Abundance Counting for Non-Model Organism for edgeR

Question

[sdbaney]

I am working with a non-model organism (no annotated genome or the option to use an annotation file in my abundance counting). I performed de novo transcriptome assembly of the Illumina HiSeq4000 reads.

I originally used StringTie as my abundance counter because I didn't need a gtf annotation file and it created one I could use for the differential expression analysis but I am now seeing that because of its feature to predict isoforms, I am having trouble matching up their IDs with my assembly IDs (there being more stringtie IDs than assembly because of its novel prediction).

Is there any way to disable this feature in StringTie? Is there another program I can use to create a gtf file so that I can use featureCounts or just count abundances that will not require a gtf?

Answer 1

[ATpoint]

You can use salmon to quantify reads against a transcriptome in fasta format. From there on you can summarize transcript count to the gene level with tximport.

[sdbaney]

Thank you, I think that’s what I’m going to end up doing. As stringtie seems to be causing more headache than anticipated.