DEVELOPMENT OF A SYSTEM FOR THE DETECTION OF HIGH IMPACT PRRSV EMERGENT STRAINS

• Sponsoring Institution: State Agricultural Experiment Station
• Project Status: COMPLETE
• Funding Source: STATE
• Reporting Frequency: Annual
• Accession No.: 1026984
• Project Start Date: Jun 22, 2021
• Project End Date: Jun 30, 2023
• Program Code: [(N/A)]- (N/A)

Recipient Organization
UNIV OF MINNESOTA
(N/A)
ST PAUL,MN 55108

Performing Department
Veterinary Population Medicine

Non Technical Summary
Background/Problem:•The emergence and spread of new genetic variants hinders control of porcine reproductive and respiratory syndrome virus(PRRSV)• Identifying and tracking variants in large-scale sequence databases in real-time is challenging, but necessary to generateuseful and actionable insights for swine producer and practitioner stakeholders.• Early indicators of emergence (reflective of emergence/widespread transmission) are detectable in viral phylogenies, thus itmay be possible to anticipate which PRRS variants may become widespreadObjective: Develop a PRRSV molecular surveillance platform for detection of emerging genetic variants and tracking geneticdiversityResearch plan:Data: Creation of NextStrain platform for analysis of 45,000 ORF5 sequences from UMN VDL and MSHMP databases (2001-present) Approach: Evaluate correlations between the future success of variants with evolutionary attributes measurable fromthe phylogeny at time t. Future success of a variant will be measured by quantifying its expansion in frequency, geographicextent, or genetic diversity.Our team has experience in phylogenetics, PRRSV epidemiology and evolutions, and development of novel analyticalplatforms.

Animal Health Component

50%

Research Effort Categories

Basic

50%

Applied

50%

Developmental

0%

Classification

Knowledge Area (KA)	Subject of Investigation (SOI)	Field of Science (FOS)	Percent
311	3510	1170	100%

Knowledge Area
311 - Animal Diseases;

Subject Of Investigation
3510 - Swine, live animal;

Field Of Science
1170 - Epidemiology;

Keywords

epidemiology

phylogenetics

porcine reproductive and respiratory syndrome virus

multi-strain dynamics

virus evolution

Goals / Objectives
Porcine reproductive and respiratory syndrome virus (PRRSV) is the most important endemic pathogen to the United Statesswine industry [1]. PRRSV causes seasonal outbreaks in 20-30% of breeding herds (Tousignant et al., 2015), spreads readilybetween farms through animal movements and indirect transmission, and results in an economic burden of >$600 mill/year. Theimpact of PRRSV is primarily due to reductions in birthing rates, number of weaned pigs, poor growth, and mortality. A majorobstacle for the control of the PRRSV is its rapid rate of evolution and broad genetic and antigenic heterogeneity. As the virus evolves and new genetic variants emerge, immune responses generated against a past viral strain likely become less effective.Therefore, individual and population immunity may drive the emergence and spread of variants that can evade host immunity,resulting in the co-circulation of numerous PRRSV genetic variants. Evolving genetic variability, combined with variable crossimmunity, are primary contributors to why control of the disease has been challenging in the U.S. swine industry.The overarching objective of this proposal is to develop a PRRSV molecular surveillance platform for detectionof emerginggenetic variants and tracking genetic diversity in the U.S. For the purposes of this proposal, a variant is defined as a clade of closely related PRRS viruses present at a particular point of time (<2% genetic distance - see research plan), and the future success of a given variant will be measured by quantifying its expansion through time. Following , expansions in frequency, geographic extent, and/or genetic diversity of a variant are characteristics of widespread transmission, and such variants will be considered emerging. However, there may be an evolutionary signature detectable in the phylogenetic tree early in the emergence period, thus it may be possible to detect variants with high emergence potential before or earlier during emergence.Therefore, we outline the following specific aims:1. Identifying early indicators of the emergence potential of PRRSV variants in large-scale sequence datasets• H1: The future success of variants is correlated with measurable attributes that can be quantified from the structure ofphylogenetic trees.• Outcome 1a: Validated methodology that can be applied prospectively to large-scale phylogenetic datasets to rank variantsaccording to their emergence potential.• Outcome 1b: HTML-based platform (NextStrain) for phylogenetic analysis and visualization for routine use, includingexpansion to WGS tree visualization for enhanced interpretation.2. Integrating epidemiological data into phylogenies for enhanced surveillance and epidemiological investigations• H2: Highly successful variants have distinctive spatiotemporal properties (e.g., rate of geographic expansion) that arequantifiable early in emergence• Outcome 2a: Validated methodology to identify emerging variants based on spatiotemporal properties• Outcome 2b: Implementation of spatiotemporal sequence analysis on NextStrain for prospective use in routine surveillance,and for rapid analysis and communication of the spatiotemporal distribution of any virus of concern to stakeholders

