Comparative Analysis of H1N1 Avian Influenza Virus by Multiple Sequence Alignment and Support Vector Machine

The H1N1 is a subtype of avian influenza virus (AIV) that is able to break the host barrier to seriously endanger human health. Investigating the molecular mechanisms of AIV interspecies transmission is important for preventing the influenza epidemics. In this study, we used bioinformatics approaches to identify factors that may cause the avian-to-human transmission in hemagglutinin (HA) sequences of H1N1. First, the multiple sequence alignment analysis reveals 10 signature regions of HA that are highly conserved in intra-species, but largely divergent between interspecies. Then, the avian-to-human transformation was modeled as a binary classification problem in a machine learning (ML) context. A computational prediction model was developed to predict the avian-to-human transmission of H1N1 with advanced ML techniques by characterizing amino acid residues in these signatures regions. The evaluation results suggested that these amino acid residues have a discrimination ability to distinguish H1N1 strains isolated from human to those from avian. The proposed bioinformatic framework would be helpful for further understanding the transmission mechanisms of H1N1 and other AIV viruses.