哔哩哔哩的生信技能树R语言视频教程学习笔记

https://www.bilibili.com/video/av25643438
哔哩哔哩，劈里啪啦，1.5倍速飞快听完，然后主要整理了R语言最新十道题，压力就是第一生产力啊。http://www.bio-info-trainee.com/3750.html

准备工作--安装所需的包

cran_packages <- c('tidyverse',
                   'ggpubr',
                   'ggstatsplot')
Biocductor_packages <- c('org.Hs.eg.db',
                         'hgu133a.db',
                         'CLL',
                         'hgu95av2.db',
                         'survminer',
                         'survival',
                         'hugene10sttranscriptcluster',
                         'limma')
if(length(getOption("CRAN"))==0) options(CRAN="https://mirrors.tuna.tsinghua.edu.cn/CRAN/")
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(length(getOption("CRAN"))==0) options(CRAN="https://mirrors.tuna.tsinghua.edu.cn/CRAN/")
if(!require("BiocManager")) install.packages("BiocManager",update = F,ask = F)
if(length(getOption("BioC_mirror"))==0) options(BioC_mirror="https://mirrors.ustc.edu.cn/bioc/")
# 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) 
  }
}

1.找到指定ensembl ID与symbol的对应关系

ENSG00000000003.13
ENSG00000000005.5
ENSG00000000419.11
ENSG00000000457.12
ENSG00000000460.15
ENSG00000000938.11

思路：在注释包中有gene_id与symbol、gene_id与ensembl_id的对应关系。

#将以上基因id保存在a.txt，存放于工作目录下。
rm(list=ls())
options(stringsAsFactors = F)

a=read.table('e1.txt')

g2s <- toTable(org.Hs.egSYMBOL)
g2e <- toTable(org.Hs.egENSEMBL)
a$V1 = apply(a[1], 1,function(x){
  str_split(x,'[.]')[[1]][1]
}) %>% unlist
colnames(a) <- 'ensembl_id'
tmp <- merge(a,g2e, by="ensembl_id")
result <- merge(tmp,g2s, by="gene_id")[-1]

2.根据探针名找对应symbol ID

1053_at
117_at
121_at
1255_g_at
1316_at
1320_at
1405_i_at
1431_at
1438_at
1487_at
1494_f_at
1598_g_at
160020_at
1729_at
177_at

思路：找到注释包中探针与symbol的对应关系然后取子集

rm(list=ls())
options(stringsAsFactors = F)
library(hgu133a.db)
p2s=toTable(hgu133aSYMBOL)
a=read.table('e2.txt')
colnames(a) <- colnames(p2s)[1]
# 方法一：利用merge
tmp1 <- merge(a,p2s, by="probe_id")
# 方法二：利用match得到第一组向量在第二组中的坐标
tmp2 <- p2s[match(a$probe_id,p2s$probe_id),]
## 附：判断得到的两组结果是否一致
# 法一：
identical(tmp1,tmp2) #返回逻辑值
# 法二：
dplyr::setdiff(tmp1,tmp2) #返回两组的差别【没差就返回空】

3.根据symbol找基因表达量并作图

找到R包CLL内置的数据集的表达矩阵里面的TP53基因的表达量，并且绘制在 progres.-stable分组的boxplot图,并通过 ggpubr 进行美化。
探针的三大内容：表达矩阵assay/exprs、探针信息featureData、样本信息phenoData

# 从内置数据集的表达矩阵中找TP53基因的表达量
rm(list=ls())
options(stringsAsFactors = F)
suppressMessages(library(CLL))
data(sCLLex)
# sCLLex
exprSet <- exprs(sCLLex) #探针的表达量
pdata <- pData(sCLLex) #sampleID与disease的对应关系
p2s <- toTable(hgu95av2SYMBOL) #探针与symbol的对应关系
p3 <- filter(p2s,symbol=='TP53')
# boxplot [find TP53 has 3 probe IDs]
probe_tp53 <- c("1939_at","1974_s_at","31618_at")
i = 3 #可换1，2
boxplot(exprSet[probe_tp53[i],] ~ pdata$Disease)

#用ggpubr作图
#http://www.sthda.com/english/articles/24-ggpubr-publication-ready-plots/
exp_tab <- rownames_to_column(as.data.frame(exprSet))
exp_tab2 <- gather(exp_tab,
                   key = 'sample',
                   value = 'exp',-1)
pdata <- rownames_to_column(pdata)
exp_tab3 <- merge(exp_tab2,pdata,by.x='sample',by.y='rowname')
i=1 ###可换1，2
dev.off()
p <- ggboxplot(filter(exp_tab3,rowname==probe_tp53[i]), 
               x = 'Disease',
               y = 'exp',
               color = "Disease", palette =c("#00AFBB", "#E7B800", "#FC4E07"),
               add = "jitter", shape = "Disease")
p

