MSI-3, an improved version of the MSI application that includes both fungal and bacterial identifications for MALDI-ToF spectra

Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-ToF MS) is now central to routine diagnostics, enabling accurate identification of microbial species. We assessed the performance of MSI-3, the updated release of the free online mass spectrometry identification (MSI) platform that now supports bacterial, in addition to fungal, identification. MSI-3 was benchmarked against MSI-2 and the Bruker MALDI Biotyper using two fungal data sets: (i) a reference panel of 669 sequenced isolates (Panel 1), and (ii) a real-world set of user-submitted spectra (Panel 2). MSI-3 and MSI-2 demonstrated equivalent performance across all fungal groups in Panel 1, with 97.8% (mean of 892/912) concordance at the species or species complex level, and a massive added value compared to the Bruker database for filamentous fungi (from about 41% to about 98%). MSI-3 exhibited slightly longer identification times than MSI-2 (1.52 vs 0.84 s per spectrum), but remained compatible with real-time online usage. The increased identification time is offset by a twofold expansion of the reference database and algorithmic improvements. In Panel 2, MSI-3 achieved comparable performance to Bruker for bacterial and yeast spectra, and outperformed Bruker for filamentous fungi, especially Aspergillus spp., Fusarium spp., and dermatophytes. A new “uninterpretable” category was introduced in MSI-3 to flag potential mixed or poor-quality spectra, facilitating result interpretation. Overall, MSI-3 marks a substantial advance in the MSI platform, uniting flexibility, broad taxonomic coverage, and artificial intelligence (AI) readiness. Its combined use with Bruker enhances diagnostic performance, positioning MSI-3 as a powerful and scalable standard for microbial identification in clinical and research settings.