Deep Learning-based Phylogenetic Analysis of Influenza Protein Sequences: A Siamese Neural Network Approach

Influenza, a highly mutable virus that can infect multiple species, has been a persistent global health challenge. The virus’s ability to recombine and evolve necessitates robust and efficient tools for phylogenetic analysis. Traditional methods, however, often rely on sequence alignment, which can be computationally intensive and less effective when dealing with large-scale sequence data. This paper introduces a new approach to phylogenetic analysis based on Siamese Neural Network (SNN). Trained on the influenza virus’s Neuraminidase (NA) protein, the proposed SNN model can accurately classify the primary host species—human, avian, and swine—and construct phylogenetic trees without requiring sequence alignment. The empirical results demonstrate the model’s robustness, achieving an accuracy of 95.2% in host species classification.