iSGTW Feature - The path more travelled by: grids help track human migration

Feature - The path more travelled by: grids help track human migration

Nicolas Ray uses Approximate Bayesian Computation-developed in 2002-to track the migration of humans through the centuries.
Images courtesy of Rakesh Rampertab

Today, families interested in an intercontinental move will probably go by plane, train or automobile. Several thousand years ago, adventurers had a harder time of it.

Thanks to the plucky, pioneering efforts of early families, humans have managed to explore and settle every habitable region of the globe.

Surviving the "bottlenecks"

Two ancient migrations have particularly affected the shape of future generations: the "Out of Africa" event and the North to South colonization of the Americas. These migrations left traces in our genes still observable today.

Described as "bottlenecks"-events where only a few individuals get through-the migrations may have caused a drastic reduction in population size and a corresponding drop in genetic diversity.

To estimate the strength and timing of these bottlenecks, specialists like Nicolas Ray at the Computational and Molecular Population Genetics Lab at the University of Bern, Switzerland, try to reconstruct past human movements using the genetic diversity of current populations.

New technologies track old paths

Ray says that new statistical tools, larger data sets and the robust computing power of computing grids mean he can now examine human migration in greater detail than ever before.

"This is a very exciting field right now-we have so much to study," he says. "The technology required to obtain genetic data is much cheaper now. We can acquire a large number of genetic markers in many individuals, and obtain data much more rapidly than before."

The Computational and Molecular Population Genetics Lab has been involved in the Enabling Grids for E-sciencE project since 2005, providing Ray and his colleagues with access to previously impossible levels of computing power.

"For a given project, where four alternative models are compared, I would typically run about 20 million simulations," says Ray. "This might take a single computer five years; it is done in a couple weeks on the EGEE grid."

The SPLATCHE program uses genetic data to simulate patterns of human dispersal. This simulation shows a range expansion with specific long-distance dispersal along the west coast of America.
Images courtesy of Nicolas Ray (left) and (right)

SPLATCHE-ing out

With funding from the Swiss National Science Foundation, Ray and his colleagues are now working on SPLATCHE, a program designed to integrate environmental factors-like vegetation patterns, topography and the locations of fresh water and coastlines-in simulations of Amerindian migration.

The ultimate goal is to create a spatially explicit demographic model that accurately describes the genetic distribution found in the real world.

The coastal route

Ray recently looked at colonization of the American continent, using genetic patterns to determine migratory paths.

"Usually we assume that populations connected by migration are more genetically similar," he explains. "We observed this pattern in coastal populations, so Amerindians probably used coastlines as migratory corridors, probably because these landscape features tend to be richer in resources and easier to cross than dense forests."

Ray's recent work will appear in a paper in press at PLoS Genetics, along with the work of 26 other co-authors who also worked to characterize the genetic diversity of Amerindians.

- Danielle Venton, EGEE