#已有UCSCXena数据，进行后续分析
rm(list=ls())
options(stringsAsFactors = F)
##可以直接读没有解压的数据，mRNA-SEQ数据
mRNA_HiSeqV2=read.table(“HiSeqV2”,header = T,sep = ‘t‘)
dim(mRNA_HiSeqV2)
mRNA_HiSeqV2[1:4,1:4]
#查看NA的数据
na.omit(mRNA_HiSeqV2)[1:4,1:4]
dim(na.omit(mRNA_HiSeqV2))
save(mRNA_HiSeqV2,file=“mRNA_HiSeqV2.Rdata”)
##临床信息
mRNA_clinical=read.table(“CESC_clinicalMatrix” ,header = T,sep = ‘t‘, quote = “”)
dim(mRNA_clinical)
mRNA_clinical[1:4,1:4]
save(mRNA_clinical,file=“mRNA_clinical.Rdata”)
##生存相关信息
mRNA_survival=read.table(“CESC_survival.txt.gz” ,header = T,sep = ‘t‘)
dim(mRNA_survival)
mRNA_survival[1:4,1:4]
save(mRNA_survival,file=“mRNA_survival.Rdata”)
rm(list=ls())
options(stringsAsFactors = F)
load(“mRNA_HiSeqV2.Rdata”)
###########
# 这里需要解析TCGA数据库的ID规律，来判断样本归类问题。https://mp.weixin.qq.com/s/Ph1O6V5RkxkyrKpVmB5ODA
#01–09是癌症，10–19是正常，20–29是癌旁
expr <- mRNA_HiSeqV2
expr[1:4,1:4]
rownames(expr)=expr[,1]
expr=expr[,-1]
expr[1:4,1:4]
#通过数字来判断样本类型
ls <- unlist(substr(colnames(expr),14,15))
table(ls)
group_list=ifelse(as.numeric(substr(colnames(expr),14,15)) < 10,‘tumor’,‘normal’)
table(group_list)
exprSet=na.omit(expr)
dim(exprSet)
exprSet[1:4,1:4]
exprSet=exprSet[,group_list==“tumor”]
dim(exprSet)
##xena下载的数据是经过log的count值，所以要还原才可以进行差异分析
exprSet <- 2^exprSet-1
dim(exprSet)
exprSet[1:4,1:4]
#并且还要取整数
exprSet <- floor(exprSet)
exprSet[1:4,1:4]
save(exprSet,file=“exprSet_DEG.Rdata”)
#四、分组
#根据感兴趣的基因的表达量进行分组，之后进行差异分析。
load(“exprSet_DEG.Rdata”)
####根据感兴趣的基因高低表达做差异分析,这里用了中位数，也可以用平均值
group_list=ifelse(as.numeric(exprSet[“MYC”,])>median(as.numeric(exprSet[“MYC”,])),‘high’,‘low’)
table(group_list)
class(group_list)
group_list <- factor(group_list)
class(group_list)
#五、差异分析
#转录组数据最常用DESeq2包进行差异分析
library(DESeq2)
colData <- data.frame(row.names=colnames(exprSet),
group_list=group_list)
colData
dds <- DESeqDataSetFromMatrix(countData = exprSet,
colData = colData,
design = ~ group_list)
dds <- DESeq(dds)
res <- results(dds,
contrast=c(“group_list”,“high”,“low”))
head(res)
##sort by padj 进一步优化，安装标准化的P值排序
resOrdered <- res[order(res$padj),]
head(resOrdered)
DEG =as.data.frame(resOrdered)
DESeq2_DEG = na.omit(DEG)
head(DESeq2_DEG)
nrDEG=DESeq2_DEG#[,c(2,6)]
#colnames(nrDEG)=c(‘log2FoldChange’,‘pvalue’)
head(nrDEG)
#六、筛选差异基因
#对得到的差异基因，按照fold change和P值，制定自己的标准，进行筛选。
##筛选
#get diff_gene 得到差异基因,选P值及logFC大于1的
diff_gene <- subset(nrDEG,pvalue<0.01 & (log2FoldChange>=1 | log2FoldChange<=-1))
head(diff_gene)
#给数据标记上调下调
