第二次RNA-seq实战总结(3)-用DESeq2进行基因表达差
2018-11-04 本文已影响7人
邱俊辉
DESeq2是一个用于分析基因表达差异的R包，具体操作姚在R语言中运行
1.R语言安装DESeq2
>source("https://bioconductor.org/biocLite.R")
>biocLite("DESeq2")
2.载入基因表达量文件，添加列名
> setwd("C:\\Users\\18019\\Desktop\\counts")
> options(stringsAsFactors=FALSE)
> control1<-read.table("SRR957677_counts.txt",sep = "\t",col.names = c("gene_id","control1"))
> head(control1)
gene_id control1
1 ENSG00000000003.14_2 1576
2 ENSG00000000005.5_2 0
3 ENSG00000000419.12_2 756
4 ENSG00000000457.13_3 301
5 ENSG00000000460.16_5 764
6 ENSG00000000938.12_2 0
> control2<-read.table("SRR957678_counts.txt",sep = "\t",col.names = c("gene_id","control2"))
> treat1<-read.table("SRR957679_counts.txt",sep = "\t",col.names = c("gene_id","treat1"))
>treat2<-read.table("SRR957680_counts.txt",sep = "\t",col.names = c("gene_id","treat2"))
3.数据整合
> raw_count <- merge(merge(control1, control2, by="gene_id"), merge(treat1, treat2, by="gene_id"))
> head(raw_count)
gene_id control1 control2 treat1 treat2
1 __alignment_not_unique 7440131 2973831 7861484 8676884
2 __ambiguous 976485 412543 1014239 1179051
3 __no_feature 1860117 768637 1289737 1812056
4 __not_aligned 1198545 572588 1256232 1348068
5 __too_low_aQual 0 0 0 0
6 ENSG00000000003.14_2 1576 713 1589 1969
#删除前五行
>raw_count_filt <- raw_count[-1:-5,]
#因为我们无法在EBI数据库上直接搜索找到ENSMUSG00000024045.5这样的基因，只能是ENSMUSG00000024045的整数，没有小数点，所以需要进一步替换为整数的形式。
#将_后面的数字替换为空赋值给a
>a<- gsub("\\_\\d*", "", raw_count_filt$gene_id)
#将.后面的数字替换为空赋值给ENSEMBL
>ENSEMBL <- gsub("\\.\\d*", "", a)
# 将ENSEMBL重新添加到raw_count_filt1矩阵
>row.names(raw_count_filt) <- ENSEMBL
> raw_count_filt <- cbind(ENSEMBL,raw_count_filt)
> colnames(raw_count_filt)[1] <- c("ensembl_gene_id")
>head(raraw_count_filt )
ensembl_gene_id gene_id control1 control2 treat1 treat2
ENSG00000000003 ENSG00000000003 ENSG00000000003.14_2 1576 713 1589 1969
ENSG00000000005 ENSG00000000005 ENSG00000000005.5_2 0 0 0 1
ENSG00000000419 ENSG00000000419 ENSG00000000419.12_2 756 384 806 984
ENSG00000000457 ENSG00000000457 ENSG00000000457.13_3 301 151 217 324
ENSG00000000460 ENSG00000000460 ENSG00000000460.16_5 764 312 564 784
ENSG00000000938 ENSG00000000938 ENSG00000000938.12_2 0 0 0 0
4.对基因进行注释-获取gene_symbol
用bioMart对ensembl_id转换成gene_symbol
> library("biomaRt")
> library("curl")
> mart <- useDataset("hsapiens_gene_ensembl", useMart("ensembl"))
> my_ensembl_gene_id <- row.names(raw_count_filt)
> options(timeout = 4000000)
> hg_symbols<- getBM(attributes=c('ensembl_gene_id','hgnc_symbol',"chromosome_name", "start_position","end_position", "band"), filters= 'ensembl_gene_id', values = my_ensembl_gene_id, mart = mart)
#将合并后的表达数据框raw_count_filt和注释得到的hg_symbols整合为一
readcount <- merge(raw_count_filt, hg_symbols, by="ensembl_gene_id")
> head(readcount)
ensembl_gene_id gene_id control1 control2 treat1 treat2 hgnc_symbol chromosome_name start_position
1 ENSG00000000003 ENSG00000000003.14_2 1576 713 1589 1969 TSPAN6 X 100627109
2 ENSG00000000005 ENSG00000000005.5_2 0 0 0 1 TNMD X 100584802
3 ENSG00000000419 ENSG00000000419.12_2 756 384 806 984 DPM1 20 50934867
4 ENSG00000000457 ENSG00000000457.13_3 301 151 217 324 SCYL3 1 169849631
5 ENSG00000000460 ENSG00000000460.16_5 764 312 564 784 C1orf112 1 169662007
6 ENSG00000000938 ENSG00000000938.12_2 0 0 0 0 FGR 1 27612064
end_position band
1 100639991 q22.1
2 100599885 q22.1
3 50958555 q13.13
4 169894267 q24.2
5 169854080 q24.2
6 27635277 p35.3
#输出count表达矩阵
> write.csv(readcount, file='readcount_all,csv')
> readcount<-raw_count_filt[ ,-1:-2]
> write.csv(readcount, file='readcount.csv')
