CryptoType - Public Datasets for MALDI-TOF-MS Based Differentiation of Cryptococcus neoformans/gattii Complexes

Yeasts of the Cryptococcus neoformans/gattii species complexes are human pathogens mostly in immune compromised individuals, and can cause infections from dermal lesions to fungal meningitis. Differences in virulence and antifungal drug susceptibility of species in these complexes indicate the value of full differentiation to species level in diagnostic procedures. MALDI-TOF MS has been reported to sufficiently discriminate these species. Here, we sought to re-evaluate sample pre-processing procedures and create a set of publicly available references for use with the MALDI Biotyper system. Peak content using four different pre-processing protocols was assessed, and database entries for 13 reference strains created. These were evaluated against a collection of 153 clinical isolates, typed by conventional means. The use of decapsulating protocols or mechanical disruption did not sufficiently increase the information content to justify the extra hands-on-time. Using the set of 13 reference entries created with the standard formic acid extraction, we were able to correctly classify 143/153 (93.5%) of our test isolates. The majority of the remaining ten isolates still gave correct top matches; only two isolates did not give reproducible identifications. This indicates that the log score cut-off can be lowered also in this context. Ease to identify cryptococcal isolates to the species level is improved by the workflow evaluated here. The database references are freely available from https://github.com/oliverbader/BioTyper-libraries for incorporation into local diagnostic systems.

Automated summary

This study re-evaluated sample pre-processing procedures for yeasts of the Cryptococcus neoformans/gattii species complexes and created a set of publicly available references for use with the MALDI Biotyper system. By assessing peak content using four different pre-processing protocols and creating 13 reference entries with standard formic acid extraction, correct classification of 93.5% of test isolates was achieved. The freely available database references can improve species-level identification of cryptococcal isolates.