比较不同流程（limma/voom，edgeR，DESeq2 ）差异分析的区别

# R包太多，这里略了。
for (pkg in cran_packages){
  if (! require(pkg,character.only=T) ) {
    install.packages(pkg,ask = F,update = F)
    require(pkg,character.only=T) 
  }
}

# first prepare BioManager on CRAN
if(!require("BiocManager")) install.packages("BiocManager",update = F,ask = F)

# use BiocManager to install
for (pkg in Biocductor_packages){
  if (! require(pkg,character.only=T) ) {
    BiocManager::install(pkg,ask = F,update = F)
    require(pkg,character.only=T) 
  }
}

#前面的报错都先不要管。主要看这里
for (pkg in c(Biocductor_packages,cran_packages)){
  require(pkg,character.only=T) 
}

#哪个报错，就回去安装哪个。如果你没有安装xx包，却提示你xx包不存在，这也正常，是因为复杂的依赖关系，缺啥补啥。

if(!require(tinyarray))devtools::install_local("tinyarray-master.zip",upgrade = F)

#step1
options(stringsAsFactors = F)
library(stringr)
cancer_type="TCGA-BRCA"
if(!dir.exists("clinical"))dir.create("clinical")
if(!dir.exists("expdata"))dir.create("expdata")
dir()

#此代码块是在Terminal中使用
conda create -n gdc python=3.7
source activate gdc
git clone https://github.com/NCI-GDC/gdc-client
cd gdc-client
pip install -r requirements.txt
python setup.py install 2>&1 | tee -a install.log
###安装完毕后开始下载数据
gdc-client download -m gdc_manifest.2020-12-17_clinic.txt -d clinical
gdc-client download -m gdc_manifest.2020-12-17_exp.txt -d expdata

#下载结束后
length(dir("./clinical/"))
length(dir("./expdata/"))

library(XML)
result <- xmlParse("./clinical/00049989-fa21-48fb-8dda-710c0dd5932e/nationwidechildrens.org_clinical.TCGA-A2-A0CT.xml")   #单个运行
rootnode <- xmlRoot(result)
rootsize <- xmlSize(rootnode)
print(rootnode[1])
#print(rootnode[2])
xmldataframe <- xmlToDataFrame(rootnode[2])
head(t(xmlToDataFrame(rootnode[2])))

xmls = dir("clinical/",pattern = "*.xml$",recursive = T)   #可层级

td = function(x){
  result <- xmlParse(file.path("clinical/",x))
  rootnode <- xmlRoot(result)
  xmldataframe <- xmlToDataFrame(rootnode[2])
  return(t(xmldataframe))
}

cl = lapply(xmls,td)
cl_df <- t(do.call(cbind,cl))
cl_df[1:3,1:3]
clinical = data.frame(cl_df)
clinical[1:4,1:4]

count_files = dir("expdata/",pattern = "*.htseq.counts.gz$",recursive = T)

ex = function(x){
  result <- read.table(file.path("expdata/",x),row.names = 1,sep = "\t")
  return(result)
}
#  head(ex("03aee74e-4e37-4a58-a720-c90e807d2f40/be5bc6a0-9720-47ac-953e-fa8d0c32cd82.htseq.counts.gz"))

exp = lapply(count_files,ex)
exp <- do.call(cbind,exp)
dim(exp)
exp[1:4,1:4]

meta <- jsonlite::fromJSON("metadata.cart.2020-12-17.json")
colnames(meta)
ids <- meta$associated_entities;class(ids)
ids[[1]]
class(ids[[1]][,1])
ID = sapply(ids,function(x){x[,2]})
file2id = data.frame(file_name = meta$file_name,
                     ID = ID)
head(file2id$file_name)
head(count_files)
count_files2 = stringr::str_split(count_files,"/",simplify = T)[,2]
count_files2[1] %in% file2id$file_name
file2id = file2id[match(count_files2,file2id$file_name),]
colnames(exp) = file2id$ID
exp[1:4,1:4]

dim(exp)
exp = exp[apply(exp, 1, function(x) sum(x > 1) > 9), ]
dim(exp)
exp[1:4,1:4]

# 过滤在至少在75%的样本中都有表达的基因
keep_exp <- rowSums(exp>0) >= floor(0.75*ncol(exp))
table(keep_exp)
exp <- exp[keep_exp,]
exp[1:4,1:4]
dim(exp)

table(str_sub(colnames(exp),14,15))
group_list = ifelse(as.numeric(str_sub(colnames(exp),14,15)) < 10,'tumor','normal')
group_list = factor(group_list,levels = c("normal","tumor"))
table(group_list)
save(exp,clinical,group_list,cancer_type,file = paste0(cancer_type,"gdc.Rdata"))

