J. William Bell
 A Flower's Family Tree 1 2 3 4
 Gene data allow researchers to recover the evolutionary history of plants, but even the smallest dataset can require impossibly large computations. Using an Alliance Linux cluster and newly designed software, a team from the University of New Mexico and the University of Texas have increased the speed of the process millionfold for one family of plants.

If you're looking for extreme diversity, consider bluebells, officially the Campanulaceae family. The 2,000 species of these plants with bell-shaped flowers show amazing variety. They live everywhere on Earth except the Sahara, Antarctica, and the northern extremes of Greenland. Some are annuals, and others are perennials. Despite their common name, the flowers can be blue, purple, red, or yellow. The bluebells of North America typically grow close to the ground, while Asian species can grow as tall as eight feet.

As fascinating as that diversity is, it's not the sort of thing that computational scientists usually get excited about. Uncovering how that diversity came to be has captured the attention of a team of researchers at Alliance partner University of New Mexico and the University of Texas, though. Using the 512-processor LosLobos Linux Pentium III supercomputing cluster at the Albuquerque High Performance Computing Center, the team has created a phylogeny reconstruction—or evolutionary history—of 12 bluebell species, predicting all of the steps that take these species back to a single common ancestor. To meet the challenge, they created a whole new piece of software known as GRAPPA.

"In our context, we can answer the question of why we do this work by saying, 'We just want to know, and we like a challenge.' But phylogeny reconstruction has very significant implications to pharmaceutical design and in other industries," says Bernard Moret, a computer science professor at the University of New Mexico.


 1 2 3 4 up