transcriptome_analysis.md

Transcriptome (cDNA) Analysis

Pipeline overview

There are three different endpoints for individual parts of the cDNA analysis which should defined as steps in the config file:

1. cdna: This step only runs the read classification with pychopper and splice aware mapping
2. transcriptome: This step runs the isoform analysis with stringtie or flair or both, depending on the config option transcriptome.methods . When additionally, the qc step is defined, a detailed classification report of the transcripts is generated with SQANTI3. Additionally a comparison to the reference annotation is created with gffcompare
3. expression: Expression tables are created on multiple levels. Featurecount is used to create a gene expression table for each sample. If additionally, the step transcriptome is selected, the flair quantify module is used to create an count matrix containing all samples defined in the metadata file reads_manifest.tsv with reference to the transcriptome generated by flair for each individual sample. When the stringtie method is selected, in a first step, a consensus transcriptome is generated with stringtie merge which is then used as reference for quantification of all samples using salmon quant. The joined count table is available in isoform_counts/merged_counts_stringtie.tsv