#三大R包差异分析
if(!require(stringr))install.packages('stringr')
if(!require(ggplotify))install.packages("ggplotify")
if(!require(patchwork))install.packages("patchwork")
if(!require(cowplot))install.packages("cowplot")
if(!require(DESeq2))BiocManager::install('DESeq2')
if(!require(edgeR))BiocManager::install('edgeR')
if(!require(limma))BiocManager::install('limma')

rm(list = ls())
load("TCGA-BRCAgdc.Rdata")
table(group_list)

#deseq2----
library(DESeq2)
colData <- data.frame(row.names =colnames(exp), 
                      condition=group_list)
if(!file.exists(paste0(cancer_type,"dd.Rdata"))){
  dds <- DESeqDataSetFromMatrix(
    countData = exp,
    colData = colData,
    design = ~ condition)
  dds <- DESeq(dds)
  save(dds,file = paste0(cancer_type,"dd.Rdata"))
}
load(paste0(cancer_type,"dd.Rdata"))

# 两两比较
res <- results(dds, contrast = c("condition",rev(levels(group_list))))
resOrdered <- res[order(res$pvalue),] # 按照P值排序
DEG <- as.data.frame(resOrdered)
head(DEG)

#添加change列标记基因上调下调
logFC_cutoff <- with(DEG,mean(abs(log2FoldChange)) + 2*sd(abs(log2FoldChange)) )
#logFC_cutoff <- 2
k1 = (DEG$pvalue < 0.05)&(DEG$log2FoldChange < -logFC_cutoff)
k2 = (DEG$pvalue < 0.05)&(DEG$log2FoldChange > logFC_cutoff)
DEG$change = ifelse(k1,"DOWN",ifelse(k2,"UP","NOT"))
table(DEG$change)
head(DEG)
DESeq2_DEG <- DEG

#edgeR----
library(edgeR)

dge <- DGEList(counts=exp,group=group_list)
dge$samples$lib.size <- colSums(dge$counts)
dge <- calcNormFactors(dge) 

design <- model.matrix(~0+group_list)
rownames(design)<-colnames(dge)
colnames(design)<-levels(group_list)

dge <- estimateGLMCommonDisp(dge,design)
dge <- estimateGLMTrendedDisp(dge, design)
dge <- estimateGLMTagwiseDisp(dge, design)

fit <- glmFit(dge, design)
fit2 <- glmLRT(fit, contrast=c(-1,1)) 

DEG=topTags(fit2, n=nrow(exp))
DEG=as.data.frame(DEG)
logFC_cutoff <- with(DEG,mean(abs(logFC)) + 2*sd(abs(logFC)) )

#logFC_cutoff <- 2
k1 = (DEG$PValue < 0.05)&(DEG$logFC < -logFC_cutoff)
k2 = (DEG$PValue < 0.05)&(DEG$logFC > logFC_cutoff)
DEG$change = ifelse(k1,"DOWN",ifelse(k2,"UP","NOT"))

head(DEG)
table(DEG$change)
edgeR_DEG <- DEG


###limma----
library(limma)

design <- model.matrix(~0+group_list)
colnames(design)=levels(group_list)
rownames(design)=colnames(exp)

dge <- DGEList(counts=exp)
dge <- calcNormFactors(dge)

v <- voom(dge,design, normalize="quantile")
fit <- lmFit(v, design)

constrasts = paste(rev(levels(group_list)),collapse = "-")
cont.matrix <- makeContrasts(contrasts=constrasts,levels = design) 
fit2=contrasts.fit(fit,cont.matrix)
fit2=eBayes(fit2)

DEG = topTable(fit2, coef=constrasts, n=Inf)
DEG = na.omit(DEG)

logFC_cutoff <- with(DEG,mean(abs(logFC)) + 2*sd(abs(logFC)) )

#logFC_cutoff <- 2
k1 = (DEG$P.Value < 0.05)&(DEG$logFC < -logFC_cutoff)
k2 = (DEG$P.Value < 0.05)&(DEG$logFC > logFC_cutoff)
DEG$change = ifelse(k1,"DOWN",ifelse(k2,"UP","NOT"))
table(DEG$change)
head(DEG)
limma_voom_DEG <- DEG

tj = data.frame(deseq2 = as.integer(table(DESeq2_DEG$change)),
                edgeR = as.integer(table(edgeR_DEG$change)),
                limma_voom = as.integer(table(limma_voom_DEG$change)),
                row.names = c("down","not","up")
);tj

save(DESeq2_DEG,edgeR_DEG,limma_voom_DEG,group_list,tj,file = paste0(cancer_type,"DEG.Rdata"))


