Genetics Times, April 12, 2008
A team of researchers from Washington University in St. Louis and the Israeli Institute of Technology (Technion) in Haifa has developed a technique to detect the ancestry of disease genes in hybrid, or mixed, human populations.
The technique, called expected mutual information (EMI), determines how a set of DNA markers is likely to show the ancestral origin of locations on each chromosome. The team constructed an algorithm for the technique that selects panels of DNA markers in order to render the best picture of ancestral origin of disease genes. They then tested the algorithm to show that it is more powerful and accurate than standard algorithms that are currently used.
The result is easier identification of inherited genes that cause diseases in people of mixed races, which researchers call “population admixture.” Nephrologists, for instance, have noted that African-Americans are far more likely than Europeans to die rapidly of end-stage, progressive renal failure due to kidney disease. Many African-Americans, though, have genes that originated in Europe due to ethnic mixing. The technique helps researchers isolate the genetic causes of disease by detecting from which continent the recurrent disease genes originated.
It’s a good bet, Templeton said, that the disease genes are highly likely to have emerged from Africa, as African-Americans have shown the tendency to die more quickly of the disease.
The technique and algorithm apply beyond this particular disease, Templeton added.
“Our novel approach extends previous methods by incorporating knowledge on population admixture, drawing a more precise picture of the mosaic of ancestries along an individual’s genome,” said Sivan Bercovici, Templeton’s colleague at Technion and primary author of a research paper published in Genome Research.
The researchers analyzed DNA from 575 cases of African-Americans with end-stage progressive renal failure and compared it to controls that did not have the disease. They came up with a panel of approximately 2,000 genetic markers. Enough, Templeton said, “to cover the whole genome.”
A paper discussing the technique and algorithm is published in the current issue of Genome Research 18, 661-667.