Provide pre-built genedb (.db) files for common annotations

[amarinderthind]

Feature request
Would it be possible to host pre-built .db files for commonly used annotations (e.g., GENCODE v49 for GRCh38, RefSeq for T2T-CHM13v2.0)?

Motivation
When integrating IsoQuant into automated pipelines (e.g., Nextflow/Snakemake), the --complete_genedb GTF-to-database conversion adds ~16 minutes per sample. Providing a pre-built --genedb avoids this, but:

The .db files are not hosted anywhere — users must build them locally
The correct gffutils.create_db() parameters to match IsoQuant's expectations making it difficult to build them independently outside of IsoQuant
A standalone --build_db_only mode (or similar) that builds the .db and exits without requiring input reads would also solve this

Suggested options (any would help)
Host pre-built .db files for common annotations (GENCODE, RefSeq) alongside IsoQuant releases
Document the exact gffutils.create_db() parameters IsoQuant uses, so users can build them externally
Add a --build_db_only CLI mode that creates the .db from a GTF and exits

[andrewprzh]

Dear @amarinderthind

Thanks for the suggestion!

Would it be possible to host pre-built .db files for commonly used annotations (e.g., GENCODE v49 for GRCh38, RefSeq for T2T-CHM13v2.0)?

As there are quite a lot of versions of human and mouse annotations (Gencode, Ensemle etc), and new releases come out rather frequently, I don't think storing all kinds .db files somewhere makes a lot sense.

When integrating IsoQuant into automated pipelines (e.g., Nextflow/Snakemake), the --complete_genedb GTF-to-database conversion adds ~16 minutes per sample.

Once the GTF is converted, IsoQuant stores paths to converted .db files in a global config. Thus, when running IsoQuant again with the same GTF, no repeated conversion will take place (unless .db file is moved or deleted of course).

A standalone --build_db_only mode (or similar) that builds the .db and exits without requiring input reads would also solve this

There is a simple script isoquant_lib/gtf2db.py which does exactly this. It has the same --complete_genedb option as IsoQuant itself. I will add this to the documentation.

Best
Andrey

[amarinderthind]

@andrewprzh Thank you very much for your response and clarification. @calizilla FYI

[calizilla]

Thanks both. Andrey we will add this isoquant_lib/gtf2db.py step to our nextflow workflow in the next round op updates. Thanks for your fast reply!