Skip to content

Latest commit

 

History

History
85 lines (57 loc) · 2.97 KB

README.md

File metadata and controls

85 lines (57 loc) · 2.97 KB
Raw

star-rsem

Summary

This pipeline is for RNASeq analysis and runs STAR, followed by RNASeQC, RSEM and optionally Cuffdiff. It should be suitable for most types of RNASeq, except small RNASeq.

Note that this pipeline is not suitable for bacterial data as its parameters are tuned towards large, spliced genomes and it requires good quality genome annotation.

Reads are aligned to given reference genome using the STAR mapper. See cfg/references.yaml for references used by default (also refer to option --references-cfg). Output of STAR includes the uniquely mapped genome bam file, transcripts mapped bam file, gene based read count matrix, bigwig files etc. (see below). For running STAR we follow recipes given here.

The transcripts/genes expression abundance are estimated by STAR and RSEM (reusing STAR's BAM file). The RSEM results matrix contains mapped reads count and TPM (normalized value) of genes and isoforms.

The pipeline also provides generic stats, coverage, mappability, QC e.g. by running RNA-SeQC.

Cuffdiff can be run optionally (slow!): it will run in Cufflinks mode, with no differential analysis carried out, to get raw fragment count of genes and isoforms in addition to Cufflinks' FPKM. The different options for the stranded argument are translated as follows (see also http://chipster.csc.fi/manual/library-type-summary.html):

  • none (default): fr-unstranded
  • reverse: fr-firststrand
  • forward: fr-secondstrand

Expression values of genes and isoforms are provided with annotation in all run methods.

Note, STAR is very memory hungry. Because its shared memory option can cause trouble when jobs fail (memory needs to be cleared manually), we do not make use of it, even in a multi-sample setting.

Output

The following lists the most important files/directories that are created in correspondingly named subfolders:

STAR

  • Mapped genome BAM: {sample}_{genome}_Aligned.sortedByCoord.out.bam
  • Mapped transcriptome BAM (RSEM input): {sample}_{genome}_Aligned.toTranscriptome.out.bam
  • Visualization: Bigwig files (*.bw)
  • Read count (genes): {sample}_{genome}_ReadsPerGene.out.tab
  • Mappability stats: {sample}_{genome}_Log.final.out

Exact STAR mapping parameters can be looked up in the Snakefile.

RSEM

  • Genes expression values with annotation: {sample}_{genome}_RSEM.genes.results.desc
  • Isoforms expression values with annotation: {sample}_{genome}_RSEM.isoforms.results.desc
  • Visualization: Wiggle files (*.wig)
  • Plots: {sample}_{genome}_RSEM.pdf

RNA-SeQC

QC and rate of rRNA and distribution of reads on transcripts:

  • countMetrics.html
  • metrics.tsv

Cuffdiff

  • Genes expression values with annotation: {sample}_{genome}_genes_FPKM_Rawreadcount_GIS.txt
  • Genes with raw fragment and fpkm value: genes.read_group_tracking