跟着Nature Communications学数据分析：R语言DESeq2包做otu差异丰度分析及火山图展示结果

论文

Microbiome differential abundance methods produce different results across 38 datasets

数据链接

https://figshare.com/articles/dataset/16S_rRNA_Microbiome_Datasets/14531724

代码链接

https://github.com/nearinj/Comparison_of_DA_microbiome_methods

这个人的github主页还有其他论文的数据和代码

https://github.com/jnmacdonald/differential-abundance-analysis 这个链接有很多冠以差异丰度分析的代码

这两天在看宏基因组的利用otu丰度数据做差异丰度分析，找到了这篇论文，看了摘要，好像是比较了不同差异丰度分析方法获得结果的异同。重复一下这里利用DESeq2做差异丰度分析的代码

这里我用到的数据集是

丰度数据 ArcticFireSoils_genus_table.tsv
分组数据 ArcticFireSoils_meta.tsv

这里有一个疑问：论文提供的丰度表格数据有两个，还有一个是带rare后缀的，暂时不知道这两个有啥区别

首先是读取数据集

ASV_table <- read.table("metagenomics/dat01/ArcticFireSoils_genus_table.tsv", 
                        sep="\t", 
                        skip=1, 
                        header=T, 
                        row.names = 1,
                        comment.char = "", 
                        quote="", check.names = F)
groupings <- read.table("metagenomics/dat01/ArcticFireSoils_meta.tsv", 
                        sep="\t", 
                        row.names = 1, 
                        header=T, 
                        comment.char = "", 
                        quote="", 
                        check.names = F)
dim(ASV_table)
dim(groupings)
groupings$Fire<-factor(groupings$Fire)

这里需要对表示分组的赋予因子，要不然后面deseq2的步骤会有警告信息

判断一下丰度数据的样本名和分组数据的样本名顺序是否一致

identical(colnames(ASV_table), rownames(groupings))

返回false不一致

对两个样本名取交集

rows_to_keep <- intersect(colnames(ASV_table), rownames(groupings))

根据取交集的结果重新选择样本

groupings <- groupings[rows_to_keep,,drop=F]
ASV_table <- ASV_table[,rows_to_keep]

疑问：这里中括号里的drop参数是啥作用了？

再次判断两个数据集中样本名的顺序

identical(colnames(ASV_table), rownames(groupings))

这次返回TRUE

对分组文件的列名进行修改

colnames(groupings)[1] <- "Groupings"

差异丰度分析

library(DESeq2)
dds <- DESeq2::DESeqDataSetFromMatrix(countData = ASV_table,
                                      colData=groupings,
                                      design = ~ Groupings)
dds_res <- DESeq2::DESeq(dds, sfType = "poscounts")

res <- results(dds_res, 
               tidy=T, 
               format="DataFrame",
               contrast = c("Groupings","Fire","Control"))
head(res)

image.png

火山图代码

DEG<-res
logFC_cutoff<-2
DEG$change<-as.factor(ifelse(DEG$pvalue<0.05&abs(DEG$log2FoldChange)>logFC_cutoff,
                             ifelse(DEG$log2FoldChange>logFC_cutoff,"UP","DOWN"),
                             "NOT"))
this_title <- paste0('Cutoff for logFC is ',round(logFC_cutoff,3),
                     '\nThe number of up gene is ',nrow(DEG[DEG$change =='UP',]) ,
                     '\nThe number of down gene is ',nrow(DEG[DEG$change =='DOWN',]))
DEG<-na.omit(DEG)
library(ggplot2)
ggplot(data=DEG,aes(x=log2FoldChange,
                    y=-log10(pvalue),
                    color=change))+
  geom_point(alpha=0.8,size=3)+
  labs(x="log2 fold change")+ ylab("-log10 pvalue")+
  ggtitle(this_title)+theme_bw(base_size = 20)+
  theme(plot.title = element_text(size=15,hjust=0.5),)+
  scale_color_manual(values=c('#a121f0','#bebebe','#ffad21')) -> p1
p1+xlim(NA,10)+ylim(NA,30) -> p2

library(patchwork)
p1+p2