Sampling strategies and biodiversity of influenza A subtypes in wild birds

Abstract

Wild aquatic birds are recognized as the natural reservoir of avian influenza A viruses (AIV), but across high and low pathogenic AIV strains, scientists have yet to rigorously identify most competent hosts for the various subtypes. We examined 11,870 GenBank records to provide a baseline inventory and insight into patterns of global AIV subtype diversity and richness. Further, we conducted an extensive literature review and communicated directly with scientists to accumulate data from 50 non-overlapping studies and over 250,000 birds to assess the status of historic sampling effort. We then built virus subtype sample-based accumulation curves to better estimate sample size targets that capture a specific percentage of virus subtype richness at seven sampling locations. Our study identifies a sampling methodology that will detect an estimated 75% of circulating virus subtypes from a targeted bird population and outlines future surveillance and research priorities that are needed to explore the influence of host and virus biodiversity on emergence and transmission.

Figure 1. GenBank data on subtypes displayed with host and geographic information.

(a) Subtypes found in Anseriformes (green), Charadriiformes (purple), Procellariiformes (blue), more than one order (gray), and occurrence in domestic birds (***); (b) richness of identifiable wild bird genera (of 81 examined) associated with each subtype (+ indicates found in at least one family where genus was not determinable) and occurrence in domestic birds (red); (c) distribution of subtypes unique to the continents of North America (blue), Europe (brown), Asia (tan), Australasia (green with black X), and across multiple continents (gray); and (d) distribution of subtypes unique to the North American Pacific Flyway (light green), Central Flyway (yellow), Mississippi Flyway (dark green), and Atlantic Flyway (red). Also indicated are subtypes unique to North America but found in multiple flyways (blue), subtypes unique to other continents (brown), and subtypes found across multiple continents (gray).

Figure 2. Maps of global sampling effort and observed richness for nine regions.

(a) Global sampling effort (total number of birds tested) by region reported in 50 studies that non-discriminately tested for AIV subtype. If we did not identify studies that met the inclusion criteria for a country, state (USA only), or province (Canada only) we report it as not observed. (b) Total richness (number of AIV subtypes detected) by region based on GenBank records and 50 studies. A country, state (USA only), or province (Canada only) is not observed if we did not identify studies that met the inclusion criteria and if no subtypes were reported in GenBank.

Figure 3. Subtype richness and sampling effort varies among studies.

(a) Predicted AIV subtype richness (diamonds) with 95% confidence intervals based on the mean Chao2 richness estimator for seven distinct study locations for which the criteria of five sampling events and >5,000 birds tested non-discriminately for AIV subtypes were met (Table? 1). Further, predicted AIV subtype Chao2 richness estimates are displayed from GenBank. (b) The sample size at each study location (blue bar) and minimum sufficient sampling size necessary to capture 75% of total estimated AIV richness (black dot). EGY  =  Egypt , PRT  =  Portugal , MNG  =  Mongolia (this study), EU  =  European Union (this study), CAN  =  Alberta, Canada , SWE  =  Sweden , TWN  =  Taiwan , and GB  =  sequences with subtypes deposited in GenBank (1959–2012).