resdata <- diff_gene
resdata$significant <- c(rep(“unchange”,nrow(resdata)))
resdata$significant[resdata$pvalue<=0.01&resdata$log2FoldChange>=1] <- “up”
resdata$significant[resdata$pvalue<=0.01&resdata$log2FoldChange<=-1] <- “down”
#查看最新数据的前五行
head(resdata)
resdata[1:40,]
diff_gene <- resdata
table(diff_gene$significant)
#输出文件为TXT格式
write.table(x=diff_gene,
file=“diff_result_CESC_MYC.txt”,
quote=F,
sep = “t“,
row.names = T,
col.names = T)
save(nrDEG,DESeq2_DEG,diff_gene,file=“DESeq2_DEG_result.Rdata”)
load(“DESeq2_DEG_result.Rdata”)
#七、热图
rm(list=ls())
options(stringsAsFactors = F)
load(“DESeq2_DEG_result.Rdata”)
load(“exprSet_DEG.Rdata”)
####根据感兴趣的基因高低表达做差异分析,这里用了中位数，也可以用平均值
group_list=ifelse(as.numeric(exprSet[“MYC”,])>median(as.numeric(exprSet[“MYC”,])),‘high’,‘low’)
table(group_list)
class(group_list)
group_list <- factor(group_list)
class(group_list)
######3提取差异最大的50个基因画热图,用P值判断
## heatmap
library(pheatmap)
choose_gene=head(rownames(diff_gene),50) ## 50 maybe better
choose_matrix=exprSet[choose_gene,]
#把太小的值去掉
boxplot(choose_matrix)
n=t(scale(t(choose_matrix)))
boxplot(n)
n[n>2]=2
n[n< –2]=-2
n[1:4,1:4]
boxplot(n)
#分组信息
annotation_col=data.frame(group=group_list)
row.names(annotation_col) <- colnames(exprSet)
annotation_col
#pheatmap(choose_matrix,filename = paste0(n,‘_need_DEG_top50_heatmap.png’))
pheatmap(n,annotation_col=annotation_col,show_rownames=F,show_colnames=F)
######3提取差异最大的50个基因画热图,用logFC判断
## heatmap
library(pheatmap)
diff_gene<- diff_gene[order(abs(diff_gene$log2FoldChange),decreasing = T),]
head(diff_gene)
exprSet <- exprSet[,order(group_list)]
exprSet[1:4,1:4]
choose_gene=head(rownames(diff_gene),50) ## 50 maybe better
choose_matrix=exprSet[choose_gene,]
#把太小的值去掉
boxplot(choose_matrix)
n=t(scale(t(choose_matrix)))
boxplot(n)
n[n>2]=2
n[n< –2]=-2
n[1:4,1:4]
boxplot(n)
#分组信息
annotation_col=data.frame(group=group_list)
row.names(annotation_col) <- colnames(exprSet)
annotation_col
#pheatmap(choose_matrix,filename = paste0(n,‘_need_DEG_top50_heatmap.png’))
pheatmap(n,annotation_col=annotation_col,
show_rownames=F,show_colnames=F)
#八、火山图
#做火山图数据处理
######火山图
rm(list = ls()) ## 魔幻操作，一键清空~
options(stringsAsFactors = F)
#导入数据,火山图需要的数据前期处理
load(“DESeq2_DEG_result.Rdata”)
head(DESeq2_DEG)
library(ggpubr)
df<-DESeq2_DEG
df<-DESeq2_DEG[-1,] ##MYC的P值太小，为了作图好看
attach(df)
##火山图横坐标是logFC，纵坐标是-log10(P.Value)
plot(log2FoldChange,-log10(pvalue))
