Rapid and accurate discrimination of Mycobacterium abscessus subspecies based on matrix-assisted laser desorption ionization-time of flight spectrum and machine learning algorithms

Abstract

Mycobacterium abscessus complex (MABC) has been reported to cause considerable complicated infections. Subspecies identification of MABC is crucial for adequate treatment due to different antimicrobial resistance property amid the subspecies. However, long incubation days is needed for the traditional antibiotic susceptibility testing (AST) method. Effective antibiotics administration often delayed considerably and caused unfavorable outcomes. Thus, we proposed a novel and accurate method to identify subspecies and its potential antibiotics resistance, to guide clinical treatment within hours. Subspecies of the MABC isolates were determined by secA1, rpoB, and hsp65. AST was tested by using microdilution method, as well as sequencing of erm(41) and rrl genes. MALDI-TOF mass spectrometry (MS) spectra were analyzed. The informative peaks on MS spectra were detected by random forest (RF) importance. Machine learning (ML) algorithms were used to build models for classifying MABC subspecies based on MALDI-TOF spectrum. The models were developed and validated by nested five-fold cross-validation to avoid over-fitting. In total, 102 MABC isolates (52 subspecies abscessus and 50 subspecies massiliense) were analyzed. Top informative peaks including m/z 6715, 4739, 2805, etc. were identified. RF model attained AUROC of 0.9166 (95% CI: 0.9072-0.9196) and outperformed other algorithms in discriminating subspecies abscessus from massiliense. We developed a MALDI-TOF based ML model for rapid and accurate MABC subspecies identification. The novel diagnostic tool would guide a more accurate and timely MABC subspecies-specific treatment.