Project Methods
Availability of large-scale sequence datasetsPRRSV ORF5 sequence databases available for this work come from the UMN VDL and MSHMP. Together, these twodatabases archive >4,000 sequences annually, with a combined total of ~45,000 sequences between 2001-2020, andoriginating from at least 50% of the U.S. breeding and growing pig herd. VDL data includes data on state, and includessequences for viruses collected from throughout the U.S., whereas MSHMP includes sequences from breeding and growfinishing sites of its participants that were sequenced by the UMN, ISU, SDSU, and KSU VDLs, with additional meta-data on farm location, herd size, production system, breeding herd PRRSV status and air filtration status. WGS is becoming increasinglycommon in both databases. Despite the availability of these databases, we currently lack a platform to visualize and analyzethese data, both for the purposes of molecular surveillance and epidemiological investigations.Analysis of early indicators related to emergence potential of PRRSV variantsIn order to achieve the objectives of this proposal we will 1) classify all available ORF5 sequences at time t into variants, 2)measure each variant's evolutionary attributes based on the structure of the phylogenetic tree at time t, 3) evaluate the successof each variant during a follow-up period of n months, and 4) correlate evolutionary attributes at time t to a variant's futuresuccess during the follow-up period to identify attributes that can serve as early indicators of the emergence potential of avariant. This analysis will be conducted both for the VDL database (Aim 1- sequence data only), and for the MSHMP database(Aims 1 - 2: inclusion of spatial meta-data)1. Classifying variants: A variant is defined as a clade of closely related PRRSV, such that at least one sequence within theclade was detected during time period t (i.e., a particular year, half-year, or quarter). Several alternative variant classificationstrategies will be compared. For example, simple rule-based approaches based on patristic distance between sequences (i.e.,sum of the branch lengths between two sequences, where branch length corresponds to genetic distance) could be used tocluster sequences, whereby clades are defined such that all sequences within the clade must be within a threshold distance,conventionally 2%, from at least one other sequence. Based on this heuristic, a crude analysis with a subset of data suggeststhat we may expect ~60 variants per 1000 sequences, therefore the analysis may include >200 variants per year over the pastdecade. Alternative clustering methods include discriminant analysis of principal components and ClusterPicker software.2. Measurement of attributes: Evolutionary attributes of each variant that will be measured for the phylogenetic tree at time tinclude the local branching index - an estimate of evolutionary fitness based on the supposition that rapid diversification of avariant is indicative of population expansion; tree imbalance - a hallmark of evolutionary dynamics characterized by immuneescape, characterized by a skewed distribution of a tree's internal nodes due to variation in branching rates ; episodic positiveselection - relative rate of non-synonymous to synonymous mutations calculated for each branch in the tree , among others. Forthe MSHMP database where information on farm identity, location, and production system are available (Aim 2), a variant'sspatiotemporal properties prior to time t will also be measured, such as rate of geographic expansion, and number offarms/production systems affected (Aim 2).3. Quantifying success: The success of a variant will be measured in a follow-up period of n months, with n corresponding toeither 6 or 12 months (akin to the next seasonal peak in PRRS incidence), 3 or 5 years (timeframe appropriate to captureemergence of new sub-lineages, Figure 1). Variability in success will be alternatively quantified as the relative frequency orgenetic diversity of a variant's descendants in the follow-up period, or how closely related a variant is to the most prevalentvariant identified in the follow-up period (Aim 1). For the MSHMP database, success will additionally be approximated by avariant's spatiotemporal properties in the follow-up period, such as changes in the geographic extent, and number offarms/production systems affected (Aim 2).Early indicators of emergence potential: Rapidly increasing frequency, diversity, geographic extent, and/or number offarms/production systems affected during the follow-up period are indicative of highly successful emerging variants withexpanding viral population sizes [14]. Variant attributes at time t will be correlated with measures of success in the follow-upperiod to identify attributes that are predictive of future success. Methods include bivariate approaches (Spearman'scorrelations) and multi-variable approaches (regression). We expect that variants with high local branching indices, highmutation rates, or evidence of selection (tree imbalance or episodic selection) will be correlated with future success. We alsoexpect that detection of highly similar sequences across multiple production systems (or in multiple states) is reflective of rapidtransmission, and thus serve as an indicator of emergence.NextStrain platform for prospective use in routine surveillance, outbreak investigations, and for rapid analysis andcommunicationWe will develop a PRRS-specific NextStrain platform for the analysis and visualization of large-scale PRRSV databases. Theend-user of this platform will be the UMN VDL and MSHMP, who will communicate with producers/clients per their confidentialityagreements (representing >50% of the U.S. swine herd). This platform will construct time-scaled phylogenetic trees, and allowfor the visualization of meta-data such as location, RFLP-type, lineage classification, point mutations, and may provideopportunities for the detection of emerging variants in a timely fashion.As an interactive platform, it will also allow for zooming in to particular clades of interest, and for identification of specificmutations associated with that clade (and thus mutations associated with particular variants associated with particularoutbreaks). Through this platform, users will rapidly be able to assess the relatedness of any sequence to all other sequences inthe dataset, visualize the spatial distribution of a clade, date of emergence, and other metadata that may be relevant forpractitioners and industry stakeholders. In addition, calculations of metrics identified as early indicators of a variant's emergencepotential can be incorporated into the platform and visualized as a heatmap onto the tree to rapidly identify variants with highemergence potential at a particular point of time. This platform can help meet the demands of rapid analysis and stakeholdercommunication during outbreak investigations, and will also be used to generate routine quarterly reports on the relativefrequency of different variants and potential emergence of new variants.Observing the evolutionary trajectories of variants highlights viruses that are of particular concern and should be targeted foradditional whole genome sequencing and analysis. WGS data is notoriously difficult to interpret. Thus, we will adapt NextStainfunctionalities so that users can rapidly switch between trees based on WGS, structural genes, and ORF5, which is a NextStrainfeature that was originally developed for segmented viruses such as influenza. Branch-specific mutations can be rapidlyidentified and visualized on the WGS tree. These features will enhance our ability to interpret whole-genome sequence data.