df$v=-log10(pvalue) #df新增加一列‘v’,值为-log10(P.Value)
ggscatter(df, x = “log2FoldChange”, y = “v”,size=0.8)
##加上上调下调信息
df$group=ifelse(df$pvalue>0.01,‘stable’, #if 判断：如果这一基因的P.Value>0.01，则为stable基因
ifelse( df$log2FoldChange >=1,‘up’, #接上句else 否则：接下来开始判断那些P.Value<0.01的基因，再if 判断：如果logFC >1.5,则为up（上调）基因
ifelse( df$log2FoldChange <= –1,‘down’,‘stable’) )#接上句else 否则：接下来开始判断那些logFC <1.5 的基因，再if 判断：如果logFC <1.5，则为down（下调）基因，否则为stable基因
)
table(df$g)
df$name=rownames(df)
head(df)
用gpubr画火山图
##方法一：ggpubr画火山图
ggscatter(df, x = “log2FoldChange”, y = “v”,size=0.8,color = ‘group’)
##挑选一些感兴趣的基因标记出来
ggscatter(df, x = “log2FoldChange”, y = “v”, color = “group”,size = 1,
label = “name”, repel = T,
#label.select = rownames(df)[df$g != ‘stable’] ,
label.select = head(rownames(df)), #挑选一些基因在图中显示出来
palette = c(“#00AFBB”, “#999999”, “#FC4E07”) )
ggsave(‘volcano.png’)
###方法二：ggplot画火山图
library(ggplot2)
p <- ggplot(data = df,
aes(x = log2FoldChange,
y = v)) +
geom_point(alpha=0.4, size=3.5,
aes(color=group)) +
scale_color_manual(values=c(“blue”, “grey”,“red”))+
geom_vline(xintercept=c(-1,1),lty=4,col=“black”,lwd=0.8) +
geom_hline(yintercept = –log10(0.01),lty=4,col=“black”,lwd=0.8) +
theme_bw()
p
#如果想要标记基因名字，需要生成用于添加图层的新数据
#新数据框的内容是你想要标记的基因，这里根据logFC和Pvalue的大小来筛选，可以自定义阈值来调整要显示的基因的数量：
library(dplyr)
head(df)
for_label <- df %>%
filter(abs(logFC) >4 & v>-log10(0.01))
for_label <- data.frame(head(df)) #挑选一些基因在图中显示出来
class(for_label)
head(for_label)
###新图层叠加到原图上去
p +
geom_point(size = 3, shape = 1, data = for_label) +
ggrepel::geom_label_repel(
aes(label = name),
data = for_label,
color=“black”
)
ggsave(“ggplot2_volcano.png”)
九、GO_KEGG分析前数据处理
rm(list = ls()) ## 魔幻操作，一键清空~
options(stringsAsFactors = F)
#导入数据
load(“DESeq2_DEG_result.Rdata”)
head(DESeq2_DEG)
df <- DESeq2_DEG
##加上上调下调信息
df$g=ifelse(df$pvalue>0.01,‘stable’, #if 判断：如果这一基因的P.Value>0.01，则为stable基因
ifelse( df$log2FoldChange >=1,‘up’, #接上句else 否则：接下来开始判断那些P.Value<0.01的基因，再if 判断：如果logFC >1.5,则为up（上调）基因
ifelse( df$log2FoldChange <= –1,‘down’,‘stable’) )#接上句else 否则：接下来开始判断那些logFC <1.5 的基因，再if 判断：如果logFC <1.5，则为down（下调）基因，否则为stable基因
)
table(df$g)
#增加一列基因名字
df$symbol=rownames(df)
head(df)
#载入GO、KEGG、GSEA分析需要的包
library(ggplot2)
library(clusterProfiler)
library(org.Hs.eg.db)
#将基因名转化成ENTREZID才能进行GO分析
deg <- bitr(unique(df$symbol), fromType = “SYMBOL”,
toType = c( “ENTREZID”),
OrgDb = org.Hs.eg.db)
head(deg)
DEG=df
head(DEG)
#通过合并完成SYMBOL和ENTREZID的转换
