Analysis of high-molecular-weight proteins using MALDI-TOF MS and machine learning for the differentiation of

Ana Candela1, David Rodriguez-Temporal2, Mario Blázquez-Sánchez3

  • 1Clinical Microbiology Department, Complexo Hospitalario Universitario A Coruña, Institute of Biomedical Research A Coruña (INIBIC), A Coruña, Spain. acandelagon@gmail.com.

Abstract

Insights

Matrix-assisted laser desorption/ionization-time of flight mass spectrometry (MALDI-TOF MS) can rapidly identify high-risk Clostridioides difficile ribotypes. This method offers a faster, cheaper, and less laborious alternative to current C. difficile typing techniques.

Area of Science:

  • Microbiology
  • Infectious Diseases
  • Mass Spectrometry

Background:

  • Clostridioides difficile is a primary cause of antibiotic-associated diarrhea.
  • High-risk C. difficile ribotypes (e.g., RT027, RT181, RT078) necessitate rapid identification for effective clinical management.
  • Current C. difficile ribotyping methods are time-consuming and labor-intensive.

Purpose of the Study:

  • To evaluate the potential of MALDI-TOF MS for rapid C. difficile ribotyping.
  • To analyze high-molecular-weight proteins for discriminatory C. difficile ribotype analysis.
  • To compare the performance of machine learning algorithms for C. difficile ribotyping using MALDI-TOF MS data.

Main Methods:

  • Sixty-nine C. difficile isolates representing common European ribotypes were analyzed.
  • MALDI-TOF MS was used to analyze proteins in the 17,000-65,000 m/z range.
  • Five supervised machine learning algorithms (KNN, SVM, PLS-DA, RF, GBM) were applied for classification into RT027, RT181, RT078, and 'Other RTs' categories.

Main Results:

  • All tested machine learning algorithms achieved cross-validation results greater than 70%.
  • Random Forest (RF) and Light-Gradient Boosting Machine (GBM) demonstrated the highest performance with 88% agreement.
  • RT078 was classified with 100% accuracy; RT181 and RT027 isolates were not reliably differentiated.

Conclusions:

  • MALDI-TOF MS coupled with machine learning enables rapid discrimination of key C. difficile ribotypes.
  • This approach significantly reduces the time, cost, and labor associated with conventional C. difficile typing.
  • MALDI-TOF MS offers a promising alternative for timely clinical decision-making in C. difficile infections.