> head(readcount)
control1 control2 treat1 treat2
ENSG00000000003 1576 713 1589 1969
ENSG00000000005 0 0 0 1
ENSG00000000419 756 384 806 984
ENSG00000000457 301 151 217 324
ENSG00000000460 764 312 564 784
ENSG00000000938 0 0 0 0
5.DEseq2筛选差异表达基因并注释(bioMart)
#载入数据（countData和colData）
> mycounts<-readcount
> head(mycounts)
control1 control2 treat1 treat2
ENSG00000000003 1576 713 1589 1969
ENSG00000000005 0 0 0 1
ENSG00000000419 756 384 806 984
ENSG00000000457 301 151 217 324
ENSG00000000460 764 312 564 784
ENSG00000000938 0 0 0 0
> condition <- factor(c(rep("control",2),rep("treat",2)), levels = c("control","treat"))
> condition
[1] control control treat treat
Levels: control treat
> colData <- data.frame(row.names=colnames(mycounts), condition)
> colData
condition
control1 control
control2 control
treat1 treat
treat2 treat
构建dds对象,开始DESeq流程
>library("DESeq2")
> dds <- DESeqDataSetFromMatrix(mycounts, colData, design= ~ condition)
> dds <- DESeq(dds)
estimating size factors
estimating dispersions
gene-wise dispersion estimates
mean-dispersion relationship
final dispersion estimates
fitting model and testing
> dds
class: DESeqDataSet
dim: 60880 4
metadata(1): version
assays(4): counts mu H cooks
rownames(60880): ENSG00000000003 ENSG00000000005 ... ENSG00000285993 ENSG00000285994
rowData names(22): baseMean baseVar ... deviance maxCooks
colnames(4): control1 control2 treat1 treat2
colData names(2): condition sizeFactor
#查看总体结果
> res = results(dds, contrast=c("condition", "control", "treat"))
> res = res[order(res$pvalue),]
> head(res)
log2 fold change (MLE): condition control vs treat
Wald test p-value: condition control vs treat
DataFrame with 6 rows and 6 columns
baseMean log2FoldChange lfcSE stat pvalue
<numeric> <numeric> <numeric> <numeric> <numeric>
ENSG00000178691 1025.66218695436 2.83012875791025 0.225513526042636 12.5497073615672 3.98981786210676e-36
ENSG00000135535 2415.77359618136 1.22406336488047 0.183431131037356 6.67314952460929 2.5037106203736e-11
ENSG00000164172 531.425786834548 1.30449018960413 0.207785830749451 6.27805170785243 3.42841906697453e-10
ENSG00000172239 483.998634607265 1.31701332235233 0.215453141699223 6.11275988813803 9.79226522597759e-10
ENSG00000237296 53.0114998109978 2.70139282483841 0.480033904207378 5.62750422660019 1.82835684560772e-08
ENSG00000196504 3592.67315807893 1.09372324353448 0.200308218929736 5.46020153031335 4.75594407815571e-08
padj
<numeric>
ENSG00000178691 3.90682965057494e-32
ENSG00000135535 1.22581671973492e-07
ENSG00000164172 1.11903598346049e-06
ENSG00000172239 2.39714652731931e-06
ENSG00000237296 NA
ENSG00000196504 9.31404088266014e-05
> summary(res)
out of 33100 with nonzero total read count
adjusted p-value < 0.1
LFC > 0 (up) : 78, 0.24%
LFC < 0 (down) : 15, 0.045%
outliers [1] : 0, 0%
low counts [2] : 23308, 70%
(mean count < 135)
[1] see 'cooksCutoff' argument of ?results
[2] see 'independentFiltering' argument of ?results
#这里可以看到有78个基因上调，15个基因下调
#将分析结果输出
> write.csv(res,file="All_results.csv")
提取差异表达基因
这里我用的方法是倍差法
获取padj（p值经过多重校验校正后的值）小于0.05，表达倍数取以2为对数后大于1或者小于-1的差异表达基因
> diff_gene_deseq2 <-subset(res, padj < 0.05 & abs(log2FoldChange) > 1)
> dim(diff_gene_deseq2)
[1] 21 6
> head(diff_gene_deseq2)
log2 fold change (MLE): condition control vs treat
Wald test p-value: condition control vs treat
DataFrame with 6 rows and 6 columns
baseMean log2FoldChange lfcSE stat pvalue
<numeric> <numeric> <numeric> <numeric> <numeric>
ENSG00000178691 1025.66218695436 2.83012875791025 0.225513526042636 12.5497073615672 3.98981786210676e-36
