Digital World Reveals Architecture of Evolution

August 7, 2006
Digital World Reveals Architecture of Evolution
Diagram of a scale-free network that contains components with a highly diverse level of connectivity. Some components form highly interconnected hubs, while other components have few connections, and there are many levels of interconnectivity in between. Scale-free networks are pervasive in biology. Computer simulations at the University of Chicago show that scale-free networks are able to evolve to perform new functions more rapidly than an alternative network design. Image courtesy of Panos Oikonomou and Philippe Cluzel, University of Chicago

The architecture that pervades biological networks gives them an evolutionary edge by allowing them to evolve to perform new functions more rapidly than an alternative network design, according to computer simulations conducted at the University of Chicago. The finding is published in the August issue of the journal Nature Physics.

Scientists have found the same intricate network architecture of evolution just about everywhere they look. This architecture characterizes the interaction network of proteins in yeast, worms, fruit flies and viruses, to name a few. But this same architecture also pervades social networks and even computer networks, affecting, for example, the functioning of the World Wide Web.

"These results highlight an organizing principle that governs the evolution of complex networks and that can improve the design of engineered systems," wrote the article's co-authors, graduate student Panos Oikonomou and Philippe Cluzel, Assistant Professor in Physics.

This organizing principle is what scientists call a "scale-free design." A diagram of this design resembles the route maps of airline companies. "You have hubs that are highly linked with airplanes going in and out of those hubs," Oikonomou said. But then smaller airports also exist that have far fewer connections, and there are various scales of connections in between.

Digital World Reveals Architecture of Evolution
Diagram of a random network containing components that all have approximately the same number of connectivity. Computer simulations at the University of Chicago show that random networks evolve more slowly and in spurts when compared to scale-free networks. Image courtesy of Panos Oikonomou and Philippe Cluzel, University of Chicago

Oikonomou and Cluzel initiated his project to find out if network design conferred any kind of evolutionary advantage. They created a Darwinian computer simulation to compare the evolvability of this scale-free network design with a more random design in which all network components have approximately the same number of connections. They programmed this computer world to have random mutations and natural selection operate on its digital populations, then compared how long it took the two types of networks to evolve the ability to perform a new task.

The populations organized in scale-free networks evolved rapidly and smoothly, while randomly organized networks evolved slowly and in spurts following a succession of rare and beneficial random events.

"They followed drastically different evolutionary paths," Cluzel said.

Cluzel plans to conduct laboratory experiments on bacteria to test the validity of the organizing principle he and Oikonomou have identified via their simulations.

Their goal was to better understand biological evolution, but social and economic networks also display a scale-free architecture. "These networks can be people, they can be molecules, they can be whatever you like," Cluzel said.

In the engineering arena, the findings of Oikonomou and Cluzel indicate that using a scale-free architecture rather than a random network design in new electronic devices will likely produce better results.

Many engineers specialize in finding the best way to train artificial neural networks using an array of powerful computer programs. "The problem is that it usually takes time to find which connections you have to change within a network to achieve your target function," Cluzel said.

Noted Oikonomou: "If you start with an architecture that is scale-free, maybe this would give you better results, quicker."

Source: University of Chicago

Explore further: Researchers use Titan to accelerate design, training of deep learning networks

Related Stories

Walk this way—a better way to identify gait differences

November 9, 2017

Biometric-based person recognition methods have been extensively explored for various applications, such as access control, surveillance, and forensics. Biometric verification involves any means by which a person can be uniquely ...

Recommended for you

Information engine operates with nearly perfect efficiency

January 19, 2018

Physicists have experimentally demonstrated an information engine—a device that converts information into work—with an efficiency that exceeds the conventional second law of thermodynamics. Instead, the engine's efficiency ...

Team takes a deep look at memristors

January 19, 2018

In the race to build a computer that mimics the massive computational power of the human brain, researchers are increasingly turning to memristors, which can vary their electrical resistance based on the memory of past activity. ...

Artificial agent designs quantum experiments

January 19, 2018

On the way to an intelligent laboratory, physicists from Innsbruck and Vienna present an artificial agent that autonomously designs quantum experiments. In initial experiments, the system has independently (re)discovered ...

0 comments

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Click here to reset your password.
Sign in to get notified via email when new comments are made.