Host markers and correlated mutations in the overlapping genes of influenza viruses: M1, M2; NS1, NS2; and PB1, PB1-F2

Abstract

The influenza A viruses have three gene segments, M, NS, and PB1, which code for more than one protein. The overlapping genes from the same segment entail their interdependence, which could be reflected in the evolutionary constraints, host distinction, and co-mutations of influenza. Most previous studies of overlapping genes focused on their unique evolutionary constraints, and very little was achieved to assess the potential impact of the overlap on other biological aspects of influenza. In this study, our aim was to explore the mutual dependence in host differentiation and co-mutations in M, NS, and PB1 of avian, human, 2009 H1N1, and swine viruses, with Random Forests, information entropy, and mutual information. The host markers and highly co-mutated individual sites and site pairs (P values < 0.035) in the three gene segments were identified with their relative significance between the overlapping genes calculated. Further, Random Forests predicted that among the three stop codons in the current PB1-F2 gene of 2009 H1N1, the significance of a mutation at these sites for host differentiation was, in order from most to least, that at 12, 58, and 88, i.e., the closer to the start of the gene the more important the mutation was. Finally, our sequence analysis surprisingly revealed that the full-length PB1-F2, if the three stop codons were all mutated, would function more as a swine protein than a human protein, although the PB1 of 2009 H1N1 was derived from human H3N2.