Abstract
<jats:p>Accurate species-level identification of Aspergillus is clinically important because closely related species may differ in virulence and antifungal susceptibility. This study evaluated the discriminatory capacities of the ITS, β-tubulin (benA), and calmodulin (CaM) loci using an in silico multilocus approach. Whole-genome sequences from the NCBI GenBank and RefSeq databases were used to construct an in-house reference library. Barcode regions were extracted in silico using published primer pairs and analyzed for sequence length variation, nucleotide differences, single nucleotide polymorphisms, insertions/deletions, and candidate species reduction patterns. Principal coordinate analysis (PCoA) was performed to visualize genetic relationships among species. Most clinically important species exhibited similar barcode length distributions and clustered according to Aspergillus sections. PCoA results were largely consistent with section classification, with the first two coordinates explaining 20.4% and 18.2% of the total variation. Sequential analysis using a ≥99% sequence identity threshold showed that ITS had the lowest discriminatory capacity, whereas benA achieved the greatest reduction in candidate species and resolved the highest number of taxa. CaM displayed the highest sequence variation in pairwise comparisons of closely related species. Although A. flavus and A. austwickii remained indistinguishable across all three loci, additional differences were detected in the cyp51A region. These findings indicate that section-oriented marker selection and prior in silico evaluation can improve the efficiency and discriminatory performance of sequence-based Aspergillus identification workflows.</jats:p>