Leveraging Graph Neural Networks for MIC Prediction in Antimicrobial Resistance Studies
Student Contest:
No
Affiliation Type:
Academia
Keywords:
Graph Neural Networks, genomics, K-mer, Antibiotics, MIC
Abstract:
Antimicrobial resistance poses major healthcare challenges as organisms become resistant to antimicrobial agents. Conventional testing methods like MIC brothels are slow and labor-intensive. Machine learning offers a revolutionary approach to predict MICs and improve therapies. This paper explores using graph neural networks (GNNs) to correlate gene similarities and MICs. We introduce the K-mer GNN model, which identifies k-mer similarities and incorporates them into GNN with k-mer features. This boosts MIC prediction accuracy and provides insights into genomic drivers of resistance.
