Rather than focus on errors and alterations in DNA sequence, the new technique highlights variations in the number of copies of a particular gene. Additional copies of a gene may lead to overproduction of that gene's protein, and this may affect both easily identifiable traits such as body size or more difficult-to-discern traits such as cancer risk.

Scientists report in Public Library of Science Genetics that they refined an analysis technique to assess variations in gene copy number in 20 different mouse strains. According to the paper's lead author, this budding area of study is likely to have wide-ranging implications for scientists' understanding of how DNA variations contribute to human health and illness.

"Right now, our results and other early assessments of human and other mammalian genomes are suggesting that about 10 percent of the genome features copy number variations," says Timothy Graubert, M.D., assistant professor of pathology and immunology and of medicine. "That's a huge number. As a percentage of the genome, variations in gene copy number could explain more person-to-person variability than the single-letter changes in the genetic code known as SNPs [single nucleotide polymorphisms]."

Graubert's lab uses human samples and mouse models to study leukemia, cancer that occurs in the bone marrow cells that make blood cells. Using Nimblegen's technique for assessing gene copy number, they identified approximately 80 variations in the number of gene copies in each of the mouse genomes. Graubert will incorporate the results into his lab's search for genetic factors that protect against or increase susceptibility to leukemia.

Much of the analytic work was led by graduate student Patrick Cahan and postdoctoral fellow Deepa Edwin, Ph.D. The 20 mouse strains were previously selected by the Mouse Phenome Project, which is assembling a database of how changes in mouse DNA affect mouse characteristics. The project is headquartered at Jackson Laboratory in Bar Harbor, Maine. For their analysis, researchers compared the genome of each of the 20 mouse strains against that of the prototypical research mouse strain, C57BL/6J.

"That's the 'plain vanilla' mouse genome," Graubert explains. "Just like the reference human genome sequence that is used to identify genetic differences between human individuals, the C57BL/6J mouse genome is the one we understand best and the standard against which other mouse genomes can be compared." Nimblegen's technique for rapid analysis is known as oligonucleotide array comparative genomic hybridization.

"The copy number variants we describe in this paper are numerous and fairly large - they vary in length between two thousand and two million DNA base pairs," Graubert says. "Datasets this large require a lot of analysis to be sure that what you're seeing is real, so we really worked hard to prove that these gene copy number variations are real and validated many of them using other technologies."

Graubert is working with Nimblegen to conduct a follow-up analysis of gene copy number variation in the mouse strains using an even more sensitive version of the technique. They are also testing how changes in gene copy number are reflected in RNA, the order slips for assembly of a gene's protein that are copied from DNA.

"The prediction is that if you have a higher gene copy number count, you'll see more RNA from that gene," he says. "But we need to test that on a genome-wide scale."

wustl