A complete analysis of HA and NA genes of influenza A viruses

Abstract

Background: More and more nucleotide sequences of type A influenza virus are available in public databases. Although these sequences have been the focus of many molecular epidemiological and phylogenetic analyses, most studies only deal with a few representative sequences. In this paper, we present a complete analysis of all Haemagglutinin (HA) and Neuraminidase (NA) gene sequences available to allow large scale analyses of the evolution and epidemiology of type A influenza.

Methodology/principal findings: This paper describes an analysis and complete classification of all HA and NA gene sequences available in public databases using multivariate and phylogenetic methods.

Conclusions/significance: We analyzed 18,975 HA sequences and divided them into 280 subgroups according to multivariate and phylogenetic analyses. Similarly, we divided 11,362 NA sequences into 202 subgroups. Compared to previous analyses, this work is more detailed and comprehensive, especially for the bigger datasets. Therefore, it can be used to show the full and complex phylogenetic diversity and provides a framework for studying the molecular evolution and epidemiology of type A influenza virus. For more than 85% of type A influenza HA and NA sequences into GenBank, they are categorized in one unambiguous and unique group. Therefore, our results are a kind of genetic and phylogenetic annotation for influenza HA and NA sequences. In addition, sequences of swine influenza viruses come from 56 HA and 45 NA subgroups. Most of these subgroups also include viruses from other hosts indicating cross species transmission of the viruses between pigs and other hosts. Furthermore, the phylogenetic diversity of swine influenza viruses from Eurasia is greater than that of North American strains and both of them are becoming more diverse. Apart from viruses from human, pigs, birds and horses, viruses from other species show very low phylogenetic diversity. This might indicate that viruses have not become established in these species. Based on current evidence, there is no simple pattern of inter-hemisphere transmission of avian influenza viruses and it appears to happen sporadically. However, for H6 subtype avian influenza viruses, such transmissions might have happened very frequently and multiple and bidirectional transmission events might exist.

Figure 1. The workflow.

This figure takes HA sequences of H2 subtype as an example to illustrate the workflow. In step I, we carry out a PCOORD (Ai) and a phylogenetic analysis (Aii) using the H2HA dataset. Results from the two methods support their division into two groups, H2g1 and H2g2. In step II, we repeat step I using two sub-datasets, H2g1 (Bi and Bii) and H2g2 (Ci and Cii). Bi and Bii show the results from PCOORD and the phylogenetic tree using H2g1. Group H2g1 is further divided into 5 subgroups. Similarly, Ci and Cii display the results from PCOORD and the phylogenetic tree using H2g2, and this group can also be further divided into some smaller subgroups. In step III, we summarize the results into a table (D) and in a tree-like figure (E). Panel E summarizes the phylogenetic diversity of HA sequences of H2 influenza A virus and values after the underlines indicate the numbers of sequences in the group or subgroup. If there are groups can be further divided based on the results of step II, we will repeat step II until there are no distinctly separated groups or the bootstrap values are too low to support further sub-division.

Figure 2. Phylogenetic diversity indicated by the association of HA and NA subtypes.

Panel A shows the phylogenetic diversity indicated by the NA subtype distribution among HA subgroups, while panel B shows the phylogenetic diversity indicated by the HA subtype distribution among NA subgroups.

Figure 3. PCCORD of H1 subtype influenza viruses.

In this figure, each sequence is shown as a dot using shape to signify host and color to indicate geographic region (the continent this virus was isolated from). For Figure 4 to 8 and all PCCORD figures available from our webpage, we use the same shape and color coding.

Figure 4. PCCORD of H3 subtype influenza viruses.

The sequences are coded for host (shape of dot) and geographic origin (color) as for figure 3.

Figure 5. PCCORD of H5 subtype influenza viruses.

The sequences are coded for host (shape of dot) and geographic origin (color) as for figure 3.

Figure 6. PCCORD of H9 subtype influenza viruses.

The sequences are coded for host (shape of dot) and geographic origin (color) as for figure 3.

Figure 7. PCCORD of N1 subtype influenza viruses.

The sequences are coded for host (shape of dot) and geographic origin (color) as for figure 3.

Figure 8. PCCORD of N2 subtype influenza viruses.

The sequences are coded for host (shape of dot) and geographic origin (color) as for figure 3.

Figure 9. Numbers of subgroups associated with SIV.

Panel A shows the numbers of subgroups associated with swine influenza viruses among different HA subtypes, while panel B shows the numbers of subgroups among different NA subtypes.