DESeq2 after salmon

[nchut]

Hello,
I am trying to learn RNA-Seq from Biostar handbook. After we do quantification using salmon, do I need to make a tx2gene file from your GTF/FASTA to convert transcript IDs → gene IDs. I have used the NumReads column from quant.sf file. Or if I make a .csv file with NumReads, I can run DESeq with that file.
Could you please help me here. I am new to this field and I kind of lost here.
Thanks for your help in advance,

[ialbert]

The Biostar Workflows has a detailed explanation of the process of assigning informative gene names to transcripts:

https://www.biostarhandbook.com/books/workflows/rnaseq/rnaseq-using-salmon/

[nchut]

Thank you very much for your reply. Yes, I am doing this following Biostar book. But my question is it required to assign gene names to transcript before running DESeq? Can we just run salmon output as transcript in DESeq?
Thanks again for your prompt reply.

[ialbert]

I see,

you don't need to assign gene names to consolidate the transcripts as genes.

I like to do it both ways and recommend that in the book, but in reality, only transcripts exist (genes are an abstract concept).

We have to be a little more careful with transcripts only since the assignment to transcripts is a still not a completely solved problem. Salmon implements an expectation maximization algorithm, but that is an approximation. As long as the reads are not completely covering the transcripts, we are only approximating.

[nchut]

Thank you very much for your help.