Each row is one resistance gene × pathogen species: how many sequenced isolates of that species carry the gene, in how many countries, and the earliest collection year seen. Critical mechanisms (carbapenemase, colistin mcr, ESBL, vancomycin) are flagged and sorted first.
Source: NCBI Pathogen Detection — the open per-isolate metadata tables (AMR_genotypes, the precomputed AMRFinderPlus gene calls, plus collection_date, geo_loc_name, scientific_name). No assemblies are downloaded and AMRFinderPlus is not re-run — we ingest NCBI's own calls.
Read this honestly:
- Two organisms carrying the same gene is co-occurrence, not proof of transfer between them — horizontal gene transfer needs plasmid/phylogenetic analysis this view does not do.
- Sampling is heavily surveillance-biased: the US and Europe sequence far more isolates, so a dark country may mean more surveillance, not more resistance. Country-years with fewer than 5 sequenced isolates are greyed.
- A gene's presence is a genotype, not a phenotype — carriage does not guarantee clinical resistance.
- Scope is NCBI's foodborne + healthcare pathogen panel, not all bacteria.
Generated by Facetwork · view FFL · source repo.