Jane wants Get me all AMRF+ results for species-X Note from jane: would be good to be able to specify a filter here (eg high quality only) Get me all the MLST for species X
https://github.com/immem-hackathon-2025/atb-amr-shiny AMR parquet files is in https://github.com/immem-hackathon-2025/atb-amr-shiny/tree/main/data/amr_by_genus
Other ideas: Get me all the ST131 E coli Get me 100 evenly spread Salmonella Get me this genome Get me the closest genome to my genome Get me for <this set of genomes: species, MLST, list?> Return the command to download them? Get me all info on sample X Get me all genomes for species X Restrict searches by high quality, chekm2, etc etc High level stats: number of genomes, number per species etc
(out of scope?: How does my sample compare with it)
Shing’s thoughts/questions about CLI CLI commands: fetch: Fetch a parquet database from an URL (constant and version-specific). update query: Query the database by a user-specified set of filters (in a TOML file) to get a set of sample IDs/accessions (dump to CSV for further analysis?). summarise: Generate summary statistics of the samples (number of total samples, available collection dates, available geographical location, sequencing platform, etc.; customisable) in the CSV file. download: Download the sequences for a set of sample IDs/accessions in a CSV file. Pipeline to do query/summarise/download in a single command. (Limit to a max number of samples?). Version control Tied to a specific version of ATB/incremental updates and the parquet database. What to name this CLI? atb-cli Goals: Make queries reproducible. The query process should be well documented. data schema location: ~/atb/metadata/parquet. This parquet files contain metadata of the data/genome as well as ulr