geo差异表达分析_GEO数据库下载表达矩阵之后做下游分析

首先打开网页：

https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE104168

然后下载处理好的数据：

Excel打开：

然后就可以开始进行比较了：

新建一个文件，并且从刚才的表达矩阵提取我们需要的几列数据：

本例选择：

然后就在R进行操作：

#安装limma包，需要上网搜索最新的安装方式if (!requireNamespace("BiocManager", quietly = TRUE))  install.packages("BiocManager")BiocManager::install("limma")#导入包library("limma")library("impute")#设置参数logFoldChange=1adjustP=0.05getwd()setwd("")          #设置工作目录rt=read.table("compare.txt",sep="\t",header=T,check.names=F)    #读取文件rt=as.matrix(rt)rownames(rt)=rt[,1]exp=rt[,2:ncol(rt)]dimnames=list(rownames(exp),colnames(exp))rt=matrix(as.numeric(as.matrix(exp)),nrow=nrow(exp),dimnames=dimnames)rt=avereps(rt)rt=normalizeBetweenArrays(as.matrix(rt))#rt=log2(rt+1)#根据矩阵的值来确定，如果比较大则需要做log转换。#differential #分组信息Type=c(rep("DOCK5.H3N2.D1",3),rep("WT.H3N2.D1",3))design colnames(design) "DOCK5.H3N2.D1",fit cont.matrixfit2 fit2 allDiff=topTable(fit2,adjust='fdr',number=200000)write.table(allDiff,file="ko.xls",sep="\t",quote=F)type=sapply(strsplit(rownames(allDiff),"\\|"),"[",2)rownames(allDiff)=gsub("(.*?)\\|.*","\\1",rownames(allDiff))rownames(rt)=gsub("(.*?)\\|.*","\\1",rownames(rt))#差异表达diffSig logFoldChange & adj.P.Val < adjustP )), ]write.table(diffSig,file="diff_ko.xls",sep="\t",quote=F)#上调差异表达diffup logFoldChange & adj.P.Val < adjustP )), ]write.table(diffup,file="diff_ko_up.xls",sep="\t",quote=F)#下调差异表达diffdown write.table(diffdown,file="diff_ko_down.xls",sep="\t",quote=F)

得到差异表达矩阵：

xMax=max(-log10(allDiff$adj.P.Val))yMax=max(abs(allDiff$logFC))plot(-log10(allDiff$adj.P.Val), allDiff$logFC, xlab="-log10(adj.P.Val)",ylab="logFC",     main="Volcano", xlim=c(0,xMax),ylim=c(-yMax,yMax),yaxs="i",pch=20, cex=0.8)diffSub=subset(allDiff, adj.P.VallogFoldChange)points(-log10(diffSub$adj.P.Val), diffSub$logFC, pch=20, col="red",cex=0.8)diffSub=subset(allDiff, adj.P.Valpoints(-log10(diffSub$adj.P.Val), diffSub$logFC, pch=20, col="green",cex=0.8)abline(h=0,lty=2,lwd=3)

#差异基因的表达水平，这是经过校正的hmExp=rt[rownames(diffSig),]diffExp=rbind(id=colnames(hmExp),hmExp)write.table(diffExp,file="heatmap.txt",sep="\t",quote=F,col.names=F)library(pheatmap)rt=read.table("heatmap.txt",sep="\t",header=T,row.names=1,check.names=F)Type=c(rep("DOCK5.H3N2.D1",3),rep("WT.H3N2.D1",3))     #修改对照和处理组样品数目fac=levels(factor(Type))rt1=rt[,Type==fac[1]]rt2=rt[,Type==fac[2]]rt=cbind(rt1,rt2)Type=c(rep(as.character(fac[1]),ncol(rt1)),rep(as.character(fac[2]),ncol(rt2)))names(Type)=colnames(rt)Type=as.data.frame(Type)tiff(file="heatmap.tiff",     width = 30,            #图片的宽度     height =30,            #图片的高度     units ="cm",     compression="lzw",     bg="white",     res=600)pheatmap(rt,          annotation=Type,          color = colorRampPalette(c("green", "black", "red"))(50),         cluster_cols =F,         fontsize = 8,         fontsize_row=4,         fontsize_col=8)dev.off()

然后就可以搞其他的了：

将差异基因复制到deg.txt文件中。

富集分析：

if (!requireNamespace("BiocManager", quietly = TRUE))  install.packages("BiocManager")BiocManager::install("topGo")BiocManager::install("clusterProfiler")BiocManager::install("org.Hs.eg.db")library(clusterProfiler)library(org.Hs.eg.db)library(topGO)getwd()a b eg = bitr(b, fromType="SYMBOL", toType="ENTREZID", OrgDb ="org.Hs.eg.db")data(geneList, package = "DOSE")gene gene ego   gene          = gene,  keyType = "ENTREZID",  OrgDb         = org.Hs.eg.db,  ont           = "MF",  pAdjustMethod = "BH",  pvalueCutoff  = 0.05,  qvalueCutoff  = 0.05,  readable      = TRUE)barplot(ego)ego   gene          = gene,  keyType = "ENTREZID",  OrgDb         = org.Hs.eg.db,  ont           = "BP",  pAdjustMethod = "BH",  pvalueCutoff  = 0.05,  qvalueCutoff  = 0.05,  readable      = TRUE)barplot(ego)ego   gene          = gene,  keyType = "ENTREZID",  OrgDb         = org.Hs.eg.db,  ont           = "cc",  pAdjustMethod = "BH",  pvalueCutoff  = 0.05,  qvalueCutoff  = 0.05,  readable      = TRUE)barplot(ego)dotplot(ego)ekegg   gene          = gene,  keyType     = "kegg",  organism   = 'hsa',  pvalueCutoff      = 0.05,  pAdjustMethod     = "BH",  qvalueCutoff  = 0.05)barplot(ekegg)dotplot(ekegg)

网络分析：

STRING: functional protein association networks

https://string-db.org/

得到网络图：

进行一下调节：

也有富集分析的结果：

下载tsv文件：

打开cytoscape，导入文件：

得到中心节点：

Top 10 in network string_interactions (13).tsv ranked by MCC  method
Rank	Name	Score
1	C3	6.23E+09
2	FGG	6.23E+09
2	FGA	6.23E+09
4	IGFBP3	6.23E+09
4	IGFBP1	6.23E+09
6	APOE	6.23E+09
7	MSLN	6.23E+09
8	BMP4	6.23E+09
8	VCAN	6.23E+09
10	SPP1	6.23E+09

参考资料：

不用编程的一篇SCI是怎么弄出来的？

https://mp.weixin.qq.com/s/0pmZ1EFHZDvVsZ4rdSv2Iw

从富集分析到统计学假设检验

https://mp.weixin.qq.com/s/3rermga6zgT7RWZyaAPn7A