Progress 06/22/21 to 06/30/23

Outputs
Target Audience:The target audience of this project was molecular epidemiologists, swine diagnosticians (veterinary diagnostic labs), vaccine developers, and swine health professionals. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?The project was developed by a veterinarian PhD candidate (Dr. Nakarin Pamornchainavakul), supervised by the PI. Opportunities associated to development in areas as diverse as coding, virology, phylogenetics analysis and epidemiology were provided by this project. Oral presentations directly related to this project were accepted in international conferences, and the PhD candidate above mentioned performed the presentations. Poster presentations with products of this project were also done, the PhD candidate presenting all of those. These events provides a unique exposure opportunity and greater contact with peers in a professional environment. This work formed two chapters of the student's dissertation, which he expects to submit in January 2024. How have the results been disseminated to communities of interest? Results from this project have been disseminated to communities of interest. This was achieved by academic publications (2 published, one submitted), presentations at academic conferences (1) and industry meetings (3),lay articles in National Hog Farmer (2), and by podcast participation. What do you plan to do during the next reporting period to accomplish the goals? Nothing Reported

Impacts
What was accomplished under these goals? In order to achieve goals 1a and 1b, we developed an algorithm that can be prospectively applied to large-scale datasets to estimate the odds of a PRRSV variant to emerge. This algorithm is based on the achievements of objective 2 (see paragraphs below) but the access to this tool is not public yet. Essentially, steps to identify emerging variants of PRRSV were programmed to allow the input of collection of sequences and metadata such as that the output contains how likely a variant is to emerge as compared to other variants. In order to achieve goal 2a, we developed an algorithm that can be applied to datasets to estimate the odds of a PRRSV variant to emerge.We analyzed a decade's worth of virus ORF5 sequences (n = 20,700) and corresponding metadata to identify phylogenetic-based early indicators for short-term (12 months) and long-term (24 months) variant emergence. A total of 74 unique sets of time-scaled phylogenetic trees with a median size of 4,247.5 (IQR = 2,688 - 6,712.75) taxa were reconstructed to across 6-month intervals throughout 2011 - 2020.Classified by 2% average pairwise patristic distance, the median number of variants per tree was 151 (IQR = 96 - 204), the median size of a variant was 12 (IQR = 5 - 30) taxa per variant.We evaluated population expansion, spatial distribution, and genetic diversity as key success metrics for variant emergence. For population expansion, the successfulvariants were typically at least twice as many sequences in the follwoup period compared to the time of analysis [median relative increase in number of taxa = 400 (IQR = 283.33 - 804.69)%]. Similarly for genetic diversification, successful variants increased in their genetic diversity from the by a median of 0.01 (IQR = 0.008 - 0.018), while the diversity of non-successful variants decreased by the median of 0.004 (IQR = 0.002 - 0.009). Geographic expansion metrics were also well stratified between successful and non-successful variants, particularly when considering measures based on estimated geographical distance; successful variants often doubled their geographic distribution with distances increasing by up to 1000 km or more, whereas non-successful variants frequently displayed no increase whatsoever. Conditional logistic regression revealed the local branching index (LBI) as the sole informative indicator for predicting population expansion;variants in the upper quartile for ancestral LBI had approximately 12 - 13 times higher odds of being successful (AbIn.Taxa > 10, or having at least 10 more taxa in the follow-up period) in the next 12 or 24 months compared to variants in the lower quartile.Ancestral mutation length was strongly linked to future genetic diversity. Predicting spatial dispersion relied on multiple predictors, but their causal relationships remain unclear. Although the predictive models effectively captured most emerging variants, they exhibited relatively low positive predictive value due to high false positivity. The pipeline for spatiotemporal sequence analysis on NextStrain is developed, and has been uitlized to visualize the emergence potentical of variants (aim 2b). As an internal grant at the University of Minnesota, part of the intent of this grant program is to help leverage external funding opportunities. Results from this project were used as preliminary data to successfully apply and receive funding from the USDA Critical Agricultural Research and Extension Program, and the USDA Data Science for Food and Agricultural Systems program.

