Timing the regional spread of PRRSV-2 variants across the United States

By analyzing 14,835 PRRSV-2 sequences from 2015 to 2024, this study quantifies the regional spread dynamics of 156 viral variants across the United States, revealing that the Upper Midwest serves as a central hub for both emergence and introduction while interior regions experience faster dispersal than coastal pathways, thereby providing critical data to enhance epidemiological models and disease preparedness.

The Gist

Imagine the U.S. swine industry as a massive, bustling city made up of five distinct neighborhoods: the Upper Midwest, Lower Midwest, Great Plains, Northeast, and Atlantic Seaboard. In this city, there is a tiny, invisible troublemaker called PRRSV-2 (a virus that makes pigs sick). This virus is a master shapeshifter; it constantly changes its "outfit" (genetic variants) to stay one step ahead of vaccines and defenses.

This paper is like a detective story where the authors tracked how these different "outfits" of the virus travel between neighborhoods over the last decade. Here is the breakdown of their findings in plain English:

1. The "Fashion Show" of Viruses

The virus doesn't just have one look; it has hundreds of different "variants" (like different fashion styles). The researchers looked at a massive database of virus samples (like checking a fashion logbook) to see which styles were popular in which neighborhoods.

• The Upper Midwest is the fashion capital. It has the most variety of virus styles (123 different ones!). However, it's a bit like a runway where one or two super-popular styles dominate the crowd, while the rest are just rare, niche trends.
• The Lower Midwest is the eclectic mix. It has fewer total styles than the Upper Midwest, but they are all worn more evenly. No single style completely takes over; it's a balanced community of different viruses.

2. The Travel Routes: Who Visits Whom?

The virus spreads mostly because pigs are shipped from one farm to another (like people taking a bus or train). The researchers mapped out who visits whom and how fast.

• The Upper Midwest is the Hub: It's the central train station. It sends virus variants to almost every other neighborhood, and it also receives visitors from everywhere. It's the busiest intersection in the city.
• The Great Plains and Lower Midwest are Fast Connectors: If a new virus style pops up in the Great Plains, it travels to the Upper Midwest very quickly (about 1.3 years). It's like a high-speed train between two major cities.
• The Atlantic Seaboard is the Slow Lane: This region is on the coast. When a new virus style emerges here, it takes much longer (2–3 years) to travel inland to the rest of the country. It's like a slow ferry crossing a wide ocean.
• The Northeast is a Bit Isolated: It mostly receives visitors but rarely sends its own unique virus styles out to the rest of the country.

3. The "Time Lag" Game

The most important question the authors asked was: "How long does it take for a new virus style to travel from its birthplace to a new neighborhood?"

They found that the speed depends entirely on the route:

• Fast Corridors: Moving between the interior regions (like Great Plains to Upper Midwest) is fast. It's like driving on an open highway.
• Slow Corridors: Moving from the coast (Atlantic Seaboard) to the interior is slow. It's like driving through heavy traffic and winding country roads.

4. Why Does This Matter?

Think of this research as creating a traffic map for a pandemic.

• Early Warning Systems: If a dangerous new virus style appears in the "Fast Corridor" (like the Great Plains), officials know it will hit the Upper Midwest in just over a year. They can sound the alarm early.
• Better Defense: If a virus comes from the "Slow Lane" (the coast), there is more time to prepare, test, and strengthen biosecurity before it arrives.
• Smarter Models: By knowing exactly how long the "bus ride" takes for the virus, scientists can build better computer models to predict future outbreaks.

The Big Takeaway

The virus is constantly evolving and moving, but it doesn't move randomly. It follows the "roads" of the swine industry. By understanding which neighborhoods are the fast lanes and which are the slow lanes, farmers and veterinarians can stop the virus in its tracks before it spreads to the whole city. It's about knowing the traffic patterns to avoid the jam.

Technical Summary

1. Problem Statement

Porcine Reproductive and Respiratory Syndrome Virus type 2 (PRRSV-2) is a major threat to the global swine industry, causing over $1.2 billion in annual losses in the USA alone. The virus is characterized by rapid genetic diversification, with dozens of new variants emerging annually. While previous phylogeographic studies have mapped long-term dispersal pathways at the sub-lineage level (spanning decades), there is a critical knowledge gap regarding the temporal dynamics of specific genetic variants. Specifically, it remains unclear how quickly new variants spread from their region of origin to other regions after emergence. This lack of temporal resolution hinders the ability to predict outbreak trajectories, calibrate epidemiological models, and implement timely biosecurity measures.

2. Methodology

The study utilized a comprehensive dataset and employed two complementary analytical frameworks to assess spatiotemporal dynamics.

Data Source and Curation:

• Dataset: 14,835 ORF5 gene sequences retrieved from the Morrison Swine Health Monitoring Project (MSHMP), representing >60% of the US breeding population.
• Timeframe: 2015–2024.
• Filtering: Sequences were classified into 156 distinct variants using the PRRSLoom-Variants tool. Vaccine-like variants and sequences assigned to undetermined sub-lineages were removed. Variants were defined based on an average genetic distance of ~2.5%.
• Regions: Five major swine-producing regions were analyzed: Upper Midwest (UM), Lower Midwest (LM), Atlantic Seaboard (AS), Northeast (NE), and Great Plains (GP).

