Towards an interpretable machine learning model for predicting antimicrobial resistance

Abstract

This article explores the main stages of developing an interpretable machine learning (ML) model for predicting antimicrobial resistance (AMR), highlighting the importance of model interpretability in enhancing the prediction performance. By integrating phenotype-genotype synergy, our goal is to better understand AMR mechanisms. Such an approach combines ML with biological insights, offering a pathway towards more reliable AMR predictions and advancing the discovery of effective treatments against resistant pathogens. The challenges and opportunities related to incorporating this synergy into an ML model are discussed.