dWe developed an Australian database for the identification of Aspergillus, Scedosporium, and Fusarium species (n ؍ 28) by matrix-assisted laser desorption ionization؊time of flight mass spectrometry (MALDI-TOF MS). In a challenge against 117 isolates, species identification significantly improved when the in-house-built database was combined with the Bruker Filamentous Fungi Library compared with that for the Bruker library alone (Aspergillus, 93% versus 69%; Fusarium, 84% versus 42%; and Scedosporium, 94% versus 18%, respectively).
Rapid, accurate mold identification is important due to the widening spectrum of pathogens and species-specific differences in antifungal susceptibility (1-3). Matrix-assisted laser desorption ionizationϪtime of flight mass spectrometry (MALDI-TOF MS) has proven useful, but mold identification remains challenged by the limited access to validated purpose-built databases that are necessary because of small species and strain representations in commercial libraries (4-16).Given the prior poor performance of the Bruker Filamentous Fungi Library v1.0 (Bruker Daltonics, Bremen, Germany) for mold identification using the manufacturer-recommended broth-based protein extraction methods (in our laboratory Ͼ50% of isolates were not identified; internal data) and because the geographic generalizability of in-house-built databases is not yet known, we hypothesized that a MS library of molds relevant to our region (17-21) will improve identification. Here, we constructed an in-house database containing 117 strains (see Table S1 in the supplemental material) covering 28 species of Aspergillus, Scedosporium, and Fusarium encountered in Australia. Challenge isolates (also n ϭ 117; 21 species) comprising 55 Aspergillus, 45 Fusarium, and 17 Scedosporium clinical strains (Table 1) were then used to assess the performance of the Bruker library alone versus that of the Bruker library supplemented with the in-house library for species identification.All isolates were identified using phenotypic methods (22) with definitive identification by DNA sequencing of the internal transcribed spacer (ITS) (all isolates), -tubulin (Aspergillus and Scedosporium spp.), and partial elongation factor-1alpha (EF-1␣) (to identify Fusarium to the species complex level) gene regions (23)(24)(25)(26). Sequence data were analyzed against the Centraalbureau voor Schimmelkultures (http://www.cbs.knaw.nl/Collections /BioloMICSSequences.aspx?fileϭall), International Society for Human and Animal Mycology ITS (http://its.mycologylab.org/), and Fusarium-ID (http://www.fusariumdb.org/index.php) databases, and species were assigned using published criteria (27).Protein extraction for MALDI-TOF MS was performed as previously described (11). The Bruker bacterial test standard (Bruker Daltonics) was used for calibration and Aspergillus ustus CBS 261.67T scoring of Ն2.00 was required for quality extraction and spectra acceptability (11). The in-house database was constructed using published protocols (11, 28) with 20 to 25 quality spe...