This repo contains demo and example code to use the NCBI Datasets REST API to grab accessions of all (reference) genomes under a certain taxonomic node, and save/retrieve/manipulate the resulting information for fun and profit.
The Snakefile provides a few different examples, including the use of the sourmash directsketch plugin to download all of the genomes in bulk.
Specifically, this repo contains code to:
- Download a dataset zip for one or more accessions;
- Retrieve genome accessions for all eukaryotic genomes.
- Create "subtracted" lists for polyphyletic taxonomic nodes such as invertebrates, non-bilateria, and "other" eukaryotes.
- Download 10 fungal genome sequences.
- Retrieve NCBI lineage information for a given taxid using pytaxonkit.
and maybe more.
To run, set your NCBI API key like so:
export NCBI_API_KEY=foobarbaz
Create a conda environment or otherwise install the things in
environment.yml:
conda env create -n ncbi-rest-api -f environment.yml
conda activate ncbi-rest-api
Then:
snakemake -p
to do some basic things.
Follow these instructions.
- https://github.com/sourmash-bio/2025-sourmash-eukaryotic-databases: Build eukaryotic databases for sourmash.
- https://github.com/sourmash-bio/2025-sourmash-ncbi-viral-databases: Build viral databases for sourmash.
I can't guarantee support for this code, of course, but odds are good that if you find a bug or need a fix it'll be useful to me and others. Please file an issue with any questions or comments! And feel free to say hi over on bluesky.
C. Titus Brown, 1/26/2025