One of the surprises that's come out of recent genome studies has been the significance of variations that affect large chunks of the genome, instead of single bases. At the base level, humans are well over 99 percent identical. But, when the genome structure is looked at, changes in the copy number - extra or missing copies of a section of the chromosome - cause a significant amount of variation between individuals. In some cases, these copy number variations (CNVs) may be associated with diseases.

If CNVs are this significant, it's probably worth figuring out how often they occur, but the human population isn't very suited to that; it can be hard to tell which CNVs are ancient, and which are of recent origin. To avoid this problem, researchers that have just published in Nature Genetics turned to the laboratory mouse. Strains of lab mice have been inbred for a reasonably well-defined period, so any variations among them should be recent in origin. They specifically worked with variations of the C57Bl/6 strain that have been diverging for less than 1,000 generations.

The surprise was that a lot of variation arose during that relatively short time span. Even in a coarse analysis, the researchers were able to find 23 CNVs that affected two or more consecutive DNA markers, and another 65 that affected one marker (markers were spaced roughly 17 kilobases apart). The other surprise was how often some regions were affected. Of the 38 CNVs they looked at in detail, 18 arose more than once, and 10 of those underwent copy number changes more than twice.

The data suggest that certain parts of the genomes are "hotspots" that both undergo change frequently, and produce changes that are well-tolerated by the organism. These hotspots undergo changes up to 10,000 times more frequently than quiet areas of the genome, and can undergo multiple, successive changes. The fact that the organism appears to tolerate these changes isn't due to an absence of genes in the CNVs. In the 18 mentioned above, there were a total of 43 genes, including some involved with reproduction, immunity, and brain function.

There's been an idea floating around for a while that suggests that genomes evolve to the point where they work well with evolution. A genome that, by chance, winds up with genes that are sensitive to dose effects in a region that's stable is more likely to be be inherited. In the opposite case, where different doses of a gene might help an organism adapt to different environments, having that gene located in an unstable area might be selected for. The new data doesn't directly address this proposal. But it does find that there unstable areas of the genome that are likely to undergo major changes within the span of less than 100 years, which seems to be a prerequisite for the proposal to be taken seriously.