RNA-seq 原始counts数据进行标准化

 RNA-seq的counts数据一般要进行标准化才能进行下游分析，常见的有FPKM与TPM两种方式，现在已知counts数据与基因的长度，求FPKM与TPM，代码如下：

rm(list = ls())
options(stringsAsFactors = F)
rawcount <- read.csv("./ALL_sample_gene_count.csv", header = T, row.names = 1)
rawcount$gene_name <- rownames(rawcount)
length <- read.table('./mouse mm20.exon.txt', sep = '\t', row.names = 1)
colnames(length) <- c('gene_name', 'length')


library(dplyr)
length <- distinct(length, gene_name, .keep_all = T)
mergecount <- merge(rawcount, length, by = 'gene_name')
FPKMlength <- mergecount[,c(1, ncol(mergecount))]
FPKMcount <- mergecount[, -c(2, ncol(mergecount))]
rownames(FPKMcount) <- FPKMcount$gene_name
FPKMcount <- FPKMcount[, -1]

## 计算 FPKM ##
a <- FPKMcount/FPKMlength$length
fpkm <- t(t(a)/colSums(FPKMcount))*(10^9)

## 计算TPM ##
tpm <- t(t(a)/colSums(a))*(10^6)

## fpkm to tpm ##
tpmfromfpkm <- t(t(fpkm)/colSums(fpkm))*(10^6)