Guide Dogs

Continued from page 2

It's much easier to uncover disease mutations in dogs, though. Two random rottweilers are far more closely related than two random humans. If the two rottweilers both develop bone cancer, which is common in their breed, the disease will probably be caused by several mutations carried by both dogs. But two humans with bone cancer are less likely to have disease-causing mutations in common. "For the cancer studies, we've had about 50 to 100 sick dogs," says Lindblad-Toh. You'd need a much larger sample of humans--thousands of patients and healthy people--to see similar patterns. Understanding which mutations cause a disease in dogs helps researchers figure out where to look for disease-causing mutations in humans.

Nearly two years after the publication of the dog genome, its promise as a tool for studying human disease is beginning to be borne out. Lindblad-Toh says that her group hopes to identify mutations involved in osteosarcoma, a rare but deadly cancer in adolescence. "It's going to be very exciting within the next year to see whether applying [insights from studying dog-gene mutations] back to human patients with the same diseases will also show you mutations in the same genes," she says. "My prediction is yes. I think if you can find strong risk factors in people by using dogs, that would be a great benefit."

Beyond Genes
"When we analyzed the mouse and human genomes, we found that they are 5 percent functional," says Lindblad­-Toh. That is, 5 percent of each creature's genome is very similar to 5 percent of the other's, suggesting that the related sequences must be serving some purpose. The dog genome turned out to have the same 5 percent that overlaps in humans and mice, which confirms that it's not just a coincidence. Most of that 5 percent, however, is not genes.

Our genome is made up of 46 chromosomes, which are distinct, long chains of organic compounds known as A, T, C, and G--the DNA letters. Sequencing the genome means figuring out which letter, or base, is at every point along each chain. Genes are stretches where the bases spell out a code that can be translated into proteins; they make up 1.5 percent of the genome. But an astonishing 95 percent of our genome is mutation-prone gobbledygook, the genetic equivalent of the sentences you can make by banging your head on a keyboard in frustration.

Now biologists are trying to characterize the functional parts of the genome that are not genes--regions they believe play an important role in regulating genes. They want to identify these regions and find out what elements they contain, how the regions are organized, and how they work.

The best way to find the regulatory elements amid the gobbledygook is to look at what's conserved--what stays the same--across multiple species. Comparing genomes is "almost like the Rosetta stone," says Lindblad-Toh. With the same message carved in the Greek alphabet, hieroglyphics, and demotic script (a sort of cursive hieroglyphics), the Rosetta stone made it possible to decipher the last two writing systems. In genomes, similarly, "you have a string of letters; anything that's important will stay the same."