This is the repository for the CABBAGE project, which is an acronym for:
Comprehensive Assessment of Bacterial-Based AMR prediction from GEnotypes
It aims at collecting, extracting and curating all publicly available data containing both
genotypes and matched AMR phenotypes (categorical R/I/S or quantitative values like MICs).
Unlike most existing efforts, it includes both databases as well as individual papers.
Of the former we process CDC, COMPARE-ML, NARMS, NDARO, PathogenWatch, PATRIC, and PubMLST.
Of the latter we process any paper found in a literature search with at least 100 isolates
belonging to a bacterial pathogen included in the WHO Priority Pathogen List, 2017 version.
The repository contains the R scripts used to extract the initial version of the database.
PreprocessAllPapers.R contains the URLs to the genotypes and the phenotypes in papers.
PreprocessNewPapers.R contains the URLs to the papers found in a 2nd literature search.
PreprocessDatabases.R contains the scripts required to extract and parse the databases.
PrepareTables.R contains semi-manual scripts for creating merged datatables for papers.
ExtractInfo.R focuses on extracting genotypic information from the individual databases.
ParseTables.R contains scripts used to preprocess useful tables from the AMR package.
Utilities.R contains general-purpose functions that are used by the remaining scripts.
LegacyCode.R contains full or partial functions that are not currently used (obsolete).
Replication.R is a self-contained walkthrough of the pipeline on a small representative subset of sources (two papers, the PATRIC database, and the ENA antibiogram search), intended to help reviewers reproduce key outputs from scratch.
The repository also contains the following data folders:
SelectedTables/ — processed and merged tables for papers from the first literature search.
RemainingTables/ — processed and merged tables for lower-priority papers.
NewTables/ — supplementary input files for papers from the second literature search.
AllExtractedGenotypes/ — one *_Extracted_Genotypes.csv per source, produced by getAllGenotypes.
Mappings/ — EBI-resolved accession mapping files (*_Mapped.csv) and consolidated AllMappings*.csv files.
ExtractionRecords/ — logs recording which accessions were extracted per source.
Antibiogram search/ — raw and processed files from the ENA SRA antibiogram search (BioProject PRJNA756559).