4.BRCA1基因表达量

找到BRCA1基因在TCGA数据库的乳腺癌数据集(Breast Invasive Carcinoma (TCGA, PanCancer Atlas))的表达情况

提示：使用http://www.cbioportal.org/index.do 定位数据集：http://www.cbioportal.org/datasets
该基因有四个亚型，用ggbetweenstats作图比较一下。

rm(list=ls())
options(stringsAsFactors = F)
#ID，四个亚型，表达量
f <- read.csv("e4-plot.txt", sep = "\t")
## boxplot
colnames(f) <- c("id", "subtype", "expression", "mut")
da <- f
library(ggstatsplot)
ggbetweenstats(data = da, 
               x = subtype,
               y = expression)
library(ggplot2)
ggsave("e4-BRCA1-TCGA.png")

5 .TP53 生存分析

找到TP53基因在TCGA数据库的乳腺癌数据集的表达量分组看其是否影响生存

提示使用：http://www.oncolnc.org/
思路：生存分析，TP53表达量分为高低两组做图比较

# Use http://www.oncolnc.org/ to get raw csv da

rm(list=ls())
options(stringsAsFactors = F)

f <- read.csv('e5-BRCA_7157_50_50.csv')
library(ggstatsplot)
ggbetweenstats(data = da,
               x = Group,
               y = Expression)
da <- f

library(ggplot2)
library(survival)
library(survminer)
table(da$Status)

da$Status <- ifelse(da$Status == "Dead", 1, 0)
survf <- survfit(Surv(Days,Status)~Group, data=da)
ggsurvplot(survf, conf.int = F, pval = T)

# complex survplot
ggsurvplot(survf,palette = c("orange", "navyblue"),
           risk.table = T, pval = T,
           conf.int = T, xlab = "Time of days",
           ggtheme = theme_light(),
           ncensor.plot = T)
ggsave("survival_TP53_in_BRCA_taga.png")

6.从表达矩阵中提取指定基因画热图

下载数据集GSE17215的表达矩阵并且提取下面的基因画热图

ACTR3B ANLN BAG1 BCL2 BIRC5 BLVRA CCNB1 CCNE1 CDC20 CDC6 CDCA1 CDH3 CENPF CEP55 CXXC5 EGFR ERBB2 ESR1 EXO1 FGFR4 FOXA1 FOXC1 GPR160 GRB7 KIF2C KNTC2 KRT14 KRT17 KRT5 MAPT MDM2 MELK MIA MKI67 MLPH MMP11 MYBL2 MYC NAT1 ORC6L PGR PHGDH PTTG1 RRM2 SFRP1 SLC39A6 TMEM45B TYMS UBE2C UBE2T

提示：根据基因名拿到探针ID，缩小表达矩阵绘制热图，没有检查到的基因直接忽略即可。

## Exercise 5: Retrive genes from GEO to plot heatmap

rm(list=ls())
options(stringsAsFactors = F)
#下载和表达矩阵
library(GEOquery)
GSE <- "GSE17215"
if(!file.exists(GSE)){
    geo <- getGEO(GSE, destdir = '.', getGPL = F, AnnotGPL = F)
    save(geo, file = paste0(GSE,'.eSet.Rdata'))
}
load(paste0(GSE,'.eSet.Rdata'))
expr <- exprs(geo[[1]])
p2s=toTable(hgu133aSYMBOL);head(p2s)
expr <- expr[p2s$probe_id,] #有的id找不到注释直接删掉
gp <- "ACTR3B ANLN BAG1 BCL2 BIRC5 BLVRA CCNB1 CCNE1 CDC20 CDC6 CDCA1 CDH3 CENPF CEP55 CXXC5 EGFR ERBB2 ESR1 EXO1 FGFR4 FOXA1 FOXC1 GPR160 GRB7 KIF2C KNTC2 KRT14 KRT17 KRT5 MAPT MDM2 MELK MIA MKI67 MLPH MMP11 MYBL2 MYC NAT1 ORC6L PGR PHGDH PTTG1 RRM2 SFRP1 SLC39A6 TMEM45B TYMS UBE2C UBE2T"
gp <- strsplit(gp, ' ')[[1]]
tmp <- dplyr::filter(p2s, p2s$symbol %in% gp)
tmp2 <- tibble::rownames_to_column(data.frame(expr),"probe_id")
tmp3 <- merge(tmp,tmp2,by="probe_id")

tmp3$median <- apply(tmp3[,3:ncol(tmp3)], 1, median)
tmp3 <- tmp3[order(tmp3$symbol, tmp3$median, decreasing = T),]
tmp3 <- tmp3[!duplicated(tmp3$symbol),]
rownames(tmp3) <- tmp3$symbol
tmp3 <- tmp3[,-c(1,2,ncol(tmp3))]

gp_expr <- log2(tmp3)
pheatmap::pheatmap(gp_expr, scale = "row") 

