inst/examples/analysisAsPerEmail.R

In umbibio/CIE-R-Package: Causal Inference and Directional Enrichment Methods on Biological Networks

## Conditions
files <- list.files("../CIEdraft/", "edgeR.txt")
dges <- lapply(paste("../CIEdraft/", files, sep=""), function(x) {
    read.table(x, header=T, sep="\t") } )

names <- strsplit(files, "\\.")
names(dges) <- lapply(names, function(x) {x[1]})

## Analyses to run
## 1) StringDB, using Quaternary Method
enrichmentST1 <- runCIE(databaseType="string",
                     DGEs = dges[[1]],
                     methods ="Quaternary",
                     useFile=TRUE, useMart=TRUE,
                     useBHLH=TRUE,
                     martFN="../../CIE/data/mart_human_TFs.csv",
                     BHLHFN="../../CIE/data/BHLH_TFs.txt",
                     targetsOfInterest= c("COL1A1", "COL1A2"),
                     hypTabs="1",
                     databaseDir="../../CIE/data/")

## lapply(1:length(enrichmentST), function(x) {
##     write.table(enrichmentST[[x]], paste("QuaternaryString",
##                                        names(enrichmentST)[x], sep="")) } )
## entsST <- read.table("../CIE/data/string.ents", sep="\t", stringsAsFactors=F)
## entsST <- read.table("../CIE/data/string.rels", sep="\t", stringsAsFactors=F)
## save(enrichmentST, entsST, relsST, dges)
## createCytoGraph(enrichmentST, entsST, relsST, dges)

## 2) BEL, using Quaternary method

## 3) Trust, using Fisher
enrichmentTR <- runCIE(databaseType="TRRUST",
                     DGEs = dges[[3]],
                     method="Fisher",
                     useFile=TRUE, useMart=TRUE,
                     useBHLH=TRUE,
                     martFN="../CIE/data/mart_human_TFs.csv",
                     BHLHFN="../CIE/data/BHLH_TFs.txt",
                     targetsOfInterest= c("COL1A1", "COL1A2"),
                     hypTabs="1",
                     databaseDir="../CIE/data/")

lapply(1:length(enrichmentTR), function(x) {
    write.table(enrichmentTR[[x]], paste("FisherTRRUST",
                                       names(enrichmentTR)[x], sep="")) } )
## entsTR <- read.table("../CIE/data/trrust.ents", sep="\t", 
##                      header=T, stringsAsFactors=F)
## relsTR <- read.table("../CIE/data/trrust.rels", sep="\t", 
##                      header=T, stringsAsFactors=F)
## save(enrichmentTR, entsTR, relsTR, dges)
## createCytoGraph(enrichmentTR, entsTR, relsTR, dges)

## Not currently available

## 4) Combined all above using Quaternary.

## Then move on to ChiP and perform Fisher method using
## 1) High stringency (1KB with cutoff score of 500 in at least one experiment)
ChIP1min500 <- filterChIPAtlas(1, 500, "min", writeToFile = FALSE)
enrichment1min500 <- runCIE(NULL, NULL, DGEs=dges,
                     methods="Fisher", useFile=F,
                     useMart=TRUE,
                     useBHLH=TRUE,
                     martFN="../CIE/data/mart_human_TFs.csv",
                     BHLHFN="../CIE/data/BHLH_TFs.txt",
                     targetsOfInterest= c("COL1A1", "COL1A2"),
                     ents=ChIP1min500$filteredChIP.ents,
                     rels=ChIP1min500$filteredChIP.rels, hypTabs="1")

lapply(1:length(enrichment1min500), function(x) {
   write.table(enrichment[[x]], paste("FisherChIP1min500",
                                      names(enrichment)[x], sep=""),
               sep="\t", row.names=F) } )
## createCytoGraph(enrichment1min500, ChIP1min500$filteredChIP.ents,
##                 data.frame(ChIP1min500$filteredChIP.rels), dges)

## 2) Mid stringency (5 KB with cutoff of 250 in at least one experiment)
ChIP5min250 <- filterChIPAtlas(5, 250, "min", writeToFile = FALSE)
enrichment5min250 <- runCIE(NULL, NULL, DGEs=dges,
                            methods="Fisher", useFile=F,
                            useMart=TRUE,
                            useBHLH=TRUE,
                            martFN="../CIE/data/mart_human_TFs.csv",
                            BHLHFN="../CIE/data/BHLH_TFs.txt",
                            targetsOfInterest= c("COL1A1", "COL1A2"),
                            ents=ChIP5min250$filteredChIP.ents,
                            rels=ChIP5min250$filteredChIP.rels, hypTabs="1")
lapply(1:length(enrichment5min250), function(x) {
   write.table(enrichment5min250[[x]], paste("FisherChIP5min250",
                                      names(enrichment)[x], sep=""),
               sep="\t", row.names=F) } )
## createCytoGraph(enrichment5min250, ChIP5min250$filteredChIP.ents,
##                 data.frame(ChIP5min250$filteredChIP.rels), dges)