ENSG00000135535 2415.77359618136 1.22406336488047 0.183431131037356 6.67314952460929 2.5037106203736e-11
ENSG00000164172 531.425786834548 1.30449018960413 0.207785830749451 6.27805170785243 3.42841906697453e-10
ENSG00000172239 483.998634607265 1.31701332235233 0.215453141699223 6.11275988813803 9.79226522597759e-10
ENSG00000196504 3592.67315807893 1.09372324353448 0.200308218929736 5.46020153031335 4.75594407815571e-08
ENSG00000163848 633.066990185649 1.15489622775117 0.219655131372136 5.25777030810433 1.45812478575117e-07
padj
<numeric>
ENSG00000178691 3.90682965057494e-32
ENSG00000135535 1.22581671973492e-07
ENSG00000164172 1.11903598346049e-06
ENSG00000172239 2.39714652731931e-06
ENSG00000196504 9.31404088266014e-05
ENSG00000163848 0.000230253090268928
#输出差异基因
> write.csv(diff_gene_deseq2,file= "DEG_treat_vs_control.csv")
#用bioMart对差异表达基因进行注释
> library("biomaRt")
> library("curl")
> hg_symbols<- getBM(attributes=c('ensembl_gene_id','external_gene_name',"description"),filters = 'ensembl_gene_id', values = my_ensembl_gene_id, mart = mart)
> head(hg_symbols)
ensembl_gene_id external_gene_name
1 ENSG00000011405 PIK3C2A
2 ENSG00000100731 PCNX1
3 ENSG00000128512 DOCK4
4 ENSG00000135535 CD164
5 ENSG00000140526 ABHD2
6 ENSG00000144228 SPOPL
description
1 phosphatidylinositol-4-phosphate 3-kinase catalytic subunit type 2 alpha [Source:HGNC Symbol;Acc:HGNC:8971]
2 pecanex 1 [Source:HGNC Symbol;Acc:HGNC:19740]
3 dedicator of cytokinesis 4 [Source:HGNC Symbol;Acc:HGNC:19192]
4 CD164 molecule [Source:HGNC Symbol;Acc:HGNC:1632]
5 abhydrolase domain containing 2 [Source:HGNC Symbol;Acc:HGNC:18717]
6 speckle type BTB/POZ protein like [Source:HGNC Symbol;Acc:HGNC:27934]
#合并数据:res结果hg_symbols合并成一个文件
> ensembl_gene_id<-rownames(diff_gene_deseq2)
> diff_gene_deseq2<-cbind(ensembl_gene_id,diff_gene_deseq2)
> colnames(diff_gene_deseq2)[1]<-c("ensembl_gene_id")
> diff_name<-merge(diff_gene_deseq2,hg_symbols,by="ensembl_gene_id")
> head(diff_name)
DataFrame with 6 rows and 9 columns
ensembl_gene_id baseMean log2FoldChange lfcSE stat pvalue
<character> <numeric> <numeric> <numeric> <numeric> <numeric>
1 ENSG00000011405 1600.01408863821 1.07722909393382 0.24714564887963 4.35868120202462 1.30848557424083e-05
2 ENSG00000100731 1162.93822827396 1.0006257630015 0.214393389946423 4.66724166846545 3.05270197242525e-06
3 ENSG00000128512 368.442571635954 1.19657846347522 0.262780839813213 4.5535224878867 5.27550292947225e-06
4 ENSG00000135535 2415.77359618136 1.22406336488047 0.183431131037356 6.67314952460929 2.5037106203736e-11
5 ENSG00000140526 796.447227235737 1.05296203760187 0.23492350092969 4.48214858639031 7.38952622958053e-06
6 ENSG00000144228 293.746859588111 1.10903132755747 0.283181091639851 3.91633255290906 8.9906210067011e-05
padj external_gene_name
<numeric> <character>
1 0.00533862114290257 PIK3C2A
2 0.002491004809499 PCNX1
3 0.00319558231320097 DOCK4
4 1.22581671973492e-07 CD164
5 0.00364483675888146 ABHD2
6 0.0214722343652725 SPOPL
description
<character>
1 phosphatidylinositol-4-phosphate 3-kinase catalytic subunit type 2 alpha [Source:HGNC Symbol;Acc:HGNC:8971]
2 pecanex 1 [Source:HGNC Symbol;Acc:HGNC:19740]
3 dedicator of cytokinesis 4 [Source:HGNC Symbol;Acc:HGNC:19192]
4 CD164 molecule [Source:HGNC Symbol;Acc:HGNC:1632]
5 abhydrolase domain containing 2 [Source:HGNC Symbol;Acc:HGNC:18717]
6 speckle type BTB/POZ protein like [Source:HGNC Symbol;Acc:HGNC:27934]
#输出含注释的差异基因文件
write.csv(diff_name,file= "diff_gene.csv")
到此为止就完成了RNA-seq的数据处理流程，下一步就是用pheatmap绘制热图了
