Knowledge Graph-Based Framework for Detecting Horizontal Gene Transfer Events Driving Antimicrobial Resistance
Islam, M. R.; Summers, A. O.; Arpinar, I. B.
Show abstract
Horizontal gene transfer (HGT) is a key mechanism driving the rapid dissemination of antimicrobial resistance (AMR) among bacterial populations, posing a severe threat to global public health. Traditional detection methods for HGT, such as phylogenetic incongruence and compositional analysis, are often limited in their scope and accuracy. To address these challenges, we propose a comprehensive knowledge graph (KG)-based framework that integrates genomic, environmental, biochemical, and ontological datasets to detect and analyze HGT events contributing to AMR. Our KG captures relationships among genes, plasmids, species, mobile genetic elements, antibiotics, and ecological factors using curated sources like NCBI, CARD, ICEberg, MGnify, and KEGG. We enrich nodes and edges with biological (e.g., GC content, codon bias) and chemical (e.g., binding affinities, environmental exposure) signatures and apply ontology-based semantic enrichment using ARO and GO. A Graph Neural Network (GNN) trained on the KG achieves high accuracy (92%) and F1-score (0.90) in predicting HGT events, outperforming conventional methods. Community detection identifies AMR dissemination hubs, while environmental analyses reveal strong correlations between antibiotic concentrations and resistance gene prevalence. Our framework demonstrates strong biological and chemical validation and provides a scalable, interpretable, and data-rich tool for AMR surveillance, with implications for clinical, agricultural, and environmental microbiology.
Matching journals
The top 6 journals account for 50% of the predicted probability mass.
Similar papers in this journal
Similar papers in this journal
- Over-optimism in unsupervised microbiome analysis: Insights from network learning and clustering 94%
- Learning, Visualizing and Exploring 16S rRNA Structure Using an Attention-based Deep Neural Network 94%
- Learning massive interpretable gene regulatory networks of the human brain by merging Bayesian Networks 93%
Similar papers in this journal
Similar papers in this journal
- MetaMLP: A fast word embedding based classifier to profile target gene databases in metagenomic samples 94%
- Metabolic pathway prediction using non-negative matrix factorization with improved precision 94%
- Combined topological data analysis and geometric deep learning reveal niches by the quantification of protein binding pockets 92%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.