Like progressive retinal atrophy (PRA), the term retinitis pigmentosa (RP) refers to a group of diseases with similar pathology but distinct genetic causes. Because of the clinical similarities between the canine and human diseases, Cornell University scientists have previously cloned and sequenced the canine equivalents of several human genes with known roles in the inheritance of RP. The so-called candidate gene approach was used successfully in studies of rod-cone dysplasia 1, a form of PRA that occurs only in Irish setters. A mutation-based blood test developed by the Cornell group has been in use since 1994 for identifying carriers of the Irish setter disease.

However progressive rod-cone degeneration (prcd) proved to be a tougher nut to crack. The candidate gene approach led only to the elimination of a series of likely genetic suspects. So the Cornell researchers began a collaboration with Elaine Ostrander and her post doctoral associate, Cathryn Mellersh, and others at the Fred Hutchinson Cancer Research Center. As a continuation of a project that Ostrander had begun in the laboratory of Jasper Rine at the University of California, Berkeley, Ostrander's group had developed a panel of 150 highly variable DNA sequence repeats, called microsatellite markers, to aid in developing a genetic map of the dog on which to localize gene sequences controlling the inheritance of diseases or other traits of interest. To determine the relative location of these markers, however, they needed to check them against the DNA of related dogs from fully reliable pedigrees and compare the results. CornellÕs Gregory Acland and Gustavo Aguirre, working at the James A. Baker Institute for Animal Health in the College of Veterinary Medicine, had more than a dozen highly variable and rigorously documented pedigrees that they had developed for their research on the inherited causes of blindness.

Mellersh, Amy Langston, and the rest of the Ostrander group spent several painstaking months analyzing all the markers they had generated, using powerful statistical methods to determine how the markers were ordered. The result of their work, the first canine genetic map, provided investigators throughout the world with access to the workings and organization of the canine genome.

"When our two groups first got together, it was immediately obvious that we were holding the two halves of one puzzle," recalls Acland. "Elaine Ostrander's group had the markers we needed to help us find the prcd locus, and we had the genetic material they needed to make sense of what they had."

In addition to the 150 microsatellite repeats, which are non-coding sequences collected from random locations throughout the canine genome, the Ostrander group also had information about a number of coding sequences, whole or partial genes from the human genome map. One of these was BRCA1, a breast cancer gene that had been mapped to human chromosome 17. These potential markers, were typed on the pedigrees that had been designed to yield information about prcd, some of them did just that.

One anonymous marker, now designated C09.173, had already been shown to be located near the canine BRCA1, a gene cloned and sequenced by the Ostrander group, although these markers had not been assigned to a particular dog chromosome. When this and other linked markers were tested against the prcd pedigrees, they also showed a pattern of co-inheritance with prcd. This information, plus the subsequent discovery of several other genes and anonymous markers flanking the prcd locus, told the researchers that the gene controlling the expression of prcd, though still not identified, is part of a cluster of genes and anonymous markers that tend to be inherited together.

By determining the relative odds of each of these traits following along with the prcd trait in successive generations of dogs, the researchers were able to determine their relative distance from the prcd locus. This group of markers, with the prcd locus in the middle, constitutes what is called a linkage group.

The prcd linkage group discovered in the dog is remarkably similar in composition and ordering not only to the corresponding region of human chromosome 17, but to the homologous region of the mouse genome as well. In the mapping of the human genome, it has not been difficult to assign markers to particular chromosomes, because those chromosomes are large and easily identified by their shapes.

Canine chromosomes, on the other hand, are small and unremarkable in shape, and there are a lot of them - 38 pairs of autosomes plus the X and Y, compared to the 22 autosomal pairs plus X and Y that humans carry. Due to the sheer difficulty of distinguishing one canine chromosome from the next, no heritable trait had yet been mapped to an identified canine chromosome other than the X.