#可视化
rm(list = ls())
load("TCGA-BRCAgdc.Rdata")
load("TCGA-BRCADEG.Rdata")
if(!require(tinyarray))devtools::install_local("tinyarray-master.zip",upgrade = F)
library(ggplot2)
library(tinyarray)
exp[1:4,1:4]
dat = log2(exp+1)
pca.plot = draw_pca(dat,group_list);pca.plot
save(pca.plot,file = paste0(cancer_type,"pcaplot.Rdata"))

cg1 = rownames(DESeq2_DEG)[DESeq2_DEG$change !="NOT"]
cg2 = rownames(edgeR_DEG)[edgeR_DEG$change !="NOT"]
cg3 = rownames(limma_voom_DEG)[limma_voom_DEG$change !="NOT"]

h1 = draw_heatmap(dat[cg1,],group_list,scale_before = T)
h2 = draw_heatmap(dat[cg2,],group_list,scale_before = T)
h3 = draw_heatmap(dat[cg3,],group_list,scale_before = T)

m2d = function(x){
  mean(abs(x))+2*sd(abs(x))
}

v1 = draw_volcano(DESeq2_DEG,pkg = 1,logFC_cutoff = m2d(DESeq2_DEG$log2FoldChange))
v2 = draw_volcano(edgeR_DEG,pkg = 2,logFC_cutoff = m2d(edgeR_DEG$logFC))
v3 = draw_volcano(limma_voom_DEG,pkg = 3,logFC_cutoff = m2d(limma_voom_DEG$logFC))


library(patchwork)
(h1 + h2 + h3) / (v1 + v2 + v3) +plot_layout(guides = 'collect') &theme(legend.position = "none")
ggsave(paste0(cancer_type,"heat_vo.png"),width = 15,height = 10)

#三大R包差异基因对比之相关性
#for test correlation plot between the logFC for the top 500 DE genes
head(edgeR_DEG)
head(limma_voom_DEG)
head(DESeq2_DEG)

td1=edgeR_DEG[,1:2]
head(td1)
colnames(td1)=c("logFC_edgeR","logCPM_edgeR")
td1$ID=rownames(td1)


td2=limma_voom_DEG[,1:2]
head(td2)
colnames(td2)=c("logFC_limma","logCPM_limma")
td2$ID=rownames(td2)

td3=DESeq2_DEG[,1:2]
head(td3)
colnames(td3)=c("baseMean","log2FC_DESeq2")
td3$ID=rownames(td3)

td_edgeR_limma=inner_join(td1,td2,by="ID")
td_edgeR_limma=td_edgeR_limma[,c(3,1,4)]
head(td_edgeR_limma)

td_edgeR_DESeq2=inner_join(td1,td3,by="ID")
td_edgeR_DESeq2=td_edgeR_DESeq2[,c(3,1,5)]
head(td_edgeR_DESeq2)

td_DESeq2_limma=inner_join(td3,td2,by="ID")
td_DESeq2_limma=td_DESeq2_limma[,c(3,2,4)]
head(td_DESeq2_limma)

  #选出500个logFC绝对值最大的
# for edgeR vs limma
a1=head(td_edgeR_limma[order(td_edgeR_limma$logFC_edgeR),],250)
a2=tail(td_edgeR_limma[order(td_edgeR_limma$logFC_edgeR),],250)
a=rbind(a1,a2)
head(a)
ggplot(a,aes(y=logFC_limma,x=logFC_edgeR))+geom_point()
p1=ggplot(data=a, aes(y=logFC_limma, x=logFC_edgeR))+geom_point(color="red")+stat_smooth(method="lm",se=FALSE)+stat_cor(data=a, method = "pearson")
png("cor_edgeR_limma.png")
p1
dev.off()

# for edgeR vs DESeq2
a1=head(td_edgeR_DESeq2[order(td_edgeR_DESeq2$logFC_edgeR),],250)
a2=tail(td_edgeR_DESeq2[order(td_edgeR_DESeq2$logFC_edgeR),],250)
a=rbind(a1,a2)
head(a)
ggplot(a,aes(y=log2FC_DESeq2,x=logFC_edgeR))+geom_point()
p2=ggplot(data=a, aes(y=log2FC_DESeq2, x=logFC_edgeR))+geom_point(color="red")+stat_smooth(method="lm",se=FALSE)+stat_cor(data=a, method = "pearson")
png("cor_edgeR_DESeq2.png")
p2
dev.off()

