Parameterize ncbi_id in fetch_sequences

[j23414]

Description of proposed changes

Could we parameterize the NCBI Taxon ID in the fetch sequences scripts?

Historically the NCBI Taxon ID has been hardcoded in ingest/bin/genbank-url:

https://github.com/nextstrain/monkeypox/blob/b54768ec17872eb0d898e29527785642f6b98c0d/ingest/bin/genbank-url#L21

This PR generalizes that and related scripts such that the NCBI Taxon ID can be defined in a snakemake rule:

https://github.com/nextstrain/monkeypox/blob/b54768ec17872eb0d898e29527785642f6b98c0d/ingest/workflow/snakemake_rules/fetch_sequences.smk#L17-L23

View example here! 👀

Open to discussion or other ways to parameterize this. Can test the code via:

git clone https://github.com/j23414/monkeypox.git
cd monkeypox/ingest
nextstrain build  . data/sequences.ndjson

or simpler

./bin/fetch-from-genbank 10244 > genbank.ndjson

Related issue(s)

Related commit: nextstrain/dengue@5d3281d

Testing

• Checks pass

[j23414]

Speaking of NCBI datasets, parameterizing the NCBI Taxon ID feeds in nicely to the datasets command:

 f"datasets download virus genome taxon {params.ncbi_id} --no-progressbar --filename {output.dataset_package}

https://github.com/nextstrain/ncov-ingest/blob/master/workflow/snakemake_rules/fetch_sequences.smk#L53

or with Zika example:

datasets download virus genome taxon 64320 --filename zika_datasets.zip

unzip zika_datasets.zip
#> Archive:  zika_datasets.zip
#>   inflating: README.md               
#>   inflating: ncbi_dataset/data/data_report.jsonl  
#>   inflating: ncbi_dataset/data/genomic.fna  
#>   inflating: ncbi_dataset/data/virus_dataset.md  
#>   inflating: ncbi_dataset/data/dataset_catalog.json  

[tsibley]

+1 for the parameterization. I've left a number of specific comments on small things which I think are important enough to mention, but none of them I consider blocking (e.g. if for some reason you can't/won't address them).

Additionally, I think it would be good practice to put some of the rationale you included in your PR description into the commit message itself, as that's more durable and tightly linked to the source code.