An alignment-free approach for eukaryotic ITS2 annotation and phylogenetic inference

Area of Science:

• Bioinformatics
• Molecular Biology
• Computational Biology

Background:

• The ITS2 gene region is crucial for eukaryotic phylogeny but its high sequence divergence complicates annotation and phylogenetic reconstruction.
• Existing alignment-based methods struggle to fully address ITS2 annotation quality and phylogenetic inference challenges.
• Alignment-free approaches offer a promising alternative for analyzing complex genetic regions like ITS2.

Purpose of the Study:

• To develop and evaluate novel alignment-free methods for ITS2 gene prediction and phylogenetic analysis.
• To assess the utility of topological indices (TIs) derived from ITS2 sequence and structure for classification.
• To compare the performance of alignment-free models against traditional alignment-based methods.

Main Methods:

• Utilized the TI2BioP methodology to calculate two types of topological indices from ITS2 sequences.
• Developed two alignment-free Artificial Neural Network models based on these topological indices.
• Isolated a novel ITS2 genomic sequence from Petrakia sp. for testing.
• Compared alignment-free models with Hidden Markov Model-based approaches.

Main Results:

• Both alignment-free models achieved over 95% accuracy in training and testing datasets.
• The novel alignment-free models demonstrated superior performance in identifying new ITS2 members compared to conventional methods.
• Phylogenetic analysis using both traditional and alignment-free techniques complemented the taxonomy of the Petrakia sp. isolate.

Conclusions:

• Alignment-free methods employing topological indices are effective for ITS2 gene prediction and classification.
• These novel approaches offer a robust alternative to alignment-based methods for phylogenetic inference, especially for highly divergent regions.
• The study successfully identified and classified a new ITS2 member, highlighting the potential of alignment-free techniques in fungal taxonomy.