Shadow Antibiogram: co-testing network data, Germany 2019–2023

Published: 14 July 2026| Version 1 | DOI: 10.17632/8ffg3jb2vt.1
Contributor:
Ebenezer Awotoro

Description

This dataset contains processed, non-identifiable data and reproducibility materials for the manuscript "Network-based characterisation of latent co-testing patterns in antimicrobial susceptibility testing." Deposited materials include pairwise co-testing contingency tables, antibiotic metadata, pathogen cohort definitions, figure source data, analysis configuration files, and source code. The contingency tables capture joint observation of antimicrobial susceptibility results across antibiotic pairs, stratified by pathogen, specimen type, year, care setting, and ward type. These files support reconstruction of Jaccard, Dice, Cosine, and phi similarity matrices; FDR-filtered co-testing networks; Louvain community detection results; and all manuscript figures. Raw individual-level ARS surveillance records are excluded under data protection, surveillance governance, and data provider agreements. The deposited files contain no patient identifiers, laboratory identifiers, exact dates, free-text fields, or individual-level records.

Files

Steps to reproduce

Prerequisites: Python 3.x must be available on your system. Verify with python3 --version. The pipeline runs on Linux, macOS, and HPC environments (SLURM-compatible). Step 1: Clone the code repository: > git clone https://github.com/Ebenco36/shadow-antibiogram.git Step 2: > cd shadow-antibiogram Step 3: Copy the deposited WHO_Aware_data folder into the datasets/ directory of the cloned repository, so the path reads datasets/WHO_Aware_data/. Run the full analysis pipeline: > bash run.sh This automatically creates a virtual environment, installs all dependencies from requirements.txt, and executes the complete analysis. No manual dependency installation is required. Outputs are written to the outputs/ directory, including similarity matrices, FDR-filtered co-testing networks, Louvain community assignments, and all manuscript figures. Figure outputs are located in outputs/use_cases/.

Institutions

Categories

Computer Science, Medicine, Microbiology

Licence