Support Provided ByLearn More

This algorithm is predicting where a deadly pig virus will pop up next

A swine virus that appeared in the U.S. in 2013 has proven hard to track. But an algorithm might help researchers predict the next outbreak.

ByTaylor WhiteNOVA NextNOVA Next

Scientists still don’t understand exactly what triggers, or perpetuates, a porcine epidemic diarrhea virus outbreak. But thanks to an algorithm, that could be changing. Image credit: Liz West, flickr

In 2013, a deadly swine virus popped up on a handful of U.S. pig farms and quickly spread to 30 states. Over the next year, it wiped out around 10 percent of the U.S. pig population, and pork prices jumped.

This was the first appearance of porcine epidemic diarrhea virus (PEDV) in the United States. And although this particular outbreak was mostly contained in 2014, the virus remains a lingering threat to pig farms: An outbreak can cost a farm about $300,000 once dead pigs have been replaced and barns disinfected.

The virus, which does not affect humans, is spread through pig feces. Scientists still don’t understand exactly what triggers, or perpetuates, an outbreak. But that may be changing.

After the initial outbreak, North Carolina State University epidemiologist Gustavo Machado, along with colleagues from the University of Minnesota and Brazil’s Universidade Federal do Rio Grande do Sul, set out to study PEDV in sow farms across the United States. Using data on the virus’ prevalence in farms, pigs’ movements between farms, and environmental factors like wind speed and temperatures, they developed an algorithm to predict PEDV outbreaks. They used data from 2014 to “predict” outbreaks on a week-by-week basis, testing their results against what had actually unfolded each week.

In a study published in January in the journal Nature, Machado’s team showed that its algorithm was able to predict whether a PEDV outbreak would occur during a given one-week period with about 80 percent accuracy. They also determined that the movement of pigs between farms was the most important factor in the spread of the virus, but there may also be a risk of airborne spread.

“What we know is about 75 percent of the PEDV outbreak is reasoned by the movement of animals and the proximity with other farms,” Machado says. Pigs carrying PEDV can infect both their barn mates and pigs at nearby farms.

The virus infects the cells that line pigs’ small intestines, causing severe diarrhea and dehydration. The symptoms come on quickly and run a short—and sometimes severe—course. Piglets younger than a couple weeks old are most vulnerable, and often die from dehydration, diarrhea, and vomiting.

Kristen Bernard, a virologist at the University of Wisconsin-Madison’s School of Veterinary Medicine who was not involved in the latest study, says genomic sequencing studies of PEDV suggest it originated in from a viral strain in China in 2012. Experts suggest that PEDV may have spread to the U.S. through animal feed ingredients from China.

“It wouldn’t have been carried in the wind or water or anything like that,” Bernard says of the first U.S. outbreak. “And the fact that [initial PEDV infections] happened almost within days of each other, it’s very unlikely that it was transferred from farm to farm.” But, Bernard says, concrete evidence that PEDV came from China is still lacking.

Support Provided ByLearn More

A vaccine for the virus exists, but it isn’t 100 percent effective. The vaccine does boost the herd immunity of piglets, Machado says, but researchers have found traces of PEDV in piglet fecal samples from farms where vaccination occurred.

This could be because of how the vaccine is administered, Machado says: Veterinarians vaccinate sow mothers for PEDV rather than injecting piglets directly. Mothers pass antibodies against the virus to their offspring through their milk, which should protect them from infection. But this doesn’t always pan out.

“They have immunity, but maybe it’s not enough,” Machado says of the piglets. “What we really don’t understand is if the immunity is enough to protect for a new outbreak.”

Machado says additional measures, like installing air filters, can potentially reduce the circulation of the virus. Meanwhile, his lab continues to gather data weekly from 50 percent of U.S. sow farms to model the virus’ spread.

Based on the lab’s data, Machado predicts a 10 percent decrease in PEDV in the next two years.

The swine farms that participate in Machado’s study can receive an algorithm-based prediction of any upcoming outbreaks of PEDV and other diseases.And Machado’s team is developing an online portal in which farmers can check to see if their farm is within a PEDV outbreak-expected zone. The researchers believe this will encourage precautionary measures and enable farmers to address any PEDV outbreaks immediately.

Machado also hopes to figure out why PEDV outbreaks still occur in regions where virus immunity seems to be high. “We don’t know,” he says. “Maybe the virus is inside the farm circulating at low prevalence and then something happens, and then we have [an] explosion of the virus again.”

To find answers, he and his team aim to continue to study PEDV and model its spread for another five or 10 years. Currently, they are looking into other factors that might be associated with the virus spread, including the movement of feed mill trucks into farms.

Epidemiological models like Machado’s, Bernard says, extend beyond PEDV. Analyzing pigs’ movement and population as well as factors like food and water contamination, “can be really applicable to lots of different diseases,” she says. “That’s what’s useful about these kinds of epidemiological models,” she continues. “It’s to help us know what to do in the face of a new outbreak.”

Receive emails about upcoming NOVA programs and related content, as well as featured reporting about current events through a science lens.