Inferring the inter-host transmission of influenza A virus using patterns of intra-host genetic variation

Abstract

Influenza A viruses (IAVs) cause acute, highly transmissible infections in a wide range of animal species. Understanding how these viruses are transmitted within and between susceptible host populations is critical to the development of effective control strategies. While viral gene sequences have been used to make inferences about IAV transmission dynamics at the epidemiological scale, their utility in accurately determining patterns of inter-host transmission in the short-term--i.e. who infected whom--has not been strongly established. Herein, we use intra-host sequence data from the viral HA1 (hemagglutinin) gene domain from two transmission studies employing different IAV subtypes in their natural hosts--H3N8 in horses and H1N1 in pigs-to determine how well these data recapitulate the known pattern of inter-host transmission. Although no mutations were fixed over the course of either experimental transmission chain, we show that some minor, transient alleles can provide evidence of host-to-host transmission and, importantly, can be distinguished from those that cannot.

Figure 1.

Schematic of the networks of known contacts between hosts (red lines) and sampling dates (black dots) for the (a) equine H3N8 and (b) swine H1N1 experimental transmission studies. Y-axis labels indicate the unique identifiers randomly assigned to each animal in either study. The x-axis indicates the study day (day 0 is the day in which animals were mechanically inoculated). Red dots indicate the day on which that individual animal came into contact with an infected host or was mechanically inoculated. Where multiple red lines radiate from a single red dot there is some ambiguity as to which animal transmitted to which when three (in the equine study) or four (in the swine study) animals were housed together. For example, on day 3 of the swine study, pigs 113 and 115, having shown symptoms of infection, were housed with pigs 116 and 104.

Figure 2.

Flow diagram of the allele analysis ((a) equine H3N8; (b) swine H1N1) and associated summary statistics. Boxes are arranged to show how alleles were progressively segregated based on the hosts they were observed in. Box colours match with those assigned to the alleles displayed in figures 3 and 4.

Figure 3.

Allele networks for all sequences in the (a) equine H3N8 and (b) swine H1N1 studies. Blue nodes represent the dominant allele in each host as well as minor alleles that were shared between hosts known to have had direct contact. Light green nodes represent minor alleles that were observed more than once, but never shared between hosts in direct contact. Node sizes are a function of the number of times the each allele was observed in each host and relative only to others within each network. Some nodes were shrunk to improve clarity. Dark dotted lines connect minor alleles that are shared between hosts in a manner consistent with the known host-to-host contacts. Light green dotted lines connect instances of minor alleles shared between hosts which were never in direct contact. Nodes are positioned vertically based on the first day they came into contact with an infected host. Inset. The number of edges connecting shared minor alleles that were consistent or inconsistent with the pattern of host-to-host contacts.

Figure 4.

Approximated transmission networks between hosts constructed using specific minor alleles in the (a) equine H3N8 and (b) swine H1N1 studies. Blue nodes represent the overall dominant alleles (consensus strains) specific to each study and also indirectly represent each individual host. Orange nodes represent shared minor alleles that were selected by specific criteria (see text). Node sizes are drawn relative to the number of alleles they represent. Orange lines connecting hosts represent known host-to-host contacts that were accurately inferred. Dotted, grey lines connect hosts that had no direct contact. The thickness of these lines is proportional to the number of minor alleles they share. Faint dotted lines similar in colour connect shared minor allele nodes. Nodes are positioned in top-down chronological order based on when their minor alleles were observed and are positioned horizontally so that the edges between them are apparent. Inset. The number of edges connecting shared minor alleles that were consistent or inconsistent with the pattern of host-to-host contacts.