We have a mild concern/observation about the file structure. When we build indices for a bunch of tools, some care about the transcripts and some don't. If we build a genome for a new GTF (say GENCODE 36 when we had 35 built), the genome.transcripts.fa that it generates is specific to the GTF, and there's nothing to indicate what it was based on. Or is this expecting to regenerate all of the references every time there's a new GTF?

So I do intend to regenerate some references every time there's a new GTF indeed

Some of the indexes won't change between versions: like BWA's, bowtie's, faidx, etc. The way things are organized now, it's just an inference that

genome.transcripts.fa

was from a given GTF. I suggest one of the following: 1. Create a source_version subdirectory below TranscriptFasta 2. Rename the

genome.transcripts.fa

to include the source_version

I think I'll go with 1 given the current structure. For me, in the long term, I don't really care about the folder structure as I think each file should have its own descriptor that contains meta data related to how when and why it was created