# for limma vs DESeq2
a1=head(td_DESeq2_limma[order(td_DESeq2_limma$logFC_limma),],250)
a2=tail(td_DESeq2_limma[order(td_DESeq2_limma$logFC_limma),],250)
a=rbind(a1,a2)
head(a)
ggplot(a,aes(y=log2FC_DESeq2,x=logFC_limma))+geom_point()
p3=ggplot(data=a, aes(y=log2FC_DESeq2, x=logFC_limma))+geom_point(color="red")+stat_smooth(method="lm",se=FALSE)+stat_cor(data=a, method = "pearson")
png("cor_limma_DESeq2.png")
p3
dev.off()

library(patchwork)
p1+p2+p3+ plot_layout(guides = "collect")
png("cor_limma_DESeq2_edgeR.png")
p1+p2+p3+ plot_layout(guides = "collect")
dev.off()


#三大R包差异基因对比之韦恩图
rm(list = ls())
load("TCGA-BRCAgdc.Rdata")
load("TCGA-BRCADEG.Rdata")
load("TCGA-BRCApcaplot.Rdata")
UP=function(df){
  rownames(df)[df$change=="UP"]
}
DOWN=function(df){
  rownames(df)[df$change=="DOWN"]
}

up = intersect(intersect(UP(DESeq2_DEG),UP(edgeR_DEG)),UP(limma_voom_DEG))
down = intersect(intersect(DOWN(DESeq2_DEG),DOWN(edgeR_DEG)),DOWN(limma_voom_DEG))
dat = log2(exp+1)
hp = draw_heatmap(dat[c(up,down),],group_list,scale_before = T)



#上调、下调基因分别画维恩图
up_genes = list(Deseq2 = UP(DESeq2_DEG),
                edgeR = UP(edgeR_DEG),
                limma = UP(limma_voom_DEG))

down_genes = list(Deseq2 = DOWN(DESeq2_DEG),
                  edgeR = DOWN(edgeR_DEG),
                  limma = DOWN(limma_voom_DEG))


up.plot <- draw_venn(up_genes,"UPgene")
down.plot <- draw_venn(down_genes,"DOWNgene")

#维恩图拼图，终于搞定

library(patchwork)
#up.plot + down.plot
# 拼图
pca.plot + hp+up.plot +down.plot+ plot_layout(guides = "collect")
ggsave(paste0(cancer_type,"heat_ve_pca.png"),width = 15,height = 10)


#以下又为test #for upset图  参考https://www.jianshu.com/p/b702f644cad3
#install.packages("UpSetR")
library(UpSetR)
library(dplyr)
library(tidyr)
up_DESeq2=UP(DESeq2_DEG)
up_edgeR=UP(edgeR_DEG)
up_limma=UP(limma_voom_DEG)
down_DESeq2=DOWN(DESeq2_DEG)
down_edgeR=DOWN(edgeR_DEG)
down_limma=DOWN(limma_voom_DEG)

input <- c(
  'DESeq2_up'=  length(up_DESeq2),
  'edgeR_up' =  length(up_edgeR),
  'limma_up' = length(up_limma),
  'DESeq2_down'  =length(down_DESeq2),
  'edgeR_down'  = length(down_edgeR),
  'limma_down'  =length(down_limma),
'DESeq2_up&edgeR_up'  =length(intersect(UP(DESeq2_DEG),UP(edgeR_DEG))),
'DESeq2_up&limma_up'  = length(intersect(UP(DESeq2_DEG),UP(limma_voom_DEG))),
'edgeR_up&limma_up'  =length(intersect(UP(limma_voom_DEG),UP(edgeR_DEG))),
'DESeq2_up&edgeR_up&limma_up'=length(intersect(intersect(UP(DESeq2_DEG),UP(edgeR_DEG)),UP(limma_voom_DEG))),
'DESeq2_down&edgeR_down'  =length(intersect(DOWN(DESeq2_DEG),DOWN(edgeR_DEG))),
'DESeq2_down&limma_down'  = length(intersect(DOWN(DESeq2_DEG),DOWN(limma_voom_DEG))),
'edgeR_down&limma_down'  =length(intersect(DOWN(limma_voom_DEG),DOWN(edgeR_DEG))),
'DESeq2_down&edgeR_down&limma_down'=length(intersect(intersect(DOWN(DESeq2_DEG),DOWN(edgeR_DEG)),DOWN(limma_voom_DEG)))
)

data <- fromExpression(input)
p1 <- upset(data, nsets = 6, 
            sets = c('DESeq2_up',
                     'edgeR_up',
                     'limma_up',
                     'DESeq2_down',
                     'edgeR_down',
                     'limma_down'),
            keep.order = TRUE,
            # number.angles = 30, 
            point.size = 5, 
            line.size = 1.3, 
            mainbar.y.label = "DEGs Intersection", 
            sets.x.label = "",
            mb.ratio = c(0.60, 0.40),
            text.scale = c(2, 3, 4, 1.5,1.5, 2))
png("upset.png")
p1
dev.off()