## 3) Mid stringency (5 KB with cutoff of 250 on average)
ChIP5ave250 <- filterChIPAtlas(5, 250, "average", writeToFile = FALSE)
enrichment5ave250 <- runCIE(NULL, NULL, DGEs=dges,
                            methods="Fisher", useFile=F,
                            useMart=TRUE,
                            useBHLH=TRUE,
                            martFN="../CIE/data/mart_human_TFs.csv",
                            BHLHFN="../CIE/data/BHLH_TFs.txt",
                            targetsOfInterest= c("COL1A1", "COL1A2"),
                            ents=ChIP5ave250$filteredChIP.ents,
                            rels=ChIP5ave250$filteredChIP.rels, hypTabs="1")
lapply(1:length(enrichment5ave250), function(x) {
   write.table(enrichment5ave250[[x]], paste("FisherChIP5ave250",
                                      names(enrichment)[x], sep=""),
               sep="\t", row.names=F) } )

## createCytoGraph(enrichment5ave250, ChIP5ave250$filteredChIP.ents,
##                 data.frame(ChIP5ave250$filteredChIP.rels), dges)


## 4) Matched cell lines (5 KB with cutoff of 250). 
ChIP5ave250p <- filterChIPAtlas(5, 250, "average", cellLineType="prostate",
                                writeToFile = FALSE)
enrichment5ave250p <- runCIE(NULL, NULL, DGEs=dges,
                             methods="Fisher", useFile=F,
                             useMart=TRUE,
                             useBHLH=TRUE,
                             martFN="../CIE/data/mart_human_TFs.csv",
                             BHLHFN="../CIE/data/BHLH_TFs.txt",
                             targetsOfInterest= c("COL1A1", "COL1A2"),
                             ents=ChIP5ave250p$filteredChIP.ents,
                             rels=ChIP5ave250p$filteredChIP.rels, hypTabs="1")
lapply(1:length(enrichment5ave250p), function(x) {
   write.table(enrichment5ave250p[[x]], paste("FisherChIP5ave250p",
                                      names(enrichment)[x], sep=""),
               sep="\t", row.names=F) } )

## createCytoGraph(enrichment5ave250p, ChIP5ave250p$filteredChIP.ents,
##                 data.frame(ChIP5ave250p$filteredChIP.rels), dges)

## Another analysis, this time on my own idea
ChIP1autoP <- filterChIPAtlas(1, NULL, "auto", cellLineType="prostate",
                              writeToFile = FALSE, databaseDir="../CIEdraft/")
enrichment1autoP <- runCIE(NULL, NULL, DGEs = dges,
                           methods = "Fisher", useFile=F,
                           useMart=TRUE,
                           useBHLH=TRUE,
                           martFN="../CIE/data/mart_human_TFs.csv",
                           BHLHFN="../CIE/data/BHLH_TFs.txt",
                           targetsOfInterest= c("COL1A1", "COL1A2"),
                           ents=ChIP1autoP$filteredChIP.ents,
                           rels=ChIP1autoP$filteredChIP.rels,
                           hypTabs = "1",
                           databaseDir="../CIE/data/")
createCytoGraph(enrichment1autoP, ChIP1autoP$filteredChIP.ents, 
                data.frame(ChIP1autoP$filteredChIP.rels), dges)

lapply(1:length(enrichment1autoP), function(x) {
    write.table(enrichment1autoP[[x]], paste("QuaternaryChIP1autoP",
                                       names(enrichment1autoP)[x], sep=""),
                sep="\t", row.names=F) } )


enrichments <- list(enrichmentTR, enrichment1min500, enrichment5min250,
                    enrichment5ave250, enrichment5ave250p, enrichment1autoP)
names(enrichments) <- c("enrichmentTR", "enrichment1min500", "enrichment5min250",
                        "enrichment5ave250", "enrichment5ave250p",
                        "enrichment1autoP")

top5 <- lapply(enrichments, function(x) {
    lapply(x, function(y) {
        y$name[1:5] } ) } )