Virtual ecosystem sheds new light on how species develop

Were Charles Darwin alive today, he’d probably be very interested in working with Robin Gras.

Without such modern technology as high performance computing, Darwin developed his theory of natural selection, an explanation of how all species of life on Earth have descended from common ancestors.

What science still can’t explain, however, is the actual process of speciation, or exactly how various species emerge from single gene pools. A Canada Research Chair in Heuristics for Bioinformatics and a computer science professor, Dr. Gras has developed a virtual world that attempts to shed light on that process. It’s a world he hopes will be embraced by biologists, geneticists and ecologists looking to answer the questions that remain largely unanswered since Darwin published On the Origin of Species in 1859.

“The big question is How do species emerge?” says Gras. “How is it that we have different species? What are the mechanisms that make that happen?”

Gras and his team have developed a virtual world, a complex ecosystem that runs on a high performance computing network called Sharcnet. It’s made up of virtual organisms, all programmed with genetic information and the ability to flee, fight, migrate, hunt, feed and reproduce, and it runs 24 hours a day.  Gras can check in periodically to see if new species have emerged, or if others have gone extinct.

In one of his latest experiments – the details of which were published in the academic journal Proceedings of the Royal Society B: Biological Sciences – Gras introduced obstacles in anywhere from one to 10 per cent of the one million “cells” in the virtual world. In the real world, those obstacles might be forests, rivers, or mountain ranges – anything that might reduce the mixing of certain individuals within a particular population.

Over about 80,000 hours of computing time, he repeated the experiment 60 times, introducing obstacles and monitoring how various clusters of organisms evolved over about 2,000 to 3,000 generations. The results proved that introducing obstacles actually increased speciation, with about 10 million new species being introduced, and more than 20 billion individuals born since the program began.

“The consequences were very clear,” said Gras, a native of France who came to Windsor several years ago from the Swiss Institute of Bioinformatics in Geneva, Switzerland. “We observed much more speciation by introducing these obstacles, even though they don’t completely isolate sub-populations from one another.”

Part of the explanation for increased speciation, Gras says, may have to do with reduced transferring of genetic material among populations, or “gene flow,” created by those obstacles. Reduced gene flow, he says, leads to more speciation.

“Increased gene flow leads to more mixing among the populations, while decreased gene flow actually leads to more differentiation among them,” he explained.

Gras, who would normally publish in computer science journals, hopes having his findings appear in a biology journal will introduce his to work to a more diverse group of scientists interested in experimenting in his virtual world with organisms specifically programmed to replicate the characteristics of those they study.

He has already partnered such UWindsor campus collaborators as Melania Cristescu, a professor in the Great Lakes Institute for Environmental Research who studies the ecology and evolution of aquatic invasive species and co-authored his latest paper, and Stephanie Doucet, a Biological Sciences professor who studies the ecology of various bird species.

“We’ve been recognized as being able to provide real insights for biologists,” he said. “And I have almost as many biology students in my group now as I do computer science students.”