Principal component analysis of MALDI TOF MS mass spectra separates M. abscessus (sensu stricto) from M. massiliense isolates

Abstract

Background: The discrimination of the members of the Mycobacterium abscessus complex is of clinical interest because one of the subspecies, M. massiliense, exhibits higher rates of response to antibiotic treatment for lung infection than do the other members of that complex. M. abscessus complex contains three subspecies that are laborious to identify; therefore, a routine diagnostic tool would be worthwhile. Results: We used principal component analysis, hierarchical cluster analysis, and single-peak analysis to examine peak lists derived from matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI TOF MS) mass spectra of 50 clinical M. abscessus complex isolates, including 28 M. abscessus (sensu stricto), 19 M. massiliense, and 3 M. bolletii isolates grown in mycobacterium growth indicator tube liquid medium and prepared with a bead-based protocol. Principal component analysis but not hierarchical cluster analysis separated M. abscessus (sensu stricto) isolates and M. massiliense isolates into two clusters. Furthermore, single-peak analysis displayed 4 discriminating peaks that separated M. abscessus (sensu stricto) from M. massiliense isolates. M. bolletii isolates did not exhibit specific peaks but resembled the M. abscessus (sensu stricto) peak profile and also grouped within this principal component analysis cluster. Principal component analysis of all peak lists with the exclusion of the four discriminating peaks again separated M. abscessus (sensu stricto) from M. massiliense isolates, thus relativizing the importance of these peaks for subspecies identification. Conclusions: Principal component analysis of peak lists derived from MALDI TOF mass spectra is a robust and convenient method of discriminating M. massiliense isolates from the other members of the M. abscessus complex.