Tuesday, May 17, 2005
Gene "Archeology" Gets Easier Using Carnegie Mellon University Software
PITTSBURGH—Comparing genomes of different species can tell you when new genes evolved and what they do for their respective hosts. But pinpointing the ancestry of any given gene is a complex computational task. Now, powerful new software makes gene "archeology" considerably easier, reports a team of investigators at Carnegie Mellon University. The scientists, who are making this software publicly available for the first time, presented their findings Monday, May 16, at the RECOMB meeting in Cambridge, Mass.
"Our software considers thousands of evolutionary scenarios to obtain the best guess about when a given gene arose," said Dannie Durand, associate professor of biological sciences at the Mellon College of Science (MCS) at Carnegie Mellon. "This software can help scientists use evolutionary clues to understand what genes do in modern organisms."
For example, using the program, called Notung, investigators now can identify genes that arose very recently. New genes can appear when one gene is duplicated and the second copy evolves to take on a different role inside the cell. In vertebrates, for instance, extra gene copies led to features like paired appendages. New genes often arise in response to environmental stresses, such as pesticides or drugs, and they help the organism to fight back. Over time, gene copies also can be lost.
"You could use this information to plan additional gene studies or suggest strategies for circumventing drug and pesticide resistance in parasites," Durand said. "Other uses could include identifying potential detoxification enzymes for bioremediation and designing breeding programs that would enhance pest resistance in cash crops. "
Already, Notung is helping to bridge studies of molecular evolution with laboratory research, according to Durand. Her team has collaborated with a number of scientists studying different multi-drug resistance (MDR) and detoxification genes in several species. One of Notung's strengths is its ability to use information about gene duplications and losses.
Until recently, most researchers created evolutionary trees of different gene families by comparing gene sequence data alone. These data reflect small-scale (microevolutionary) events in the form of gene mutations. But equally important in governing gene evolution are gene duplications and losses—considered large-scale (macroevolutionary) events.
Notung is the first to incorporate both sets of events to build a tree, using a powerful algorithm that finds all possible trees with the fewest gene losses and duplications.
"Using our method, we can incorporate a broader collection of evidence in the reconstruction of a gene family tree," Durand said.
Notung's user graphical interface makes it easy for the user to manipulate and interpret the output. The user interface also allows scientists to analyze trees that are too big to deal with manually. Durand has used Notung to study gene trees from families with as many as 350 member genes.
Through a series of steps, users can rearrange a massive gene tree with hundreds of branches and no apparent structure so that it becomes a highly organized tree with a distinct root and a clear, concise history of duplications.
"These parsimonious trees give us the simplest possible explanations for how a gene evolved, and simpler is usually better," Durand said.
Durand is collaborating with a team at the University of Puerto Rico in searching for MDR genes that confer drug resistance in malarial parasites. The same genes confer pesticide resistance in plants. Durand's team used Notung to find separate clusters of very old MDR genes and more recent MDR genes, suggesting that the latter MDR genes may have arisen in response to environmental selection.
In another initiative, she is working with investigators at the Pittsburgh Supercomputing Center in studies of glutathione S transferases. Some members of this superfamily of detoxification enzymes are thought to protect us from PCBs and other pollutants.
By combining ecological and biochemical data with these analyses, molecular evolutionists could identify duplicated genes that are potential sites of rapid change in response to environmental forces. And identifying these genes would be the first step in developing targeted therapies designed to thwart pesticide or drug resistance, according to Durand.
MCS maintains innovative research and educational programs in biological sciences, chemistry, physics, mathematics and several interdisciplinary areas. For more information, visit www.cmu.edu/mcs.
By: Lauren Ward