Creating a gtf annotation file from a FASTA (& bam?) for non-model organism

Question

[sdbaney]

I performed de novo transcriptome assembly. Have the FASTA file that I am using as my reference and also performed alignments of my reads back to that. I have been using stringtie to create gtf files for the differential expression analysis (counting requires a gtf file) but I am trying to do this at gene level. Stringtie creates its own IDs for alternative splicing/isoforms. I do not want this. I can't see a way around this so is there a way I can create a gtf annotation file based on my fasta file I use for my reference?

[AK]

Try SuperTranscripts (can be constructed by Lace or utility in Trinity).

[sdbaney]

![enter image description here][1]So I ran supertranscripts (which I am using that FASTA file from it) but the gtf file it gave with it, gave really really odd entries. Here is a snapshot of it:

https://ibb.co/NC9HF89

[AK]

First, let's fix the picture: How to add images to a Biostars :-)

[sdbaney]

Did this work?

[sdbaney]

or this....