Publications

• Type: Journal Articles Status: Published Year Published: 2023 Citation: Pamornchainavakul, N., I. A. D. Paploski, D. N. Makau, M. Kikuti, A. Rovira, S. Lycett, C. A. Corzo, and K. VanderWaal. 2023. 'Mapping the Dynamics of Contemporary PRRSV-2 Evolution and Its Emergence and Spreading Hotspots in the U.S. Using Phylogeography', Pathogens, 12.
• Type: Journal Articles Status: Submitted Year Published: 2023 Citation: Nakarin Pamornchainavakul, Mariana Kikuti, Igor AD Paploski, Cesar A Corzo, and Kimberly VanderWaal. In prep. Predicting Potential PRRSV-2 Variant Emergence Through Phylogenetic Inference
• Type: Journal Articles Status: Published Year Published: 2022 Citation: Pamornchainavakul, Nakarin, Mariana Kikuti, Igor A. D. Paploski, Dennis N. Makau, Albert Rovira, Cesar A. Corzo, and Kimberly VanderWaal. 2022. 'Measuring How Recombination Re-shapes the Evolutionary History of PRRSV-2: A Genome-Based Phylodynamic Analysis of the Emergence of a Novel PRRSV-2 Variant', Frontiers in Veterinary Science, 9.
• Type: Other Status: Published Year Published: 2022 Citation: Lay Article in National Hog Farmer. Mapping hotspots for spread of PRRSV-2 lineage 1 in the United States. Authors: N. Pamornchainvakul,& K. VanderWaal,&,. Aug, 2022. https://www.nationalhogfarmer.com/animal-health/mapping-hotspots-spread-prrsv-2-lineage-1-united-states
• Type: Other Status: Published Year Published: 2022 Citation: Lay Article in National Hog Farmer. Tracing the origin of the novel PRRSV-2 L1C-1-4-4 variant. Authors: N. Pamornchainavakul [&] K. VanderWaal. March, 2022. https://www.nationalhogfarmer.com/animal-health/tracing-origin-novel-prrsv-2-l1c-1-4-4-variant
• Type: Conference Papers and Presentations Status: Published Year Published: 2023 Citation: VanderWaal, K. 2023. Chasing a moving target: rapid evolution of PRRS in the U.S. Allen D. Leman Conference 2023. St. Paul, MN
• Type: Conference Papers and Presentations Status: Published Year Published: 2022 Citation: PAMORNCHAINAVAKUL, NAKARIN; PAPLOSKI, IGOR A.D.; MAKAU, DENNIS N.; KIKUTI, MARIANA; ALBERT, ROVIRA; LYCETT, SAMANTHA; CORZO, CESAR A.; VANDERWAAL, KIMBERLY. Mapping hotspots for emergence and inter-regional spread of contemporary PRRSV sub-lineages in the United States using phylogeography. The International Symposium of Veterinary Epidemiology and Economics, Halifax, Canada, 2022
• Type: Conference Papers and Presentations Status: Published Year Published: 2022 Citation: PAMORNCHAINAVAKUL, NAKARIN; MAKAU, DENNIS N.; PAPLOSKI, IGOR A.D.; KIKUTI, MARIANA; LYCETT, SAMANTHA; CORZO, CESAR A.; VANDERWAAL, KIMBERLY. Early indicators of the emergence potential of PRRSV-2 variants based on phylogenetic structure. Allen D. Leman Swine Conference, Saint Paul, US, 2022
• Type: Conference Papers and Presentations Status: Published Year Published: 2023 Citation: PAMORNCHAINAVAKUL, NAKARIN; KIKUTI, MARIANA; PAPLOSKI, IGOR A.D.; CORZO, CESAR A.; VANDERWAAL, KIMBERLY. Predicting PRRSV-2 Variant Emergence: Insights from a Decade of Genomic Analysis. Allen D. Leman Swine Conference, Saint Paul, US, 2023

