Elucidating Resistance Mechanisms in Staphylococcus epidermidis: A High-Performing MALDI-TOF MS-Based Proteomic Approach for Predictive Modeling

Abstract

The emergence of Staphylococcus epidermidis as a significant nosocomial pathogen necessitates advancements in more efficient antimicrobial resistance (AMR) profiling. However, existing culture-based antimicrobial susceptibility testing (AST) methods can take up to 96 hours, while holistic PCR assays can cost thousands of dollars. This study combines machine learning with matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) to develop predictive models for various antibiotics using a comprehensive dataset containing thousands of S. epidermidis isolates. Optimized machine learning models utilized feature selection and achieved high AUROC scores ranging from 0.80 to 0.95 while maintaining AUPRC scores up to 0.96 for balanced datasets. Shapley Additive exPlanations (SHAP) were employed to analyze relevant features and assess the significance of corresponding protein biomarkers while also verifying that predictive power was derived from the detection of proteins rather than noise. AMR prediction models were validated externally to evaluate model performance outside the original data collection site. The findings in this study demonstrate that combining machine learning with AMR profiling yields significant and relevant results, indicating that such an approach is a promising solution for rapid and cost-effective treatments for nosocomial infections. Models created to predict AMR can also aid in biomarker discovery. The workflow used in the present study is potentially applicable to other microbial pathogens in the future.