HyDRA - Identifying recently acquired antibiotic resistance genes using machine learning

The spread of antibiotic resistance poses a significant threat to humanity [1]. Rapid and accurate detection of antibiotic resistance is required to determine an appropriate antibiotic therapy. Current methods are solely based on cultivation and therefore only test a limited number of antibiotics.

Our aim is to produce a detailed resistogram that will support the physicians to find a suitable treatment and render therapy more effective.

For this purpose we are developing a fully automated Snakemake [2] workflow for reproducible genome analysis, called HyDRA - Hybrid De novo assembly for Resistance Analysis. This analysis is based on a combination of short and long reads generated by Illumina and Oxford Nanopore sequencing of bacterial isolates. These reads are used for a hybrid genome assembly, which then is screened for known antibiotic resistance genes (ARGs) using a well curated ARG specific database. In order to additionally identify newly acquired and hitherto unknown ARGs we are currently working on a machine learning classifier to reliably detect recently horizontally transferred genes.

The best classifier so far was trained on E. coli K12 genome data from HGT-DB [3] and achieved credible results for different E. coli test data. However, this model does not work well for other tested organisms (see Fig.1). We are working on improving this classifier and creating more classifiers specialized on other bacteria. Furthermore, we are extending the workflow to metagenomics sequencing data to work directly with patient samples and skip the time-consuming cultivation.

Conclusively, this workflow will rapidly provide the physicians detailed information to select an appropriate antibiotic treatment. This should prevent the unnecessary use of broad-spectrum antibiotics and thus help fight the emergence of antibiotic resistance.

[1] ECDC "Antimicrobial resistance in the EU/EEA (EARS-Net) - Annual epidemiological report for 2022", 2023

[2] J. Koster and S. Rahmann, Bioinformatics, 2012, doi: 10.1093/bioinformatics/bts480

[3] S. Garcia-Vallve et al., Nucleic Acids Res, 2003, doi: 10.1093/nar/gkg004