It’s always amazing to see the lengths people will go to get a photograph of a cheetah, lion or African hunting dog. For some, it never occurs to them that as they stumble out of their car to adjust tripods and flashes, a rare predator is quietly considering eating them. In the circle of life there are many players and while there’s no shortage of people, there are unfortunately thousands of endangered animals and plants. Biologists have been looking for ways to determine the most important species based on the number of feeding-related interactions, and surprisingly the Google search algorithm is playing a part.
There are almost 4706 animals and 4295 plants on the worldwide endangered animals list and it’s difficult to determine which extinctions would be the most devastating. Food web biologist Stefano Allesina was featured in Wired magazine for using a modified version of Google’s search algorithm PageRank to determine just that.
Allesina and Mercedes Pascual first wrote about utilizing PageRank in their recently published Googling Food Webs: Can an Eigenvector Measure Species’ Importance for Coextinctions?.
Said Allesina, “In PageRank, you’re an important website if important websites point to you. We took that idea and reversed it: Species are important if they support important species.”
The biologists use PageRank by multiplying the rate of extinction for one species and determining the rate at which others are affected. The animals and plants that link to a large number of connectors are considered keystones. In other words, if you’re an integral player in either preparing or being dinner for a large number of species, you are the keystone for an entire ecological network.
Because many species need others to survive, the worst possible case for an extinction is a domino-style co-extinction effect. Allesina and Pascual aim to determine these worst scenario cases and encourage conservationist intervention. For more information read the entire report at PLoS Computational Biology.
Photo Credit: Michael Gäbler
