If you shop online, this is a familiar scenario: You click on a product like a book, and the online merchant presents you with a list of related items. “If you like X, you might also like Y.” Behind the scenes, the merchant has assigned a series of attributes to each product. For instance, the new Sibley Guide to Birds is classified as nonfiction, recent, about nature, about birds, illustrated, a field guide, and so on. Based on that information, the merchant might suggest a different bird book, or a memoir about birding.
The newest scientist at the Cary Institute is using a similar algorithm to predict the outbreak of zoonotic diseases. Barbara Han has tapped into massive global databases that catalog various traits of mammals.
“Right now, with the Ebola outbreak, we know that the disease is hiding out in reservoir species. But if you’re talking about a continent, and you’re talking about rainforests that harbor hundreds of thousands of species, casting a wide net across that forest is just going to be ineffective at capturing the most likely reservoirs. So what you can do to start with is to narrow that scope a bit and train an algorithm, identify those species that have the highest probability of carrying something like Ebola, and then set a surveillance team loose, and say ‘Hey, I have 90% confidence that it’s going to be one of these fifteen species and they happen to occur in this type of forest, why don’t we go look there first.'”
By predicting what mammals might carry emerging infectious diseases and where these animals are located, Han is creating a tool that could be important for public health officials. Instead of reacting to frightening outbreaks, they may be able to predict and perhaps prevent them.
Full interview with Barbara Han, a disease ecologist at the Cary Institute of Ecosystems Studies
Photo, posted December 9, 2011, courtesy of John Tann via Flickr.
Earth Wise is a production of WAMC Northeast Public Radio, with script contribution from the Cary Institute of Ecosystem Studies.