Metabarcoding Characterization of Fungal Communities in Spanish Cereals with a Special Focus on Fusarium Species.

Abstract

Cereal safety is a critical public health concern, as crops are frequently compromised by mycotoxin-producing fungi, particularly those from the genus Fusarium. Although metabarcoding is a powerful tool for characterizing these fungal communities, the universal internal transcribed spacer 2 (ITS2) marker lacks the resolution required to distinguish closely related species with distinct toxigenic profiles. The objective of this work was to accurately characterize the mycobiota and specific Fusarium community structure in Spanish cereal field samples. We propose a high-resolution metabarcoding workflow within the QIIME 2 environment to optimize Fusarium identification using the translation elongation factor 1-α (TEF1) gene. We developed a TEF1 Naive Bayes classifier trained on a curated database derived from FUSARIUM-ID v.3.0, expanded with eukaryotic sequences to prevent false-positive assignments. This approach was validated through direct comparison with ITS2 in the same samples. As expected, ITS2 significantly underestimates Fusarium diversity, whereas TEF1 enables precise species-level resolution within this genus. General mycobiota analysis performed using ITS2 revealed that fungal communities are shaped primarily by geographical location rather than host cereal species. Crucially, our approach confirmed the persistence of key toxigenic species, such as F. langsethiae and F. graminearum, and revealed broader diversity through the consistent detection of species often overlooked by traditional methods, including F. equiseti, F. acuminatum, and F. culmorum. We conclude that our metabarcoding framework reveals a high Fusarium diversity in Spanish cereal grains, and this knowledge is essential for designing targeted strategies to predict and mitigate mycotoxin contamination in these crops.