Accurate prediction of antimicrobial resistance and genetic marker of Staphylococcus aureus clinical isolates using MALDI-TOF MS and machine learning - Across DRIAMS and Taiwan database.

ABSTRACT

BACKGROUND: The use of matrix-assisted laser desorption/ionization-time-of-flight mass spectra (MALDI-TOF MS) with machine learning (ML) has been explored for predicting antimicrobial resistance. This study evaluates the effectiveness of MALDI-TOF MS paired with various ML classifiers and establishes optimal models for predicting antimicrobial resistance and mecA gene existence among Staphylococcus aureus. MATERIALS AND METHODS: The antimicrobial resistance against tier 1 antibiotics and MALDI-TOF MS of S. aureus were analyzed using data from the Database of Resistance against Antimicrobials with MALDI-TOF Mass Spectrometry (DRIAMS) and one medical center (CS database). Five ML classifiers were used to analyze performance metrics. The Shapley value quantified the predictive contribution of individual feature. RESULTS: The LightGBM demonstrated superior performance in predicting antimicrobial resistance for most tier 1 antibiotics among oxacillin-resistant S. aureus (ORSA) than all and oxacillin-susceptible S. aureus (OSSA) in both databases. In DRIAMS, MLP encompassed excellent predictive performance, expressed as accuracy/AUROC/AUPR, for clindamycin (0.74/0.81/0.90), tetracycline (0.86/0.87/0.94), and trimethoprim-sulfamethoxazole (0.95/0.72/0.97). In CS database, Ada and LightGBM retained excellent performance for erythromycin (0.97/0.92/0.86) and tetracycline (0.68/0.79/0.86), respectively. Mass-to-charge ratio (m/z) features of 2,411-2,414 and 2,429-2,432 correlated with clindamycin resistance, while 5,033-5,036 was linked to erythromycin resistance in DRIAMS. In CS database, overlapping features of 2,423-2,426, 4,496-4,499, and 3,764-3,767 simultaneously predicted mecA existence and oxacillin resistance. CONCLUSION: The predictive performance of antimicrobial resistance against S. aureus using MALDI-TOF MS depends on database characteristics and ML algorithm selected. Specific and overlapping MS features are excellent predictive markers for mecA and specific antimicrobial resistance.