There have been major advances in the canine genome map over recent years, with an agreed karyotype for all 38 pairs of autosomes and an integrated linkage-radiation-hybrid map now available. An individual dog breed represents a closed, isolated population, and many suffer genetic diseases. Some are homologs of human conditions, and dogs represent naturally occurring, potentially useful large mammal models. Other conditions may be unique to this species, but study of these offers new insights into metabolic pathways. Linkage analysis may identify a locus linked to a particular condition. Dog pedigrees facilitate these studies by being relatively more informative than human families, although planned breeding strategies to maximize the informativeness may be required. A study of a late-onset, acquired heart disease, dilated cardiomyopathy (DCM), in Newfoundland dogs highlights some of the problems. The disease occurs naturally in pedigrees within the UK in dogs kept as pets and breeding stock. Problems encountered include confirmation of the mode of inheritance and determination of phenotype. A genome-wide linkage analysis study with over 200 markers failed to detect linkage. The study indicates that more detailed linkage maps are required, and further synteny information between dog and human gained before study of diseases affecting naturally occurring families of dogs from inbred pedigrees may have maximal chance of success in order to utilize this model.