Next-generation sequencing (NGS) has become a powerful and efficient tool for surveying mycorrhizal mycobiome diversity, surpassing classical methods in accuracy and throughput. Long-read NGS techniques are increasingly applied under the assumption that they provide better taxonomic resolution, yet their use often lacks a balanced evaluation against the established strengths and limitations of widely used short-read NGS technologies. This study compares Illumina MiSeq and PacBio Sequel platforms in analyzing the mycorrhizal mycobiome of Pinus densiflora roots, focusing on how sequencing platforms and database choice influence taxonomic resolution and diversity patterns. Both platforms detected mycorrhizal taxa with similar taxonomic resolution, recovering nearly all taxa previously reported from pine roots. Most mycorrhizal taxa were shared between datasets, although several taxa were detected exclusively by one platform. In terms of diversity, the short-read dataset showed higher diversity due to greater sequencing depth, whereas the long-read dataset offered improved identification of rare or closely related taxa owing to longer sequence information. Moreover, supplementing reference databases with locally derived sequences enhanced taxonomic resolution and the detection of native taxa in both approaches, with a stronger effect for the long-read dataset. Overall, our results emphasize that short- and long-read sequencing each have distinct advantages for mycorrhizal community analysis, and that the use of curated local reference databases is essential to maximize taxonomic resolution and improve the detection of regionally unique taxa.

The increase of sequence data in public nucleotide databases has made DNA sequence-based identification an indispensable tool for fungal identification. However, the large proportion of mislabeled sequence data in public databases leads to frequent misidentifications. Inaccurate identification is causing severe problems, especially for industrial and clinical fungi, and edible mushrooms. Existing species identification pipelines require separate validation of a dataset obtained from public databases containing mislabeled taxonomic identifications. To address this issue, we developed FunVIP, a fully automated phylogeny-based fungal validation and identification pipeline (https://github.com/Changwanseo/FunVIP). FunVIP employs phylogeny-based identification with validation, where the result is achievable only with a query, database, and a single command. FunVIP command comprises nine steps within a workflow: input management, sequence-set organization, alignment, trimming, concatenation, model selection, tree inference, tree interpretation, and report generation. Users may acquire identification results, phylogenetic tree evidence, and reports of conflicts and issues detected in multiple checkpoints during the analysis. The conflicting sample validation performance of FunVIP was demonstrated by re-iterating the manual revision of a fungal genus with a database with mislabeled sequences, Fuscoporia. We also compared the identification performance of FunVIP with BLAST and q2-feature-classifier with two mass double-revised fungal datasets, Sanghuangporus and Aspergillus section Terrei. Therefore, with its automatic validation ability and high identification performance, FunVIP proves to be a highly promising tool for achieving easy and accurate fungal identification.