NEW YORK (GenomeWeb News) – A relatively small set of quantitative trait loci seem to explain many physical differences between dogs, according to a study appearing in PLoS Biology last night.
Researchers from Stanford University, the National Human Genome Research Institute, and the University of California at Los Angeles assessed nearly 61,000 SNPs in more than 900 dogs from a range of wild and domestic dog groups. By mapping common genetic variation and finding signals of selection in the canine genome, the team was able to uncover loci underlying dozens of physical traits — from height and weight to snout length and ear shape.
In so doing, the researchers found a handful of sites in the genome that influenced most of the features studied — a relatively simple genetic architecture that they say may serve as a tool for studying traits and disease in other animals, including humans.
"When we study dogs we learn so much about mammalian biology as a whole," co-corresponding author Elaine Ostrander, a cancer researcher at NHGRI told GenomeWeb Daily News. "I think it's going to give us a much deeper lexicon of human variation as it pertains to growth regulation, as it pertains to proportionality, as it pertains to symmetry."
The study was done through the ongoing CanMap Project, led by Ostrander, Stanford University geneticist Carlos Bustamante, and UCLA evolutionary biologist Robert Wayne.
CanMap was established several years ago to begin exploring how genetic differences account for phenotypic differentiation between dog breeds, Bustamante told GWDN. "The idea was to genotype a common set of dogs that we would then use for a series of different studies."
Dogs are amenable to genetic studies because they fall into hundreds of isolated populations with a wide range of physical features and behavioral traits.
For the current study, researchers selected a subset of the CanMap dogs for which they had either individual measurements and/or breed standard information. They also incorporated skeletal measurements and data for several breeds examined.
Using the Affymetrix version 2 Canine GeneChip array, the researchers then genotyped 915 dogs from 80 domestic dog breeds recognized by the American Kennel Club, including several unrelated dogs from each breed, 83 wild dogs, including jackals and wolves, and 10 African village dogs from Egyptian shelters.
After applying a newly developed genotype calling algorithm called Multidimensional Analysis for Genotype Intensity Clustering (MAGIC) and doing quality control steps, the researchers were left with data at 60,968 SNPs, which they used to look for associations with 57 individual or breed distinctive traits.
By looking for parts of the dog genome that are highly differentiated between breeds, Bustamante explained, the team was able to track down signals of selection coinciding with traits that have been enhanced through breeding in specific breeds. Such differentiation would otherwise be fairly limited, he noted, since dogs share a relatively recent common ancestor.
Indeed, the researchers did detect a new signal of selection associated with floppy ears. But many of the other signals fell in parts of the genome previously linked to breed-specific features such as coat texture, body size, or snout shape, suggesting many of the same loci influence a range of dog traits.
And while the team found 51 areas of the dog genome associated with physical trait variation, the overall variation generally corresponded to body size — which was itself associated with just six or seven spots in the genome.
"This whole slew of morphological variation basically boiled down to one strong association with body size," Bustamante said. "And that association with body size you could basically explain with about six or seven different regions of the dog genome."
"In every case, there would be loci that we would see over and over and then we would always see a set of unique loci," Ostrander explained.
Given the relatively simple genetic architecture in the dog genome, the team is optimistic about using dogs to study human traits and disease.
"It helps us simplify many of the variables that, in humans, you just have to deal with," Bustamante said, noting that genes that are showing up in dog studies often coincide with those found in humans. For example, the researchers evidence that the HMGA2 gene, which contributes to mouse and human body size, is also associated with body size in dogs.
Researchers are particularly interested in using dogs to learn more about diseases that afflict both dogs and humans, including epilepsy, heart disease, cancer, or neuromuscular disease that are complex and/or difficult to study in humans.
"Dogs get many of the same diseases that humans get," Ostrander said. "Now the focus is on trying to study diseases that are tough to get a handle on in human families."
Ostrander noted that the dataset generated for the current dog study is being made publicly available as a resource for other researchers.
"It offers scientists an opportunity to study the genetics of any morphologic traits that they're interested in," she said. "It really gives us a tremendously unbiased survey of the dog world in a way that no other public dataset has done."
For their part, the CanMap team plans to continue studying the genetics underlying specific physical and behavioral traits. The researchers will likely expand their genetic analyses to include individual dog genome sequencing in the future, Bustamante noted.