DEG=merge(DEG,deg,by.y=‘SYMBOL’,by.x=‘symbol’)
head(DEG)
##将差异基因的上调下调分成不同的数据框，分别进行、KEGG、GSEA分析
gene_up= DEG[DEG$g == ‘up’,‘ENTREZID’]
gene_down=DEG[DEG$g == ‘down’,‘ENTREZID’]
gene_diff=c(gene_up,gene_down)
gene_all=as.character(DEG[ ,‘ENTREZID’] )
#data(geneList, package=”DOSE”)
#head(geneList)
#boxplot(geneList)
boxplot(DEG$log2FoldChange)
geneList=DEG$log2FoldChange
head(geneList)
names(geneList)=DEG$ENTREZID
geneList=sort(geneList,decreasing = T)
head(geneList)
save(DEG,gene_up, gene_down,gene_diff, gene_all, file = ‘anno_DEG_list.Rdata’)
GO
分析
方法一，得到3种分析‘BP’,’MF’,’CC’合并的图
rm(list = ls()) ## 魔幻操作，一键清空~
options(stringsAsFactors = F)
###完成前期的数据处理，直接load数据就可以开始进行分析
load(‘anno_DEG_list.Rdata’)
### GO database analysis
### 做GO数据集超几何分布检验分析，重点在结果的可视化及生物学意义的理解。
###方法一，得到3种分析‘BP’,’MF’,’CC’合并的图
ego <- enrichGO(gene = gene_up, ##手动换上调下调的基因
#universe = gene_all,
OrgDb = org.Hs.eg.db,
ont = “all”, #BP/CC/MF或者“ALL”，如果是“ALL”的时候要加“pool=TRUE”
pAdjustMethod = “BH”, #选哪个都差距不大
#pvalueCutoff = 0.1,
#qvalueCutoff = 0.1,
pool = TRUE,
readable = TRUE
)
head(ego[,1:8])
go_data <- data.frame(ego)
head(go_data)
write.table(x=go_data,
file=“go_data_CESC_MYC.txt”,
quote=F,
sep = “t“,
row.names = T,
col.names = T)
#barplot(ego)
#barplot(ego, showCategory = 10)
#dotplot(ego, showCategory=10)
library(ggplot2)
setting=element_text(size = 12,color=“black”,family=“Arial”)
barplot(ego, split=“ONTOLOGY”,showCategory = 5)+ facet_grid(ONTOLOGY~., scale=“free”)
ggsave(“go_barplot_5.png”)
dotplot(ego, split=“ONTOLOGY”,showCategory = 10)+ facet_grid(ONTOLOGY~., scale=“free”)
ggsave(“go_dotplot_10.png”)
#dotplot(ego, split=”ONTOLOGY”)+ facet_grid(ONTOLOGY~., scale=”free”)
#ggsave(“go_dotplot.png”)
###方法二：得到分开的图
##厉害的循环将3个列表分别进行了3种分析‘BP’,’MF’,’CC’，得到九个结果
g_list=list(gene_up=gene_up,
gene_down=gene_down,
gene_diff=gene_diff)
go_enrich_results <- lapply( g_list , function(gene) {
lapply( c(‘BP’,‘MF’,‘CC’) , function(ont) {
cat(paste(‘Now process ‘,ont ))
##最关键的函数
ego <- enrichGO(gene = gene,
universe = gene_all,
OrgDb = org.Hs.eg.db,
ont = ont , #BP/CC/MF或者“ALL”，如果是“ALL”的时候要加“pool=TRUE”
pAdjustMethod = “BH”,
pvalueCutoff = 0.99,
qvalueCutoff = 0.99,
readable = TRUE)
print( head(ego) )
return(ego)
})
})
save(go_enrich_results,file = ‘go_enrich_results.Rdata’)
load(file = ‘go_enrich_results.Rdata’)
##循环批量成图，并保存
n1= c(‘gene_up’,‘gene_down’,‘gene_diff’)
n2= c(‘BP’,‘MF’,‘CC’)
for (i in 1:3){
for (j in 1:3){
fn=paste0(‘dotplot_’,n1[i],‘_’,n2[j],‘.png’)
