Sources and sinks of influenza A virus genomic diversity in swine from 2009 to 2022 in the United States

Data and R code used to run analyses for the manuscript "Sources and sinks of influenza A virus genomic diversity in swine from 2009 to 2022 in the United States".

Janzen, G.M, Inderski, B.T., Chang, J., Arendsee, Z.W., Janas-Martindale, A., Torchetti, M.K., Baker, A.L., and Anderson, T.K. Sources and sinks of influenza A virus genomic diversity in swine from 2009 to 2022 in the United States. bioRxiv XX:XX.

Description

This manuscript summarizes 13 years of influenza A virus evolution in the swine host, focusing on the interplay between ecological concepts and swine agricultural practices, and offers tools for modeling the spread of novel virus strains across the US.

Abstract

Influenza A virus (IAV) in swine in the U.S. is surveilled to monitor genetic evolution to inform intervention efforts and aid pandemic preparedness. We describe data from the U.S. Department of Agriculture National Surveillance Plan for Influenza A Virus in Pigs from 2009 to 2022. Clinical respiratory cases were subtyped followed by sequencing of hemagglutinin (HA) and neuraminidase (NA), and a subset of viruses were whole genome sequenced. Phylogenetic analysis identified geographic and temporal IAV reassortment hotspots. Regions acting as IAV genomic diversity sources or sinks were quantified, and dissemination was qualified and modeled. The dominant IAV clades were H1N2 (1B.2.1), H3N2 (1990.4.a), and H1N1 (H1-1A.3.3.3-c3). Internal genes were classified as triple-reassortant (T) or pandemic 2009 (P), and three genome constellations represented 73.5% of detections across the last two years. In some years, the distribution of IAV diversity was so narrowly distributed that it presented a statistical signal associated with local adaptation. We also demonstrated that the source of most IAV genomic diversity was in Midwest states (IL, MO, IA), and while this was correlated with swine inventory, the emergence and persistence of diversity was tied to swine transport across the U.S. The continued regional detection of unique HA, NA, and genome constellations provides support for targeted interventions to improve animal health and enhance pandemic preparedness.

Getting Started

Dependencies

• These scripts were run using R version 4.2.0 (2022-04-22 ucrt) on Windows x86_64.

Outline

Script 1

This script brings together a lot of data, cleans it, formats it, and produces a few data frames that the following scripts will use. This script must be run once to build the IAV_Sources_and_Sinks.RData file, which all following scripts will load into memory before progressing. Some summary analyses are also run in this script. While not strictly necessary, I recommend clearing the environment before moving on to Script 2 or Script 3, either by restarting RStudio or by using rm(list=ls()), as it will clear resources.

Script 2

This script runs the Zeta diversity analyses and produces the associated plots. Core functionality is provided with the zetadiv package. You must load IAV_Sources_and_Sinks.RData from Script 1 to run this script. This script saves nothing to the .RData file.

Script 3

This script establishes a time series of detections and runs analyses and looks at temporal and geographic patterns of detections. You must load IAV_Sources_and_Sinks.RData from Script 1 to run this script.

Script 4

This script uses Markov Chains to model state-to-state IAV transition probabilities, provides code for testing real and hypothetical IAV strain detection orders, and statistics to evaluate model prediction performance. You must load IAV_Sources_and_Sinks.RData, modified and saved in Script 3, to run this script.

Script 5

This script draws out detections of three IAV clades and compares evolutionary trends between them. This script creates clade-specific Markov Chain models, testing correlation in their transition matrices via Mantel test, and plots the strongest edges of the transition matrix on a map. This script also generates Zeta diversity of these clades (not presented in the published manuscript). You must load IAV_Sources_and_Sinks.RData, modified and saved in Script 4, to run this script. This script saves nothing to the .RData file.

Microreact

The file microreact_input_5y.csv is written by script 1. The microreact figure in this manuscript can be made by following these steps:

• Go to the microreact website.
• Drag and drop the file into the page.
• Select continue.
• Select continue.
• To make the legend, select Legend to pull up the Legend, download it by clicking on the three bars in that window and downloading it as PNG.
• To make the timeline, select Create new Timeline from the three bars from the pencil icon at the top right. Change the Temporal Data Type to One column: Formatted Values, change the Temporal Data Column to Date, and click close. Download it by clicking on the three bars in that window and downloading it as PNG.
• To format the map, select the sliders icon in that window and adjust as desired.

Authors

Garrett M. Janzen @garrettjanzen
Blake T. Inderski
Jennifer Chang
Zebulun W. Arendsee
Alicia Janas-Martindale
Mia Kim Torchetti
Amy L. Baker
Tavis K. Anderson

Acknowledgments

Thanks to co-authors Blake T. Inderski, Jennifer Chang, Zebulun W. Arendsee, Alicia Janas-Martindale, Mia Torchetti, Amy L. Baker, Tavis K. Anderson, and to all members of the Flu-Crew at the National Animal Disease Center, USDA-ARS.