硬着头皮往下走PCA|GSEA / 四六文摘

一直在生信菜鸟团发布的学徒数据挖掘第二期目录如下：

因为阅读量爆棚，所以提升到生信技能树发布

本期数据挖掘任务来自于Paper：Tumor Evolution and Drug Response in Patient-Derived Organoid Models of Bladder Cancer

要重复的图表是：Figures:

PCA

ssGSEA

其实，是很简单的处理：

差异分析后提取top1000进行PCA，要表达的意思是，tumor和TCGA的tumor能聚在一起；

对差异基因进行GSEA的KEGG富集；

有了其实，必有但是；作者上传的矩阵是辣个样子的，

a.GEOquery拿不到矩阵；

b.网页下载的矩阵是DESeq2-normalized counts；

c.英文不翻之DESeq2 doesn’t actually use normalized counts, rather it uses the raw counts and models the normalization inside the Generalized Linear Model (GLM). These normalized counts will be useful for downstream visualization of results, but cannot be used as input to DESeq2 or any other tools that peform differential expression analysis which use the negative binomial model.

纯代码：

Step1-download

###一些常规的设置 rm(list = ls())#清空环境变量 options(stringsAsFactors = F)##字符不作为因子读入 #####数据下载 library(GEOquery) h<-'GSE103990.Rdata' ####getGPL获得平台的注释信息，但下载速度会慢很多 ####而且注释文件格式大多不如bioconductor包好用 if(!file.exists(h)){ gset<-getGEO('GSE103990',destdir='.', AnnotGPL=F, getGPL=F) save(gset,file=h) } load('GSE103990.Rdata') ex<- exprs(gset[[1]]) pd <- pData(gset[[1]]) #system('nohup wget ftp://ftp.ncbi.nlm.nih.gov/geo/series/GSE103nnn/GSE103990/suppl/GSE103990_Normalized_counts.txt.gz &') counts_nor <- read.table('GSE103990_Normalized_counts.txt.gz') save(counts_nor,pd,file='Nor.Rdata')

Step2-差异分析（数据不合适，但代码木有问题）

rm(list = ls()) options(stringsAsFactors = F) load('Nor.Rdata') rownames(counts_nor)<- substr(rownames(counts_nor),1,15) exprSet<- floor(counts_nor) pd <- pd[match(colnames(exprSet),pd$description.1),] group_list <- ifelse(grepl('org',pd$title),'org','tumor') suppressMessages(library(DESeq2)) #### 第一步，构建DESeq2的DESeq对象 colData <- data.frame(row.names=colnames(exprSet),group_list=group_list) dds <- DESeqDataSetFromMatrix(countData = exprSet,colData = colData, design = ~ group_list) #### 第二步，进行差异表达分析 dds2 <- DESeq(dds) res <- results(dds2,contrast=c("group_list","org","tumor")) resOrdered <- res[order(res$padj),] DEG <- as.data.frame(resOrdered) DESeq2_DEG = na.omit(DEG) nrDEG=DESeq2_DEG[,c(2,6)] colnames(nrDEG)=c('log2FoldChange','pvalue') save(nrDEG, DESeq2_DEG, file = "DEG.Rdata")

Step3-PCA

rm(list=ls()) library(edgeR) load('Nor.Rdata') load('BLC.Rdata') load('DEG.Rdata') rownames(counts_nor)<- substr(rownames(counts_nor),1,15) nor_BLCA <- edgeR::cpm(expr_BLC,log=T) nor_paper <- edgeR::cpm(floor(counts_nor),log=T) inter_gene<- intersect(rownames(nor_BLCA),rownames(nor_paper)) choose_gene <- rownames(nrDEG)[abs(nrDEG$log2FoldChange)>1.5&nrDEG$pvalue<0.01] nrDEG <- nrDEG[order(abs(nrDEG$log2FoldChange),nrDEG$pvalue,decreasing = T),] final_choose<- rownames(nrDEG)[rownames(nrDEG)%in%inter_gene][1:1000] nr_paper <- nor_paper[final_choose,] nr_BLC <- nor_BLCA[final_choose,] nr_pca <- cbind(nr_paper,nr_BLC) group_list <- ifelse(grepl('TCGA',colnames(nr_pca)),'TCGA',ifelse(grepl('org',pd$title),'org','tumor')) library("FactoMineR") library("factoextra") #### df.pca <- PCA(t(nr_pca), graph = FALSE) fviz_pca_ind(df.pca, geom.ind = "point", col.ind = group_list, addEllipses = F, legend.title = "Groups")

Step4-GSEA

如果对GSEA中的细节感兴趣，看https://www.jianshu.com/p/be8fe1318850

rm(list=ls()) load('DEG.Rdata') library(org.Hs.eg.db) library(clusterProfiler) gene <- bitr(rownames(nrDEG), fromType = "ENSEMBL", toType = "ENTREZID", OrgDb = org.Hs.eg.db) gene$logfc <- nrDEG$log2FoldChange[match(gene$ENSEMBL,rownames(nrDEG))] geneList=nrDEG$log2FoldChange names(geneList)=gene$ENTREZID geneList=sort(geneList,decreasing = T) head(geneList) library(clusterProfiler) kk_gse <- gseKEGG(geneList = geneList, organism = 'hsa', nPerm = 1000, minGSSize = 10, pvalueCutoff = 0.9, verbose = FALSE) paper_choose <- c('Cell adhesion molecules (CAMs)', 'Cell cycle', 'ErbB signaling pathway') a <- list() for (i in 1:3){ a[[i]]<- gseaplot2(kk_gse, geneSetID = kk_gse@result$ID[kk_gse@result$Description==paper_choose[i]], pvalue_table = T) } a