cat(paste0(fn,‘n‘))
png(fn,res=150,width = 1080)
print( dotplot(go_enrich_results[[i]][[j]] ))
dev.off()
}
}
###方法三，得到3种分析‘BP’,’MF’,’CC’单独的图
##BP
ego <- enrichGO(gene = gene_up, ##手动换上调下调的基因
#universe = gene_all,
OrgDb = org.Hs.eg.db,
ont = “BP”, #BP/CC/MF或者“ALL”，如果是“ALL”的时候要加“pool=TRUE”
pAdjustMethod = “BH”, #选哪个都差距不大
#pvalueCutoff = 0.1,
#qvalueCutoff = 0.1,
pool = TRUE,
readable = TRUE
)
head(ego[,1:8])
p1_BP <- dotplot(ego)
p1_BP
##CC
ego <- enrichGO(gene = gene_up, ##手动换上调下调的基因
#universe = gene_all,
OrgDb = org.Hs.eg.db,
ont = “CC”, #BP/CC/MF或者“ALL”，如果是“ALL”的时候要加“pool=TRUE”
pAdjustMethod = “BH”, #选哪个都差距不大
#pvalueCutoff = 0.1,
#qvalueCutoff = 0.1,
pool = TRUE,
readable = TRUE
)
head(ego[,1:8])
p1_CC <- dotplot(ego,showCategory = 5)
p1_CC
##MF
ego <- enrichGO(gene = gene_up, ##手动换上调下调的基因
#universe = gene_all,
OrgDb = org.Hs.eg.db,
ont = “MF”, #BP/CC/MF或者“ALL”，如果是“ALL”的时候要加“pool=TRUE”
pAdjustMethod = “BH”, #选哪个都差距不大
#pvalueCutoff = 0.1,
#qvalueCutoff = 0.1,
pool = TRUE,
readable = TRUE
)
head(ego[,1:8])
p1_MF <- dotplot(ego,showCategory = 5)
p1_MF
KEGG
分析
rm(list = ls()) ## 魔幻操作，一键清空~
options(stringsAsFactors = F)
#载入GO、KEGG、GSEA分析需要的包
library(ggplot2)
library(clusterProfiler)
library(org.Hs.eg.db)
###完成前期的数据处理，直接load数据就可以开始进行分析
load(‘anno_DEG_list.Rdata’)
## KEGG pathway analysis
### 做KEGG数据集超几何分布检验分析，重点在结果的可视化及生物学意义的理解。
### over-representation test
kk.up <- enrichKEGG(gene = gene_up,
organism = ‘hsa’,
universe = gene_all,
pvalueCutoff = 0.9,
qvalueCutoff =0.9)
head(kk.up)[,1:6]
class(kk.up)
dotplot(kk.up )
ggsave(‘kk.up.dotplot_1.png’)
#dev.off()
kk.down <- enrichKEGG(gene = gene_down,
organism = ‘hsa’,
universe = gene_all,
pvalueCutoff = 0.9,
qvalueCutoff =0.9)
head(kk.down)[,1:6]
dotplot(kk.down );ggsave(‘kk.down.dotplot.png’)
kk.diff <- enrichKEGG(gene = gene_diff,
organism = ‘hsa’,
pvalueCutoff = 0.05)
head(kk.diff)[,1:6]
dotplot(kk.diff )
ggsave(‘kk.diff.dotplot.png’)
kegg_diff_dt <- as.data.frame(kk.diff)
kegg_down_dt <- as.data.frame(kk.down)
kegg_up_dt <- as.data.frame(kk.up)
down_kegg<-kegg_down_dt[kegg_down_dt$pvalue<0.05,];down_kegg$group=-1
up_kegg<-kegg_up_dt[kegg_up_dt$pvalue<0.05,];up_kegg$group=1
##自定义了一个函数source(‘functions.R’)
kegg_plot <- function(up_kegg,down_kegg){
dat=rbind(up_kegg,down_kegg)
colnames(dat)
dat$pvalue = –log10(dat$pvalue)
dat$pvalue=dat$pvalue*dat$group
dat=dat[order(dat$pvalue,decreasing = F),]
