Biologists can now predict if a certain species can go extinct–by using Google’s algorithm.
It is well known that extinction of certain species can cause the entire ecosystem to collapse. However, due to the complexity of ecosystems, it is not possible to accurately identify the most important species. Now, computation biologists have adapted Google’s PageRank algorithm to predict extinction of species and subsequent collapse of the entire ecosystem. They have succeeded in creating a model that is very accurate.
Using a modified version of PageRank, researchers now able to identify which species’ extinction in the food web will lead to large-scale extinction of species. Unlike other models that try to predict extinction, Google’s solution not only takes the number of connections between species but also their relative importance into consideration.
When researchers compared the Google algorithm with existing models for predicting ecosystem collapse, they observed that the new model outperformed the old ones in all the 12 food webs they applied it to. In each case the model predicted total ecosystem collapse using the least number of species’ extinctions.
The PageRank algorithm is circular. Food webs, on the other hand, are unidirectional. This poses a problem for the PageRank algorithm to be applied to food webs. Researchers solved this problem by adding a root node which is based on the logic that all species contribute to the food chain through their excrement and eventual decay.
Other prediction models often underestimate the importance of species that are at the lower parts of the food chain. The new solution will encourage biologists to get a better picture of species extinctions.
For ecosystems that are on the verge of collapse such as certain marine ecosystems, a network based approach will be the best option to conserve them and prevent further damage.
Authors bio is coming up shortly.