In recent years, scientists have increasingly relied on using mitochondrial DNA to make comparisons among mammals and thereby link those that are related on the evolutionary tree.
But Killian said that mitochondrial DNA provides misleading results for a variety of reasons. Most importantly, it requires more human input to decide which information is fed to the computer, thereby raising the risk of human bias. In fact, when the data were given to three different laboratories for analysis, they generated three different family trees, Killian said.
Secondly, the "bootstrap value" a measure of the relatedness of genes -- which the computer assigns to the accuracy of its predictions has been quite low in mitochondrial studies on mammalian associations, hovering between 40 percent and 60 percent.
In contrast, the Duke study generated 97 percent to 100 percent bootstrap support in its nuclear gene comparisons of animals representing all three mammalian groups. Mammals they studied included the platypus, echidna, opossum, wallaby, hedgehog, mouse, rat, rabbit, cow, pig, bat, tree shrew, colugo, ringtail lemur and humans.
"This is the first molecular evolutionary study that seriously and powerfully says the paleontologists have been right all along in grouping mammals the way they did," said Killian. "It turns out that common sense is correct."
The Duke scientists realized the power of using large nuclear genes for molecular evolutionary studies when they were sequencing a gene called the insulin-like growth factor II receptor (IGF2R) to determine its evolutionary hist
'"/>
Contact: Richard Puff
richard.puff@duke.edu
919-684-4148
Duke University Medical Center
1-Jul-2001