Mycofier: a new machine learning-based classifier for fungal ITS sequences.

Area of Science:

• Fungal Taxonomy
• Bioinformatics
• Machine Learning

Background:

• Fungal ecology and diversity studies rely on accurate taxonomic and phylogenetic classification using the ITS1 genomic region.
• Current alignment-free classification tools for fungal ITS1 sequences are insufficient for large environmental surveys.

Purpose of the Study:

• To develop a machine learning-based classifier for the taxonomical assignment of fungal ITS1 sequences at the genus level.

Main Methods:

• A curated fungal ITS1 sequence database was constructed.
• A Naïve Bayesian classifier was trained using sequence features.
• The classifier achieved 87% accuracy in genus-level classification.

Main Results:

• The developed classifier, Mycofier, utilizes a Naïve Bayes algorithm and ITS1 sequences from 510 fungal genera.
• Mycofier demonstrates comparable accuracy to BLASTN but is more efficient due to its alignment-independent nature.

Conclusions:

• Mycofier addresses the need for an accurate, efficient, alignment-free classification tool for large fungal ITS1 datasets.
• The software and source code are publicly available for broader research application.