MDRSA

MDRSA identifies multidrug-resistant Staphylococcus aureus from MALDI-TOF MS spectra to predict susceptibility to oxacillin, clindamycin, and erythromycin.

Key Features:

• Rapid Identification: Provides rapid predictions of Staphylococcus aureus resistance to oxacillin, clindamycin, and erythromycin from MALDI-TOF MS data.
• MALDI-TOF MS Spectral Processing: Processes MALDI-TOF MS spectra to detect mass spectral peaks associated with antibiotic resistance.
• Advanced Data Analysis (KDE): Applies kernel density estimation (KDE) to mitigate peak shifting in mass spectra.
• Machine Learning Integration: Trains classifiers including decision trees, random forests, and support vector machines on processed spectra to classify resistance status.
• Validation and Performance: Reports area under the receiver operating characteristic curve (AUC) of 0.8 in internal 10-fold cross-validation and in independent external testing.

Scientific Applications:

• Clinical microbiology screening: Enables rapid antimicrobial susceptibility screening of Staphylococcus aureus in clinical microbiology laboratories.
• Antibiotic therapy guidance: Supports selection of appropriate antibiotic therapies for Staphylococcus aureus infections by predicting resistance to oxacillin, clindamycin, and erythromycin.

Methodology:

Processes MALDI-TOF MS spectra to identify resistance-associated mass spectral peaks, uses kernel density estimation (KDE) to address peak shifting, trains decision tree, random forest, and support vector machine classifiers on processed spectra, and evaluates performance via 10-fold cross-validation and independent external testing (AUC = 0.8).