7.相关性计算和热图

下载数据集GSE24673的表达矩阵计算样本的相关性并且绘制热图，需要标记上样本分组信息
相关性分析：

rm(list=ls())
options(stringsAsFactors = F)

library(GEOquery)
GSE <- "GSE24673"
if(!file.exists(GSE)){
  geo <- getGEO(GSE, destdir = '.', getGPL = F, AnnotGPL = F)
  save(geo, file = paste0(GSE,'.eSet.Rdata'))
}
load(paste0(GSE,'.eSet.Rdata'))
expr <- exprs(geo[[1]])
dim(expr)
expr[1:4,1:4]
pdata <- pData(geo[[1]]) 
# 自己根据pdata第八列做一个分组信息矩阵
grp <- c('rbc','rbc','rbc',
         'rbn','rbn','rbn',
         'rbc','rbc','rbc',
         'normal','normal')

grp_df <- data.frame(group=grp)
rownames(grp_df) <- colnames(expr)
new_cor <- cor(expr)
pheatmap::pheatmap(new_cor, annotation_col = grp_df)

8.找到芯片平台对应的注释包

找到 GPL6244 platform of Affymetrix Human Gene 1.0 ST Array 对应的R的bioconductor注释包，并且安装它！
https://mp.weixin.qq.com/s/sVSsV2fWZOQmNd72Vb3SmQ
https://www.jianshu.com/p/40b560755cdf

options()$repos
options()$BioC_mirror 
options(BioC_mirror="https://mirrors.ustc.edu.cn/bioc/")
options("repos" = c(CRAN="https://mirrors.tuna.tsinghua.edu.cn/CRAN/"))
BiocManager::install("hugene10sttranscriptcluster",ask = F,update = F)
options()$repos
options()$BioC_mirror

9.找到指定探针和对应的基因

rm(list=ls())
options(stringsAsFactors = F)

library(GEOquery)
GSE <- "GSE42872"
if(!file.exists(GSE)){
  geo <- getGEO(GSE, destdir = '.', getGPL = F, AnnotGPL = F)
  save(geo, file = paste0(GSE,'.eSet.Rdata'))
}
load(paste0(GSE,'.eSet.Rdata'))
expr <- exprs(geo[[1]])

dim(expr)
expr[1:4,1:4]

# 选出所有样本的(mean/sd/mad/)最大的探针

sort(apply(expr,1,mean),decreasing = T)[1]
sort(apply(expr,1,sd),decreasing = T)[1]
sort(apply(expr,1,mad),decreasing = T)[1]

下载数据集GSE42872的表达矩阵，并且分别挑选出所有样本的(平均表达量/sd/mad/)最大的探针，并且找到它们对应的基因。

10.limma 差异分析

下载数据集GSE42872的表达矩阵，并且根据分组使用limma做差异分析，得到差异结果矩阵

# 接第九题，得到表型信息，然后用limma做差异分析
pdata <- pData(geo[[1]]) 
grp <- unlist(lapply(pdata$title, function(x){
  strsplit(x, ' ')[[1]][4]
}))

suppressMessages(library(limma))
#limma needs：表达矩阵（expr：需要用log后的矩阵）、分组矩阵（design）、比较矩阵（contrast）
#先做一个分组矩阵～design，说明progres是哪几个样本，stable又是哪几个
design <- model.matrix(~0+factor(grp))
colnames(design) <- levels(factor(grp))
rownames(design) <- colnames(expr)
design
#再做一个比较矩阵【一般是case比control】
contrast<-makeContrasts(paste0(unique(grp),collapse = "-"),levels = design)
contrast

##step1
fit <- lmFit(expr,design)
##step2
fit2 <- contrasts.fit(fit, contrast) 
fit2 <- eBayes(fit2)  
##step3
mtx = topTable(fit2, coef=1, n=Inf)
DEG_mtx = na.omit(mtx) 
View(DEG_mtx)

# 火山图

DEG=DEG_mtx

if(T){
  logFC_cutoff <- with(DEG,mean(abs(logFC)) + 2*sd(abs(logFC)) )
  
  DEG$change = as.factor(ifelse(DEG$P.Value < 0.05 & abs(DEG$logFC) > logFC_cutoff,
                                ifelse(DEG$logFC > logFC_cutoff ,'UP','DOWN'),'NOT')
  )
  title <- paste0('log2FoldChange cutoff: ',round(logFC_cutoff,3),
                  '\nUp-regulated genes: ',nrow(DEG[DEG$change =='UP',]) ,
                  '\nDown-regulated genes: ',nrow(DEG[DEG$change =='DOWN',])
  )
}
library(ggplot2)
vol1 = ggplot(data=DEG, aes(x=logFC, y=-log10(P.Value), color=change)) +
  geom_point(alpha=0.4, size=1.75) +
  theme_set(theme_set(theme_bw(base_size=20)))+
  xlab("log2FoldChange") + ylab("-log10 p-value") +
  ggtitle( title ) + theme(plot.title = element_text(size=15,hjust = 0.5))+
  scale_colour_manual(values = c('blue','black','red')) # according to the levels(DEG$change)
print(vol1)