10、差异基因topGO富集

参考：http://www.biotrainee.com/thread-558-1-1.html

http://bioconductor.org/packages/3.7/bioc/

http://www.bioconductor.org/packages/release/bioc/html/topGO.html

https://www.jianshu.com/p/9e21f2196178

https://rpubs.com/aemoore62/TopGo_colMap_Func_Troubleshoot

构建topGOdata对象的3个数据

1. 基因某种ID的列表（可以有另一个对应的分数值，如p值或t统计量，或者是差异表达值）
2. 基因的这种ID与GO的映射表，在ID为芯片的探针ID时，可以直接使用bioconductor的芯片注释包如hgu95av2.db包
3. GO的层次关系数据，这个结果可以从GO.db包获得，topGO也只支持GO.db包定义的层次结

topGO使用：

首先，我们制作准备文件，CC、BP、MF三个注释文件，格式为：基因ID GO:xxx,GO:yyy（topGO.map）

然后准备我们的差异基因列表，两列差异基因ID FDR。（topGO.list）

 

library("topGO")
geneID2GO<-readMappings(choose.files())　　##读取所有基因注释信息
geneNames<- names(geneID2GO)

data<-read.table(choose.files(), row.names = 1, header=TRUE,check.names =F)　　##读取差异基因的ID
geneList<-data[,1]
names(geneList) <- rownames(data)
topDiffGenes<-function(allScore){return(allScore<0.05)}
1.1、###BP
sampleGOdata <- new("topGOdata",nodeSize = 6,ontology="BP", allGenes = geneList,annot = annFUN.gene2GO, gene2GO = geneID2GO,geneSel=topDiffGenes)
resultKS.elim <- runTest(sampleGOdata, algorithm = "elim", statistic = "ks")
allRes <- GenTable(sampleGOdata,KS = resultKS.elim,ranksOf = "classic", topNodes = attributes(resultKS.elim)$geneData[4])
write.table(allRes, file="T01_vs_T02.topGO_BP.xls", sep=" ", quote=FALSE, col.names=TRUE, row.names=FALSE)

pdf("T01_vs_T02.topGO_BP.pdf")
showSigOfNodes(sampleGOdata, score(resultKS.elim), firstSigNodes = 10, useInfo = "all")　　##作图
dev.off()

png("T01_vs_T02.topGO_BP.png")
showSigOfNodes(sampleGOdata, score(resultKS.elim), firstSigNodes = 10, useInfo = "all")
dev.off()
1.2、##MF
sampleGOdata <- new("topGOdata",nodeSize = 6,ontology="MF", allGenes = geneList, annot = annFUN.gene2GO, gene2GO = geneID2GO,geneSel=topDiffGenes)
resultKS.elim <- runTest(sampleGOdata, algorithm = "elim", statistic = "ks")
allRes <- GenTable(sampleGOdata,KS = resultKS.elim,ranksOf = "classic", topNodes = attributes(resultKS.elim)$geneData[4])
write.table(allRes, file="T01_vs_T02.topGO_MF.xls", sep=" ", quote=FALSE, col.names=TRUE, row.names=FALSE)
pdf("T01_vs_T02.topGO_MF.pdf")
showSigOfNodes(sampleGOdata, score(resultKS.elim), firstSigNodes = 10, useInfo = "all")
dev.off()

png("T01_vs_T02.topGO_MF.png")
showSigOfNodes(sampleGOdata, score(resultKS.elim), firstSigNodes = 10, useInfo = "all")
dev.off()
1.3、##CC
sampleGOdata <- new("topGOdata",nodeSize = 6,ontology="CC", allGenes = geneList, annot = annFUN.gene2GO, gene2GO = geneID2GO,geneSel=topDiffGenes)
resultKS.elim <- runTest(sampleGOdata, algorithm = "elim", statistic = "ks")
allRes <- GenTable(sampleGOdata,KS = resultKS.elim,ranksOf = "classic", topNodes = attributes(resultKS.elim)$geneData[4])
write.table(allRes, file="T01_vs_T02.topGO_CC.xls", sep=" ", quote=FALSE, col.names=TRUE, row.names=FALSE)
pdf("T01_vs_T02.topGO_CC.pdf")
showSigOfNodes(sampleGOdata, score(resultKS.elim), firstSigNodes = 10, useInfo = "all")
dev.off()

png("T01_vs_T02.topGO_CC.png")
showSigOfNodes(sampleGOdata, score(resultKS.elim), firstSigNodes = 10, useInfo = "all")
dev.off()

1.4、##输出每个go的基因以及注释到这个go的差异基因[可以不做这个]

allGO =usedGO(object = sampleGOdata)

for (gos in allGO){

goID <-gos;

gene.universe <- genes(sampleGOdata);

go.genes <- genesInTerm(sampleGOdata,goID)[[1]];

sig.genes <- sigGenes(sampleGOdata);

file1=paste("GO-TMP_BP_sig_",gos,sep="");

write.table(sig.genes,file=file1);

file2=paste("GO-TMP_BP_go_",gos,sep="");

write.table(go.genes,file=file2);

}

2、来自生信技能树

###BP
sampleGOdata <- new("topGOdata",nodeSize = 6,ontology="BP", allGenes = geneList,annot = annFUN.gene2GO, gene2GO = geneID2GO,geneSel=topDiffGenes)

allGO =usedGO(object = sampleGOdata)

resultCFis<-runTest(sampleGOdata,algorithm="classic",statistic="fisher")

gtFis<-GenTable(sampleGOdata,classicFisher=resultCFis,orderBy="classic",ranksOf="classicFisher",topNodes=length(allGO))

fdr<-p.adjust(p=gtFis[,"classicFisher"],method="fdr")

r <-cbind(gtFis,fdr)

write.table(r,file="topGO_BP.xls",sep=" ")

showSigOfNodes(GOdata,score(resultCFis),firstSigNodes= 5, useInfo = "all")