transcript count after string tie merge

Question

[rthapa]

Hi, I am trying to get transcript counts with stringtie merge command. I have set of gtf output, when I checked the transcript counts of individual gtf files, I got some values of transcript counts but when I used strigtie merge command to merge all the stringtie files, I got merged output file but the transcript counts are all 0 for all the transcripts. I used the following command lines:

stringtie --merge -p 8 -o stringtie_mergedA.gtf stringtieoutputgtf.txt

where stringtieoutputgtf.txt looks like:

/scratch/user/rthapa/RNA-seq/stringtieS1Routput.gtf
/scratch/user/rthapa/RNA-seq/stringtieS2Routput.gtf
/scratch/user/rthapa/RNA-seq/stringtieS3Routput.gtf
/scratch/user/rthapa/RNA-seq/stringtieS4Routput.gtf
/scratch/user/rthapa/RNA-seq/stringtieS5Routput.gtf
/scratch/user/rthapa/RNA-seq/stringtieoutputS6R.gtf
/scratch/user/rthapa/RNA-seq/stringtieoutputS7R.gtf
/scratch/user/rthapa/RNA-seq/stringtieoutputS8R.gtf
/scratch/user/rthapa/RNA-seq/stringtieoutputS9R.gtf
/scratch/user/rthapa/RNA-seq/stringtieoutputS10R.gtf

python prepDE.py -i sample1.txt

sample1.txt looks like stringtie_mergedA.gtf /scratch/user/rthapa/RNA-seq/stringtie_mergedA.gtf

Could anyone suggest me what might have gone wrong? Thanks

Answer 1

[Kevin Blighe]

stringtie --merge is only supposed to be used to create a consensus GTF reference from multiple GTF files. The idea of doing this is that it helps with downstream analyses. For example: how could you accurately do differential expression analysis on multiple samples, each with different transcript co-ordinates in their respective GTF files? With a consensus GTF, this becomes a whole lot easier.

Once you generate the consensus GTF, you can then re-estimate abundances by re-running StringTie with the -e command-line parameter on all of your samples.

Take a look at the very helpful manual: StringTie Transcript assembly and quantification for RNA-Seq

Cross-posting: transcript count after string tie merge

Kevin

[c_u]

this was very helpful. Thank you!