Mounting genomic evidence suggests that biological contributions to psychiatric disorder susceptibility are genetically complex, environmentally mediated, and highly comorbid. By paralleling human initiatives, I propose that comparative genomics in the domestic dog offers informational and translational utility to investigations of such complex conditions. Dogs present problematic behaviors with ostensible similarity to human disorders, such as separation anxiety and compulsive behaviors. As behavior played a major role in dog evolution, canine genomes may be enriched for common genetic variants underlying many behaviors. Breeds represent a distillation of diversity in appearance and behavior, but there has been limited success in linking the latter to genes using only pedigreed dogs. The genomes of “mutts” present a natural experiment in genome-wide admixture, which I leverage to map high-dimensional phenotypes across thousands of pet dogs. While I find that most behavioral differences are heritable, especially functional behaviors characteristic of major lineages, I also show that breed alone is an unreliable and confounded predictor of dog behavior. By intersecting genomic studies in people and dogs, this work supplies a roadmap for discerning and ranking cross-species relevance for a wide array of canine and human phenotypes, including psychiatric conditions.