Progress 10/01/21 to 09/30/22

Outputs
Target Audience:Porcine reproductive and respiratory syndrome virus (PRRSV) is the most important endemic pathogen to the United States swine industry. The impact of PRRSV is primarily due to reductions in birthing rates, number of weaned pigs, poor growth, and mortality. A major obstacle for the control of the PRRSV is its rapid rate of evolution and broad genetic and antigenic heterogeneity. As the virus evolves and new genetic variants emerge, immune responses generated against a past viral strain likely become less effective. Sequencing is routinely performed for PRRSV, especially when dealing with outbreaks or elimination projects in breeding herds, and databases maintained by the University of Minnesota Diagnostic Laboratory (VDL) and the Morrison Swine Health Monitoring Project (MSHMP) contain >40,000 contemporary and historical ORF5 sequences. However, we currently lack a framework for identifying and tracking variants in these large-scale datasets. Anticipating variants that have the potential to become widespread would provide an invaluable tool for rapid response to emerging viral threats. As such, our target audience are industry practitioners and swine producers, that may be interested in anticipating PRRSV viral strains with a larger probability of causing large-scale epidemics in the U.S. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?The project is being developed by a veterinarian PhD candidate, supervised by the PI. Opportunities associated to development in areas as diverse as coding, virology, phylogenetics analysis and epidemiology were provided by this project. Oral presentations directly related to this project were accepted in international conferences, and the PhD candidate above mentioned performed the presentations. Poster presentations with products of this project were also done, the PhD candidate presenting all of those. These events provides a unique exposure opportunity and greater contact with peers in a professional environment. How have the results been disseminated to communities of interest?Results from this project have been disseminated to communities of interest. This was achieved by academic publications (one full paper already published, another paper on co-author review), oral and poster presentations in conferences and by podcast participations. These are the main ways in which the direct dissemination of the results to the community of interest. What do you plan to do during the next reporting period to accomplish the goals?We intend to fully develop a pipeline that will automate the process of sequence and metadata input and analysis to evaluate the potential of emergence of PRRSV variants. This will be achieved by automating the steps for cleaning and producing the files necessary for NextStrain illustration of the emergence potential of the PRRSV strains. This will allow for the implementation of sequence analysis prospectively, which is important for routine surveillance analysis run by systems or Veterinary Diagnostic Laboratories.