Analytical Approaches:

1. Diversity Assessment:
   • Used Hill numbers ($q=0, 1, 2$) to quantify variant richness, abundance evenness, and dominance.
   • Constructed rarefaction and extrapolation curves to evaluate sampling sufficiency.
2. Time-to-Dispersal Event Analysis (Survival Analysis):
   • Definition of Emergence: A variant was considered "emerged" in a region upon detection of at least five sequences.
   • Metric: Calculated the time lag (in years) between the fifth detection in the origin region and the fifth detection in a recipient region.
   • Censoring: Applied right-censoring for variants emerging in the last three years (insufficient follow-up) and left-censoring for older variants that did not spread.
   • Tools: Kaplan-Meier estimators and log-rank tests using the survival package in R.
3. Phylogeographic Waiting Time Analysis:
   • Subset: Focused on 15 variants with >100 sequences and presence in $\ge$2 regions.
   • Method: Reconstructed time-scaled phylogenies using BEAST 1.10.4 (GTR+I+G4 model, uncorrelated log-normal relaxed clock, Skygrid coalescent model).
   • Metric: Estimated "waiting times" (time lag between emergence/introduction and subsequent dispersal) using Posterior Analysis of Coalescent Trees (PACT). This approach captures complex routes, including indirect transmission through intermediate regions.

3. Key Contributions

• High-Resolution Variant Tracking: Shifted the focus from broad sub-lineages to specific genetic variants, providing a finer scale for tracking emergence and spread.
• Quantification of Dispersal Lags: Provided the first empirical estimates of the time required for PRRSV-2 variants to traverse between major US swine regions (ranging from ~1 to 3 years).
• Dual-Method Framework: Demonstrated the value of combining survival analysis (high sample size, direct observation) with structured coalescent phylogeography (lower sample size, but capable of reconstructing complex transmission pathways and indirect spread).
• Regional Connectivity Mapping: Identified specific "fast" and "slow" corridors of viral transmission, highlighting asymmetries in inter-regional connectivity.

4. Key Results

A. Regional Diversity Profiles:

• Upper Midwest (UM): Highest variant richness ($n=123$) but low evenness; dominated by a few variants (e.g., 1C.5 accounted for ~30% of sequences).
• Lower Midwest (LM): Moderate richness ($n=47$) but the highest evenness and dominance diversity, indicating a more balanced distribution of variants.
• Other Regions: Atlantic Seaboard (35), Northeast (45), and Great Plains (38) showed intermediate diversity profiles.

B. Emergence and Dispersal Patterns:

• UM as a Hub: The UM was the primary source of local emergence (62 variants) and a major recipient (24 introductions). It exported variants to all other regions.
• Asymmetry:
  • Atlantic Seaboard: Functioned primarily as a source, exporting to most regions but receiving few introductions.
  • Northeast: Received introductions from all other regions but acted as a minor source.
  • Great Plains & Lower Midwest: Exhibited balanced exchange patterns.

C. Dispersal Time Estimates:

• Fast Corridors (Interior Regions): Variants spread rapidly between interior regions.
  • Great Plains $\to$ Upper Midwest: ~1.3 years (median).
  • Lower Midwest $\to$ Upper Midwest: ~1.15 years (median).
• Slow Corridors (Coast to Interior): Spread from coastal regions took significantly longer.
  • Atlantic Seaboard $\to$ Interior regions: ~2.4–2.9 years (median).
• Overall Median: The median time for a variant to spread from its origin to a new region was approximately 1.53 years (Time-to-Dispersal) and 0.86–1.45 years (Phylogeographic Waiting Time), depending on the specific variant and route.

D. Phylogeographic Insights:

• Reintroduction: Some variants (e.g., 1A.28, 1C.2) showed evidence of reintroduction to their region of origin after spreading elsewhere.
• Indirect Spread: The phylogeographic approach revealed transmission routes not captured by direct emergence models (e.g., Atlantic Seaboard $\to$ Northeast $\to$ Upper Midwest).

5. Significance and Implications

• Surveillance Optimization: The identification of rapid dispersal corridors (e.g., interior Midwest) suggests that surveillance and biosecurity efforts should be intensified in these high-connectivity zones to detect variants early.
• Early Warning Systems: The estimated time lags (1–3 years) provide a temporal window for implementing risk-based alert systems. If a high-risk variant emerges in a source region, stakeholders in connected regions have a predictable window to prepare before the variant likely arrives.
• Model Calibration: These empirical dispersal times offer critical parameters for refining epidemiological models, moving beyond static assumptions to dynamic, time-dependent spread scenarios.
• Biocontainment Strategy: Understanding that variants can persist in a region for years before spreading (or spread rapidly depending on the route) supports the development of targeted, region-specific containment strategies rather than a "one-size-fits-all" approach.

In conclusion, this study establishes that PRRSV-2 variant spread is not uniform; it is heavily influenced by regional connectivity, with interior regions acting as rapid conduits and coastal regions exhibiting slower, more asymmetric transmission dynamics. These findings are vital for improving disease preparedness in the US swine industry.