g_kegg<- ggplot(dat, aes(x=reorder(Description,order(pvalue, decreasing = F)), y=pvalue, fill=group)) +
geom_bar(stat=“identity”) +
scale_fill_gradient(low=“blue”,high=“red”,guide = FALSE) +
scale_x_discrete(name =“Pathway names”) +
scale_y_continuous(name =“log10P-value”) +
coord_flip() + theme_bw()+theme(plot.title = element_text(hjust = 0.5))+
ggtitle(“Pathway Enrichment”)
}
g_kegg=kegg_plot(up_kegg,down_kegg)
print(g_kegg)
ggsave(g_kegg,filename = ‘kegg_up_down.png’)
十、GSEA分析
rm(list = ls()) ## 魔幻操作，一键清空~
options(stringsAsFactors = F)
#载入GO、KEGG、GSEA分析需要的包
library(ggplot2)
library(clusterProfiler)
library(org.Hs.eg.db)
library(enrichplot)
#BiocManager::install(“ReactomePA”)
library(ReactomePA)
library(data.table)
###完成前期的数据处理，直接load数据就可以开始进行分析
load(‘anno_DEG_list.Rdata’)
geneList=DEG$log2FoldChange
names(geneList)=DEG$ENTREZID
geneList=sort(geneList,decreasing = T)
head(geneList)
### GSEA
kk_gse <- gseKEGG(geneList = geneList,
organism = ‘hsa’,
nPerm = 1000,
minGSSize = 120,
pvalueCutoff = 0.9,
verbose = FALSE)
dim(kk_gse)
GSEA_data <- data.frame(kk_gse)
head(GSEA_data)
write.table(x=GSEA_data,
file=“GSEA_data_CESC_MYC.txt”,
quote=F,
sep = “t“,
row.names = T,
col.names = T)
head(kk_gse)[,1:6]
###单独一个条形码
gseaplot(kk_gse, geneSetID = rownames(kk_gse[1,]))
ggsave(‘gseaplot.png’)
library(ReactomePA)
#这三种分析的结果都可以输出文本查看。有时候会出现这种情况：没有富集到任何结果。研究后发现，pvalueCutoff界定值应该背锅，虽然这里写的是P值，但这个包里实际是以FDR作为界定。经常用FDR的同学会遇到这种情况：P值很多很显著的，但是FDR算出来都是一个值，时而还都是大于0.05的相同值。所以解决办法就是pvalueCutoff设置等于1，要是不设定也不行，只要设定成1，就会输出所有结果。
###GSEA分析——GO
Go_gseresult <- gseGO(geneList, ‘org.Hs.eg.db’, keyType = “ENTREZID”, ont=“BP”,
nPerm = 1000, minGSSize = 10, maxGSSize = 1000, pvalueCutoff=0.05)
save(Go_gseresult,file = “Go_gseresult.Rdata”)
load(“Go_gseresult.Rdata”)
###GSEA分析——KEGG
KEGG_gseresult <- gseKEGG(geneList, nPerm = 1000, minGSSize = 10,
maxGSSize = 1000, pvalueCutoff=1)
save(KEGG_gseresult,file = “KEGG_gseresult.Rdata”)
load(“KEGG_gseresult.Rdata”)
##GSEA分析——Reactome
Go_Reactomeresult <- gsePathway(geneList, nPerm = 1000, minGSSize = 10,
maxGSSize = 1000, pvalueCutoff=1)
save(Go_Reactomeresult,file = “Go_Reactomeresult.Rdata”)
load(“Go_Reactomeresult.Rdata”)
##波浪图
ridgeplot(KEGG_gseresult, 5) #输出前十个结果
ggsave(‘ridgeplot_1.png’)
###还是gseaplot2看着舒服
gseaplot2(KEGG_gseresult, 1)
ggsave(‘gseaplot2.png’)
###gseaplot2还可以同时显示复数个功能组的富集曲线，并标记P值
gseaplot2(KEGG_gseresult, 1:4, pvalue_table = TRUE)
ggsave(‘KEGG_gseresult_1.png’)
