PageRank In Biological Systems

[via MALIA WOLLAN The Ninth Annual Year in Ideas]

Every
species — be it earthworm or great white shark — is entwined in a vast
and complicated system of predator and prey called a food web. To
determine which species in a food web are most important to the
survival of the larger ecosystem, scientists design computer programs
to model how the extinction of a given species would affect the other
species in the system. This year, two scientists announced that they
had found an unexpectedly useful tool for this purpose: the seminal
Google search algorithm.

Stefano
Allesina of the University of Chicago and Mercedes Pascual of the
University of Michigan began with a simple hunch. Google’s search
engine uses an algorithm called PageRank to identify the most important
Web sites on a given topic by analyzing links: a Web page is important
if other important pages link to it. How different is this, really,
Allesina and Pascual wondered, from an ecosystem, in which a species is
important if other important species eat it?

Allesina
and Pascual borrowed Google’s PageRank algorithm and modified it to
model ecosystems in the natural world. As they explained in September
in the journal PLoS Computational Biology, the modified algorithm was
more efficient than existing ecosystem-extinction models at identifying
which species’ extinction would cause the greatest number of

ILLUSTRATION BY JAN KALLWEJT

other species in the food web also to go extinct. “Our algorithm is faster and computationally simpler,” Allesina says.

The
PageRank algorithm could be useful in analyzing other networks too. The
world features countless interconnected systems ranging in size from
the minuscule (metabolic networks within a single cell) to the immense
(international financial markets). After publishing the paper, Allesina
received e-mail messages from dozens of researchers interested in
adapting the PageRank algorithm. “PageRank is a technique for finding
hidden flows in huge quantities of data,” says Yonatan Zunger, a Google
software engineer who used to work on search technology and who
contacted Allesina after seeing his research. “There are all kinds of
networks. PageRank is enormously applicable beyond the Web.”

Another example where we find out that what works on the Web is a reflection of things that work in the larger world, and vice versa.