, we build three random gene sets for demonstration
geneset1 <- sample(tniData$rowAnnotation$SYMBOL,50)
geneset2 <- sample(tniData$rowAnnotation$SYMBOL,50)
geneset3 <- sample(tniData$rowAnnotation$SYMBOL,50)
geneSetList <- list(geneset1=geneset1,
                    geneset2=geneset2,
                    geneset3=geneset3)

#run the overlap analysis
ovstats <- tni.overlap.genesets(rtni, geneSetList, pValueCutoff = 1)


## End(Not run)

Description

This function takes a TNI object and returns a transcriptional network inferred by mutual information (with multiple hypothesis testing corrections).

Usage

tni.permutation(object, pValueCutoff=0.01, pAdjustMethod="BH", globalAdjustment=TRUE,
        estimator="spearman", nPermutations=1000, pooledNullDistribution=TRUE, 
        boxcox=TRUE, parChunks=NULL, verbose=TRUE)

Arguments

Value

a mutual information matrix in the slot "results" containing a reference transcriptional network, see 'tn.ref' option in tni.get.

Author(s)

Mauro Castro

References

Dam et al. Gene co-expression analysis for functional classification and gene-disease predictions. Brief Bioinform. 2018 Jul 20;19(4):575-592. doi: 10.1093/bib/bbw139.

See Also

TNI-class

Examples

data(tniData)

## Not run: 

# preprocessing
rtni <- tni.constructor(expData=tniData$expData, 
        regulatoryElements=c("PTTG1","E2F2","FOXM1","E2F3","RUNX2"), 
        rowAnnotation=tniData$rowAnnotation)

# linear version (set nPermutations >= 1000)
rtni <- tni.permutation(rtni, nPermutations = 100)

## parallel version with SNOW package!
#library(snow)
#options(cluster=snow::makeCluster(3, "SOCK"))
#rtni<-tni.permutation(rtni)
#stopCluster(getOption("cluster"))

## End(Not run)

Description

This funtion can help to check whether the numbers of positive and negative targets are reasonably well balanced in the regulons.

Usage

tni.plot.checks(object, minRegulonSize = 15, option = c("barplot","edf","points"))

Arguments

Value

A plot showing the distribution of regulons' positive and negative targets.

Note

We have observed that transcription factor (TF) regulons reconstructed from RTN exhibit different proportions of positive and negative targets. While the proportion can vary between different regulons, we have observed a consistent higher proportion of positive targets, especially when using RNA-seq data. RTN uses mutual information (MI) to assess TF-target associations, assigning the direction of the inferred associations by Spearman's correlations. Dam et al. (2018) have acknowledged that different RNA-seq normalization methods introduce different biases in co-expression analysis, usually towards positive correlation, possibly affected by read-depth differences between samples and the large abundance of 0 values present in RNA-seq-derived expression matrices. This funtion can help to check whether the numbers of positive and negative target genes are reasonably well balanced in the regulons.

Author(s)

Mauro Castro, Gordon Robertson

References

Dam et al. Gene co-expression analysis for functional classification and gene-disease predictions. Brief Bioinform. 2018 Jul 20;19(4):575-592. doi: 10.1093/bib/bbw139.

Examples

data(tniData)

## Not run: 

# preprocessing
rtni <- tni.constructor(expData=tniData$expData, 
        regulatoryElements=c("PTTG1","E2F2","FOXM1","E2F3","RUNX2"), 
        rowAnnotation=tniData$rowAnnotation)

# compute regulons
rtni <- tni.permutation(rtni, nPermutations = 1000)
rtni <- tni.permutation(rtni)
rtni <- tni.dpi.filter(rtni)

# check target distribution
tni.plot.checks(rtni)


## End(Not run)

Description

This method plots the results of the subgroup regulon enrichment analysis in a heatmap. The rows of the heatmap represent enriched regulons, while the columns show the subgroups. The plotted values correspond to average regulon activity for a regulon in a subgroup. Enriched values can be marked.

Usage

tni.plot.sre(object, nGroupsEnriched = NULL, nTopEnriched = NULL, 
             colors = c("blue","white","red"), breaks = seq(-1.5, 1.5, 0.1), 
             markEnriched = TRUE, ...)

Arguments

Value

A heatmap of the subgroup regulon enrichment results.

See Also

tni.sre

Examples

# load tniData
data(tniData)

## Not run: 

# preprocessing
rtni <- tni.constructor(expData=tniData$expData, 
        regulatoryElements=c("PTTG1","E2F2","FOXM1","E2F3","RUNX2"), 
        rowAnnotation=tniData$rowAnnotation)
        
# permutation analysis (infers the reference/relevance network)
rtni <- tni.permutation(rtni)

# dpi filter (infers the transcriptional network)
rtni <- tni.dpi.filter(rtni)

#run GSEA2 analysis pipeline
rtni <- tni.gsea2(rtni)

# set sample groups
colAnnotation <- tni.get(rtni, "colAnnotation")
sampleGroups <- list(G1=colAnnotation$ID[1:60], 
                G2=colAnnotation$ID[61:90], 
                G3=colAnnotation$ID[91:120])

# run subgroup regulon enrichment analysis
rtni <- tni.sre(rtni, sampleGroups)

# plot results
tni.plot.sre(rtni)

## End(Not run)

Description

This is a generic function, provides all preprocessing methods for the 'tni.constructor' function.

Usage

tni.preprocess(object, rowAnnotation=NULL, colAnnotation=NULL, cvfilter=FALSE, 
    verbose=TRUE)

Arguments

Value

A pre-processed TNI-class object.

Author(s)

Mauro Castro

See Also

tni.constructor

Examples

## see 'tni.constructor'!

Description

Uses network pruning methods to compute a 'core' regulon that retains good correlation with original regulon activity.

