StringTie error generating gene and transcripts counts files

[jennaj]

Hi @tanvirai

This help will apply for you, too.

@jennaj

The instructions

@igor

A proper workflow is:

1. create gene models for every sample
2. merge individual gene models into a single “master” annotation using StringTie merge
3. and use the master annotation for read counting

The reads can be counted with StringTie or any other software, for example, featureCounts. For read counting with StringTie disable gene prediction.

Read counting should be done with the same annotation, for sake of comparison, hence, the two options were made incompatible.

More details can be found in this other Q&A

@gallardoalba

a detailed explanation of the necessary steps can be found in:

training.galaxyproject.org

Galaxy Training: De novo transcriptome reconstruction with RNA-Seq

Training material for all kinds of transcriptomics analysis.

Cheers.

---

Right now, you are calling counts against a single output from a single Stringtie run (scientific problem), and the option to call new gene models is still on (technical problem).

Hope this helps, and thanks for sending in the bug report for supporting data, very helpful!