How to create a GTF/GFF file for HTSeq OR how to count reads without GTF?

[Biol]

I am working with RNAseq of an insect under different treatments. I already have the data (Illumina HiSeq paired-end reads) and I was planning to follow the protocol from Anders et al. (2013) to analyze it (I already did it in the past with a model organism). However, the genome of this insect was recently sequenced, so it still has no GTF file available. The only files available are: the genome.fasta and the CDS.fasta.

So far, I mapped the reads on the genome using tophat (without providing a GTF file) which provided me bam/sam files. Now, I should count the reads per gene using HTSeq. However, the input for HTSeq is a GTF file that I dont have.

How can I count reads per gene without a GTF file? Is it possible to use the genome and the CDS to create a GTF/GFF file for HTSeq?

Any other idea on how to proceed would be helpful.
Thank you so much!

[Persistent LABS]

Hi Biol,
If you want to analyze your data using the pipeline HTSeq->DESeq2, you will require GTF file, and in your case probably you can write a script to create one based upon the alignment of CDS.fasta to your genome.fasta.
But if you are willing to use other pipeline, you can use TopHat->Cufflinks->Cuffmerge->Cuffdiff. Here, the reference annotation is not mandatory.