Integrating Machine Learning and Molecular Methods for Trichophyton indotineae Identification and Resistance...

Area of Science:

• Medical Mycology
• Clinical Microbiology
• Computational Biology

Background:

• Trichophyton indotineae causes difficult-to-treat, terbinafine-resistant fungal infections.
• Accurate identification and resistance profiling are crucial for effective treatment.

Purpose of the Study:

• To improve the identification and terbinafine resistance profiling of T. indotineae.
• To integrate molecular methods with machine learning for MALDI-TOF MS analysis.

Main Methods:

• Analyzed 56 clinical isolates using ITS and ERG1 gene sequencing, antifungal susceptibility testing, and MALDI-TOF MS.
• Applied unsupervised (PCA, hierarchical clustering) and supervised (PLS-DA, SVM) machine learning algorithms.
• Identified discriminatory spectral peaks using biomarker analysis.

Main Results:

• Terbinafine resistance was found in 23 isolates, linked to ERG1 mutations.
• Machine learning algorithms achieved 100% accuracy in classifying T. indotineae.
• Distinct spectral peaks were identified for T. indotineae and T. mentagrophytes.

Conclusions:

• Combining MALDI-TOF MS with machine learning enhances diagnostic accuracy for T. indotineae.
• This approach offers a practical alternative to sequencing, especially in resource-limited settings.
• Improved detection of resistant strains can guide targeted antifungal therapy.