Nucleotide host markers in the influenza A viruses

摘要

In the efforts to understand the molecular characteristics responsible for the ability of influenza viruses to cross species, various amino acid host markers in influenza viruses were uncovered. Our previous study identified a collection of novel amino acid host markers in ten proteins of 2009 pandemic H1N1. As an extension of our prior work, the objective of the current study was to employ Random Forests, a robust pattern recognition technique, to discover nucleotide host makers in the ten corresponding genes of 2009 pandemic H1N1, along with those in the genes of avian and swine viruses. Although different, there was an association between the amino acid markers in proteins and the nucleotide markers in the related genes due to codon translations. Moreover, nucleotide host markers have the capability to indicate important positions within a codon for host switches as well as the significance of synonymous mutations on host shifts, all of which amino acid markers could not provide. Our findings highlighted that two or even three nucleotide markers could coexist within a single codon, and the different importance values of these markers could further discri- minate the multiple markers within a codon. The nucleotide markers found in this study rendered a comprehensive genomic view of the complex and systemic nature of host adaptation. They verified and enriched the known amino acid markers and offered a larger set of finer host markers for further experimental confirmation.