Impacts
What was accomplished under these goals? In order to achieve goals 1a and 1b, we developed an algorithm that can be prospectively applied to large-scale datasets to estimate the odds of a PRRSV variant to emerge. This algorithm is based on the achievements of objective 2 (see paragraphs below) but the access to this tool is not public yet. Essentially, steps to identify emerging variants of PRRSV were programmed to allow the input of collection of sequences and metadata such as that the output contains how likely a variant is to emerge as compared to other variants. In order to achieve goal 2a, we developed an algorithm that can be applied to datasets to estimate the odds of a PRRSV variant to emerge. This was achieved using 28,695 ORF5 sequences with temporal and spatial metadata collected in the US from 2010-2021. The future emergence potential at each observation time point (t) in the past decade was estimated. For each time t (set as every 6 months), we constructed "pre-trees" (i.e., timed phylogenetic trees of sequences available in the previous 6, 12, or 24 months to t) to obtain potential predictors of emergence for each variant (defined as a cluster of viruses with a mean pairwise distance of 2%). Examples of predictors we used are branch length, substitution rate, local branching index (LBI), and the distinctiveness of the inferred amino acid sequence and putative N-glycosylation pattern at a variant's ancestor. A variant's future success during a follow-up period was measured based on changes in population size, spatial distribution, and genetic diversity in "post-trees" (i.e., trees created from the same set of sequences in pre-trees plus sequences available in the next 6, 12, and 24 months after t). Variants were categorized as "successful" (if they were above the top 95th percentile of each success measure) and "common" (if they were below the 75th percentile of each success measure). Variants were artificially sampled in a 1:3 proportion (successful:common) for a matched case-control study to evaluate what predictors were different between the variants with different success measures. Different scenarios considering different times t in which data was taken into account were run. The variant's potential for emergence was best predicted when using predictors from the 6-month period prior to the actual emergence in the following 24 months. The predictive performance was relatively high when predicting both population expansion and genetic diversification. The best fit model suggests that the timed tree's ancestral branch length (clock length), local branching index, and the distinctiveness of putative ancestral N-glycosylation pattern are the most informative predictors to the expansion of a variant. The pipeline for spatiotemporal sequence analysis on NextStrain is developed. In order to achieve goal 2b, sequences need to be cleaned and analyzed, with relevant metadata, in BEAST. Output of the analysis is illustrated on NextStrain. This pipeline is developed but is not currently public.

Publications

• Type: Journal Articles Status: Published Year Published: 2022 Citation: PAMORNCHAINAVAKUL, NAKARIN; KIKUTI, MARIANA; PAPLOSKI, IGOR A.D.; MAKAU, DENNIS N.; ROVIRA, ALBERT; CORZO, CESAR A.; VANDERWAAL, KIMBERLY. Measuring How Recombination Re-shapes the Evolutionary History of PRRSV-2: A Genome-Based Phylodynamic Analysis of the Emergence of a Novel PRRSV-2 Variant. FRONTIERS IN VETERINARY SCIENCE, DOI: 10.3389/fvets.2022.846904, 2022
• Type: Journal Articles Status: Other Year Published: 2023 Citation: PAMORNCHAINAVAKUL, NAKARIN; PAPLOSKI, IGOR A.D.; MAKAU, DENNIS N.; KIKUTI, MARIANA; ROVIRA, ALBERT; LYCETT, SAMANTHA; CORZO, CESAR A.; VANDERWAAL, KIMBERLY. Mapping Dynamics of Contemporary PRRSV-2 Evolution, Emergence, and Spreading Hotspots in the U.S. using Phylogeography. In Preparation.
• Type: Conference Papers and Presentations Status: Published Year Published: 2022 Citation: PAMORNCHAINAVAKUL, NAKARIN. Genomic recombination and the epidemiological emergence of novel PRRSV-2 variants: a genome-base phylodynamic approach. 16th International Symposium of Veterinary Epidemiology and Economics  ISVEE, Halifax, CA, 2022
• Type: Conference Papers and Presentations Status: Published Year Published: 2022 Citation: PAMORNCHAINAVAKUL, NAKARIN; MAKAU, DENNIS N.; PAPLOSKI, IGOR A.D.; KIKUTI, MARIANA; LYCETT, SAMANTHA; CORZO, CESAR A.; VANDERWAAL, KIMBERLY. Early indicators of the emergence potential of PRRSV-2 variants based on phylogenetic structure. Allen D. Leman Swine Conference, Saint Paul, US, 2022
• Type: Other Status: Published Year Published: 2022 Citation: PAMORNCHAINAVAKUL, NAKARIN. Tracking lineage of PRRS variants across the US. The Swine Health Blackbelt Podcast, Youtube, 29jun2022. Podcast Participation. Link: https://www.youtube.com/watch?v=w8xf1wxqQ8s

