一個完整的RNA-seq分析pipeline

目錄截圖

image.png

• 得到FPKM、TPM、Readscounts

cd /workspace/rnaseq/ref-only-expression
wget 
chmod +x stringtie_expression_matrix.pl

./stringtie_expression_matrix.pl --expression_metric=TPM --result_dirs='RNAseq_Norm_Lane1,RNAseq_Norm_Lane2,RNAseq_Tumor_Lane1,RNAseq_Tumor_Lane2' --transcript_matrix_file=transcript_tpm_all_samples.tsv --gene_matrix_file=gene_tpm_all_samples.tsv
./stringtie_expression_matrix.pl --expression_metric=FPKM --result_dirs='RNAseq_Norm_Lane1,RNAseq_Norm_Lane2,RNAseq_Tumor_Lane1,RNAseq_Tumor_Lane2' --transcript_matrix_file=transcript_fpkm_all_samples.tsv --gene_matrix_file=gene_fpkm_all_samples.tsv
./stringtie_expression_matrix.pl --expression_metric=Coverage --result_dirs='RNAseq_Norm_Lane1,RNAseq_Norm_Lane2,RNAseq_Tumor_Lane1,RNAseq_Tumor_Lane2' --transcript_matrix_file=transcript_coverage_all_samples.tsv --gene_matrix_file=gene_coverage_all_samples.tsv

head transcript_tpm_all_samples.tsv gene_tpm_all_samples.tsv

• Ballgown差異分析程式碼

# change working directorycd /workspace/rnaseq/# start RR# In R, run the following commandslibrary(ballgown)
library(genefilter)
library(dplyr)
library(devtools)# create the phenotype data for each sample#pheno_data = read.csv("RNA_data.csv")pheno_data <- data.frame(ids=c("RNAseq_Norm", "RNAseq_Norm_Lane1", "RNAseq_Norm_Lane2", "RNAseq_Tumor", "RNAseq_Tumor_Lane1", "RNAseq_Tumor_Lane2"), type=c("normal", "normal", "normal", "tumor", "tumor", "tumor"))# Load ballgown data structures for each samplebg = ballgown(dataDir = "ballgown", samplePattern = "RNAseq", pData=pheno_data)# Filter low-abundance genesbg_filt = subset (bg,"rowVars(texpr(bg)) > 1", genomesubset=TRUE)# Identify signficant differently expressed Transcriptsresults_transcripts = stattest(bg_filt, feature="transcript", covariate="type", getFC=TRUE, meas="FPKM")# Identify significant differently expressed Genesresults_genes = stattest(bg_filt, feature="gene", covariate="type", getFC=TRUE, meas="FPKM")# Add transcript/gene names and transcript/gene IDs to the results_transcripts data frameresults_transcripts = data.frame(transcriptNames=ballgown::transcriptNames(bg_filt),transcriptIDs=ballgown::transcriptIDs(bg_filt),geneNames=ballgown::geneNames(bg_filt),geneIDs=ballgown::geneIDs(bg_filt),results_transcripts)# Add common gene names ontmp <- unique(results_transcripts[,c("geneNames", "geneIDs")])
results_genes <- merge(results_genes, tmp, by.x=c("id"), by.y=c("geneIDs"), all.x=TRUE)# Sort from the smallest P value to largestresults_transcripts = arrange(results_transcripts,pval)
results_genes = arrange(results_genes,pval)# Output as CSVwrite.csv(results_transcripts,"RNAseq_transcript_results.csv",row.names=FALSE)
write.csv(results_genes,"RNAseq_genes_results.csv",row.names=FALSE)# Output as TSVwrite.table(results_transcripts,"RNAseq_transcript_results.tsv",sep="t", quote=FALSE, row.names=FALSE)
write.table(results_genes,"RNAseq_gene_results.tsv",sep="t", quote=FALSE, row.names=FALSE)# Identify genes with p value < 0.05subset(results_transcripts,results_transcripts$pval<0.05)
subset(results_genes,results_genes$pval<0.05)# quit Rq()

• 差異分析視覺化

# start RR# set the working directorysetwd("~/workspace/rnaseq")# load librarieslibrary(ggplot2)
library(viridis)# load in the ballgown DE datade_genes <- read.delim("RNAseq_gene_results.tsv")# load in the FPKM values from stringtieexpr_tumor <- read.delim("~/workspace/rnaseq/ballgown/RNAseq_Tumor_gene_abundance.out")# merge the expression and DE resultsmerged_results <- merge(de_genes, expr_tumor, by.x=c("id"), by.y=c("Gene.ID"), all.x=TRUE)# log2 the fold changemerged_results$log2_fc <- log2(as.numeric(merged_results$fc))# remove entries with an FPKM of 0merged_results <- merged_results[merged_results$FPKM > 1,]# create an MA plot for both genes and transcriptspdf(file="ma_plot.pdf", height=5, width=10)
ggplot(data=merged_results) + geom_point(aes(y=log2_fc, x=FPKM, color=qval)) + ylim(c(-10, 10)) + xlim(c(0, 1000)) + scale_colour_viridis(direction=-1, trans='sqrt') + theme_bw() + xlab("FPKM") + ylab("log2 Fold Change")
dev.off()

ma_plot

作者：二貨潛
連結：