转录组分析(11) - GO/KEGG富集分析(3)

对于没有Org.Db的非模式生物，由于某些原因不制作Org.Db，然后利用clusterProfiler做富集

数据格式

## Annotation_GO.xlsx
Gene.ID  Biological.Pathway  BP.Description  Molecular.Function  MF.Description Cellular.Component      CC.Description
gene_1  GO:0007283 spermatogenesis  GO:0003677 DNA binding  GO:0000786//GO:0005634  nucleosome//nucleus

## Annotation_KO.xlsx
Gene.ID  1st-level(pathway)  2nd-level(pathway)  3rd-level(pathway) Pathway.id  KO.id  KO.Description  EC
gene_1  Cellular Processes  Cell growth and death  Cell cycle  ko04110  K02180  cell cycle arrest protein BUB3  -

## test_list.xlsx
gene_id  FC
gene_1  1.85

数据准备

#coding:utf-8

library(openxlsx)
library(tidyr)
library(stringr)
library(dplyr)

getwd()
dir()
rm(list=ls())

#定义GO数据结构化函数
data_clean <-function(x,ont){
  colnames(x) <- c("gene","GO_id","GO_name")
  gterms <- x %>%
    dplyr::select(gene, GO_id,GO_name) %>% na.omit()
  
  gene2go <- data.frame(gene = character(),
                           GO_id = character(),
                           GO_name = character(),
                           Ont = character()
  )
  for (row in 1:nrow(gterms)){
    gene_terms_id <- str_split(gterms[row,"GO_id"], "//", simplify = FALSE)[[1]] 
    gene_terms_name <- str_split(gterms[row,"GO_name"], "//", simplify = FALSE)[[1]]
    gene_id <- gterms[row, "gene"][[1]]
    
    tmp <- data.frame(gene = rep(gene_id, length(gene_terms_id)),
                      GO_id = gene_terms_id,
                      GO_name = gene_terms_name,
                      Ont = rep(ont, length(gene_terms_id))
    )
    gene2go <- rbind(gene2go, tmp)
  }
  return(gene2go)
}
# 载入数据
go_file <- read.xlsx("Annotation_GO.xlsx")
go_file[go_file=="--"]<-NA
go_bp <- go_file[,1:3]

# 按Ont类型格式化数据
## BP
go_bp <- go_file[,1:3]
gene2go_bp <- data_clean(go_bp,"BP")
go2gene_bp <- distinct(gene2go_bp[,c(2,1)])
go2name_bp <- distinct(gene2go_bp[,c(2,3)])
## MF
go_mf <- go_file[,c(1,4,5)]
gene2go_mf <- data_clean(go_mf,"MF")
go2gene_mf <- distinct(gene2go_mf[,c(2,1)])
go2name_mf <- distinct(gene2go_mf[,c(2,3)])
## CC
go_cc <- go_file[,c(1,6,7)]
gene2go_cc <- data_clean(go_cc,"CC")
go2gene_cc <- distinct(gene2go_cc[,c(2,1)])
go2name_cc <- distinct(gene2go_cc[,c(2,3)])

#保存关键数据
save(go2gene_bp,go2name_bp,go2gene_mf,go2name_mf,go2gene_cc,go2name_cc,file="XX.go.RData")

#富集时调用
#load(file = "XX.go.RData")

#Pathway数据格式化
rm(list=ls())
ko_file <- read.xlsx("Annotation_KO.xlsx")
ko_file[ko_file=="--"]<-NA
gene2ko <- ko_file[,c(1,5,4,3,2)]
colnames(gene2ko) <- c("gene","pathway_id","pathway_name_lv3","pathway_name_lv2","pathway_name_lv1")
ko2gene <- distinct(gene2ko[,c(2,1)])
ko2name <- distinct(gene2ko[,c(2,3)])
#保存关键数据
save(gene2ko,ko2gene,ko2name,file="XX.ko.RData")
#富集时调用
#load(file = "XX.ko.RData")

富集分析绘图

if (!requireNamespace("BiocManager", quietly = TRUE))
  install.packages("BiocManager")
BiocManager::install(c("clusterProfiler"))

library(clusterProfiler)
library(openxlsx)

## __init__
rm(list = ls())
load(file = "XX.go.RData")
load(file = "AX.ko.RData")
## 导入数据
test_list <- read.xlsx("test_list.xlsx")
## 准备数据格式
test_gene <- test_list[,1]
fc_list <- test_list[,2]
names(fc_list)<- as.character(test_list[,1])
fc_list <- sort(fc_list,decreasing = T)
## GO/KO富集
enrich_go <- enricher(
  test_gene,              # 前景基因集
  pvalueCutoff = 0.05,    # p值阈值
  pAdjustMethod = "fdr",   # 调整p值校正方式。（one of "holm", "hochberg", "hommel", "bonferroni", "BH", "BY", "fdr", "none"）
  qvalueCutoff = 0.05,       # q值阈值
  TERM2GENE = go2gene_bp, # 修改以确定富集的Ont类型（go2gene_bp, go2gene_mf, go2gene_cc,ko2gene）
  TERM2NAME = go2name_bp  # 跟随上一行一起修改 (go2name_bp, go2name_mf, go2name_cc, ko2name)
  #universe,              # 背景基因集
  #minGSSize = 10,
  #maxGSSize = 500,
)
## 保存富集结果
write.csv(enrich_go@result,file="XXX.go.csv")
## 绘制条形图
plot_data <- enrich_go@result[1:10,c(2,9,5)]
library(ggplot2)
pp1=ggplot(plot_data, aes(x=reorder(plot_data$Description,plot_data$Count ),y=Count,fill=-1*log10(plot_data$pvalue)))
pbar=pp1+geom_bar(stat="identity")+theme_minimal()+coord_flip()
output_bar_pdf <- paste("XXX",".bar",".pdf",sep = '')
ggsave(pbar,file=output_bar_pdf,width=10,height=6)
## 绘制气泡图
library(ggplot2)
pp= ggplot(plot_data, aes(x=plot_data$pvalue,y=reorder(plot_data$Description,plot_data$pvalue))) 
pbubble = pp + geom_point(aes(size=plot_data$Count,color=-1*log10(plot_data$pvalue))) + scale_colour_gradient(low="green",high = "red")+ theme_minimal()
output_bb_pdf <- paste("XXX",".bb",".pdf",sep = '')
ggsave(pbubble,file=output_bb_pdf,width=10,height=6)

## GSEA富集分析
gse_result <- GSEA(
  fc_list,
  exponent = 1,
  eps = 1e-10,
  pvalueCutoff = 1,
  pAdjustMethod = "fdr",
  TERM2GENE = go2gene_bp,    # 修改以确定富集的Ont类型（go2gene_bp, go2gene_mf, go2gene_cc,ko2gene）
  TERM2NAME = go2name_bp,    # 跟随上一行一起修改 (go2name_bp, go2name_mf, go2name_cc, ko2name)
  verbose = TRUE,
  by = "fgsea"
  #minGSSize = 10,
  #maxGSSize = 500,
)
## 保存结果
write.csv(gse_result@result,file = "XXX.csv")
## 批量绘制GSEA图
for(i in 1:length(gse_result@result$ID)) {
  id <- gse_result@result$ID[i]
  gsefig<- gseaplot(gse_result,geneSetID = id)
  output_gse_pdf <- paste("XXX",i,".",id,".gse",".pdf",sep = '')
  ggsave(gsefig,file=output_gse_pdf,width=10,height=6)
}