Progress 06/22/21 to 09/30/21

Outputs
Target Audience:Swine industry veterinary practitioners and other swine industry stakeholders Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?This project supports one graduate student (Pamornchainavakul), one post-doctoral researcher (Paploski), and one research associate (Kikuti).The analysis is being lead by the graduate student, who is receiving training on bioinformatics and evolutionary modeling.Paploski and Kikuti are assisting with data curation from UMN VDL and MSHMP, respectively, as well as tointerpretation of results. How have the results been disseminated to communities of interest?Paploski presented results related to this project at the McKean Swine Conference, a conference for veterinary practicioners hosted by Iowa State University.In addition, a paperis the final preparation stages for submission, which focuses on the role of recombination and the representativeness of ORF5 data (relative to whole genome sequencing-WGS) for tracking the evolutionary trajectory of different PRRS variants. What do you plan to do during the next reporting period to accomplish the goals?We will submit the first paper comparing ORF5 and WGS phylodynamics in January 2022.In addition, now since we have compiled a curated sequence database, we willa) begin classifying sequences into distinct variants (<2% dissimilarity on ORF5),b) quantify their evolutionary attributes and c) expansion over time, and d) correlate their evolutionary attributes to future success as a means to rank the emergence potential of each variant.During the next reporting period, we expect to accomplish part a and b (aim 1), as well as complete the phylogeographic analysis (aim 2). Second, there has been numerous discussions amongst industry circles on replacing the current RFLP-type method of classifying PRRS viruses with a better system.Our project forms the foundation for establishing a means to classify variants, and thus we intend to initiate more formal discussions amongst prominent veterinary diagnostic laboratories as well as industry leaders on establishing a standardized approach to classify and name PRRS variants.

Impacts
What was accomplished under these goals? Over 45,000 PRRSV open reading frame 5 (ORF5) genetic sequences between 2002 and 2021 were obtained from public and private sources. 19,179 were obtained from the UMN Veterinary Diagnostic Laboratory, 22,427 from the Morrison Swine Health Monitoring Project, and 17,908 from GenBank. These datasets form the database fom which we can identify patterns of strain emergence, construct phylogenetic models, and evaluate predictors of emergence potential (Aims 1 and 2).The infrastructure for the NextStrain platform has been developed with a preliminary sub-sample of sequences, which can be scaled up to accommodate more data (Aim 1).From initial data exploration, it was apparent that some regions of the U.S. (particularly Minnesota) are over-represented in the dataset. In addition, some datapoints were likely duplicated across data sources.Data were de-duplicated and filtered for quality, resulting in 31,475 sequences.After data filtration and stratified random sampling, five different subsets of 500 Lineage 1 ORF5 sequences each were assembledto construct discrete phylogeographic models (Aim 2). Preliminary findings show that the L1 ancestor emerged in 1990 (1987-1992 95%HPD) and its successors peaked in population growth every ~6.2 years on average. The timing of these peaks appears to coincide with the emergence of sub-lineages known to be atypically virulent. Most sub-lineage emergence events tended to occur in the Upper Midwest, an area characterized by high pig densities and more numerous different production systems. Dispersal rates were highest between the Upper Midwest to the Central Midwest, followed by subsequent inter-regional dispersal events originating from the Central Midwest.

Publications