Usage

tni.prune(object, regulatoryElements = NULL, minRegCor = 0.95, 
tarPriorityMethod = "EC", minPrunedSize = 30, verbose = TRUE, ...)

Arguments

Value

a TNI-class object, with the pruned regulons.

Author(s)

Clarice Groeneveld

See Also

TNI-class tni.gsea2

Examples

data(tniData)

## Not run: 

rtni <- tni.constructor(expData=tniData$expData, 
        regulatoryElements=c("PTTG1","E2F2","FOXM1","E2F3","RUNX2"), 
        rowAnnotation=tniData$rowAnnotation)
rtni <- tni.permutation(rtni)
rtni <- tni.bootstrap(rtni)
rtni <- tni.dpi.filter(rtni)

# prune the PTTG1 regulon
rtni_pruned <- tni.prune(rtni, "PTTG1", tarPriorityMethod = "TC")

#parallel version with SNOW package!
#library(snow)
#options(cluster=makeCluster(3, "SOCK"))
#rtni_pruned <- tni.prune(rtni, c("PTTG1", "E2F2"))
#stopCluster(getOption("cluster"))


## End(Not run)

Description

This function takes a TNI object and optionally a list of regulatory elements and returns a summary of the network (if no regulatory elements are given) or of the chosen regulon or regulons.

Usage

tni.regulon.summary(object, regulatoryElements = NULL, verbose = TRUE)

Arguments

Value

It returns a print-out of the network summary (if verbose is TRUE) and invisibly returns a data.frame of network characteristics such as regulon size and regulon balance.

Author(s)

Clarice Groeneveld

See Also

TNI-class

Examples

data(tniData)

## Not run: 

rtni <- tni.constructor(expData=tniData$expData, 
        regulatoryElements=c("PTTG1","E2F2","FOXM1","E2F3","RUNX2"), 
        rowAnnotation=tniData$rowAnnotation)
rtni <- tni.permutation(rtni)
rtni <- tni.bootstrap(rtni)
rtni <- tni.dpi.filter(rtni)

# Summary of the network
tni.regulon.summary(rtni)

# Summary of a regulon
tni.regulon.summary(rtni, regulatoryElements = "PTTG1")


## End(Not run)

Description

This function replaces samples of an existing TNI-class objects.

Usage

tni.replace.samples(object, expData, rowAnnotation=NULL, colAnnotation=NULL, 
removeRegNotAnnotated=TRUE, verbose=TRUE)

Arguments

Value

A TNI-class object.

Author(s)

Mauro Castro

Examples

## please the package's vignette

Description

This method evaluates which regulons are enriched in sample groups, given a grouping variable. It performs Fisher's Exact Test whether a regulon is positively or negatively enriched in a subgroup using regulon activity.

Usage

tni.sre(object, sampleGroups, regulatoryElements = NULL, 
        pValueCutoff = 0.05, pAdjustMethod = "BH")

Arguments

Value

A TNI-class object with the results of the subgroup regulon enrichment added to the results slot. To recover the results, use tni.get(object, "regulonEnrichment")

See Also

tni.plot.sre

Examples

# load tniData
data(tniData)

## Not run: 

# compute regulons
rtni <- tni.constructor(expData=tniData$expData, 
        regulatoryElements=c("PTTG1","E2F2","FOXM1","E2F3","RUNX2"), 
        rowAnnotation=tniData$rowAnnotation)
rtni <- tni.permutation(rtni)
rtni <- tni.bootstrap(rtni)
rtni <- tni.dpi.filter(rtni)

#run GSEA2 analysis pipeline
rtni <- tni.gsea2(rtni)

# set sample groups
colAnnotation <- tni.get(rtni, "colAnnotation")
sampleGroups <- list(G1=colAnnotation$ID[1:60], 
                G2=colAnnotation$ID[61:90], 
                G3=colAnnotation$ID[91:120])

# run subgroup regulon enrichment analysis
rtni <- tni.sre(rtni, sampleGroups)

# get results
tni.get(rtni, "subgroupEnrichment")

# for a heatmap representation, see the tni.plot.sre() function.

## End(Not run)

Description

This is a generic function.

Usage

tni2tna.preprocess(object, phenotype=NULL, hits=NULL, phenoIDs=NULL, 
duplicateRemoverMethod="max", verbose=TRUE)

Arguments

Value

A pre-processed TNA-class object.

Author(s)

Mauro Castro

See Also

TNI-class TNA-class

Examples

data(tniData)
data(tnaData)

## Not run: 

rtni <- tni.constructor(expData=tniData$expData, 
        regulatoryElements=c("PTTG1","E2F2","FOXM1","E2F3","RUNX2"), 
        rowAnnotation=tniData$rowAnnotation)
rtni <- tni.permutation(rtni)
rtni <- tni.bootstrap(rtni)
rtni <- tni.dpi.filter(rtni)
rtna <- tni2tna.preprocess(rtni, phenotype=tnaData$phenotype, 
        hits=tnaData$hits, phenoIDs=tnaData$phenoIDs)

## End(Not run)

Description

This function provides compatibility checks for a TNI class object.

Usage

upgradeTNA(object)

Arguments

Value

An updated TNA-class object.

Author(s)

Mauro Castro

Examples

### Objects of class TNA generated by RTN (version <= 1.15.2) can be upgraded
### to the latest version by calling upgradeTNA().

Description

This function provides compatibility checks for a TNI class object.

Usage

upgradeTNI(object)

Arguments

Value

An updated TNI-class object.

Author(s)

Mauro Castro

Examples

### Objects of class TNI generated by RTN (version <= 1.15.2) can be upgraded
### to the latest version by calling upgradeTNI().