A novel bioinformatics strategy to predict directional changes of influenza A virus genome sequences

Abstract

Influenza A viruses cause a significant threat to public health as highlighted by the recent introduction of the swine-derived H1N1 virus (pandemic H1N1/09) into human populations. Pandemics were primarily initiated by introduction from animal sources and successive adaptation among humans through human-to-human transmission. We established a sequence alignment-free clustering method "BLSOM", which can analyze and compare all influenza A virus genome sequences on one map. Separation according to host animal, subtype and epidemic year could be efficiently visualized. Notably, H1N1/09 strains have oligonucleotide and codon compositions clearly distinct from those of seasonal human flu strains. This enabled us to make inferences about directional changes of H1N1/09 sequences in the near future and to list codons and oligonucleotides with the potential of reduction in H1N1/09 sequences. The strong visualization power of BLSOM also provides surveillance strategies for efficiently detecting potential precursors to pandemic viruses. © 2011 Springer-Verlag Berlin Heidelberg.