by Carol Beuchat PhD

An interesting study was just published about the genetics of behavior in the Belgian Malinois (Cao et al 2014). This is a working breed used in some of the same service environments as the German Shepherd Dog (e.g., military, security, etc), so behavior is important to the breed's function. Malinois that perform well, with good drive and initiative for work, tend to exhibit a circling behavior when in confined spaces, which is a form of obsessive-compulsive behavior. Dogs that do not display the circling behavior, and those that have very high levels of circling behavior, don't perform as well.

It turns out that a gene (Cadherin 2, CDH2; or genes in the same genomic block), that has been linked to obsessive-compulsive behavior in both Dobermans and humans might also be involved in the manifestation of these degrees of working and circling behavior in Malinois, from non-existant to extreme. Maintaining the most useful, moderate behavior in the Belgian Malinois is an example of something called "balancing selection", in which the heterozygous condition (e.g., Aa) is advantageous over either homozygous condition (AA or aa). (This is also referred to as "overdominance".) This means that breeding two dogs that are great working dogs and heterozygous won't produce better dogs, because some of the offspring will lack the drive and initiative to be good working dogs (AA), while others will have a double-dose of the CDH2 gene and be too high-strung to be useful. Because the best dogs will be heterozygous, selection tends to favor the gene combination that is the best combination of advantageous (good worker) and disadvantageous (moderate circling).

You might be familiar with other examples of overdominance in dogs. For example in the Whippet, dogs with one copy of a mutated allele of the myostatin gene (which is involved in muscle function) are significantly faster than dogs with the normal gene, but dogs with two copies of the gene are over-muscled (Mosher et al 2007). One again, the heteroygous condition is superior to either of the homozygous options.

One more interesting example is the ridge of the Rhodesian Ridgeback, which is caused by a dominant mutation (Hillbertz et al 2007). Dogs without the mutation don't have the ridge, and dogs with one copy of the mutation have the breed-typical dorsal ridge. However, dogs with two copies of the gene are predisposed to a congenital developmental disorder called dermoid sinus. Dogs without ridges are generally excluded from breeding because this is considered to be a fault, as are those with dermoid sinus. So again, the genotype resulting in the preferred phenotype is the heterozygous condition. But breeding two heterozygous dogs will result not in a litter with better ridges, but some offspring with ridges, some without, and probably some that are afflicted with dermoid sinus. (This is a simple Punnett square problem.)

These are three examples where assuming that breeding "best-to-best" will not result in "even better" because of failure to understand the underlying genetics. In fact, it can result in removing a dog from the gene pool for a genetic issue (e.g., a Malinois with extreme circling), when in fact breeding that dog to the appropriate mate (e.g., a homozygous dog with low drive) would result in heterozygous offspring that could have the perfect blend of motivation and self-control. Likewise, using Ridgebacks without ridges will produce some offspring without ridges, but it also will not produce pups with dermoid sinus.

With so many breeds facing a growing list of genetic issues as a result of the continued loss of genetic diversity, it is especially imprudent to remove dogs from the gene pool that could be used to produce offspring with the desired genotype (that is, heterozygous for the gene of interest) without the collateral damage of pups with unacceptable phenotypes.

Bibliography :

Cao X, DM Irwin, Y-H Liu, L-G Cheng, L Wang, G-D Want, & Y-P Zhang. 2014 Balancing selection on CDH2 may be related to the behavioral features of the Belgian Malinois. PLos ONE 9(10): e110075. (

pdf )
Hillbertz NHCS, M Isaksson, EK Karlsson, E Hellmen, et al 2007 Duplication of FGF3, FGF4, FGF19 and ORAOV1 causes hair ridge and predisposition to dermoid sinus in Ridgeback dogs. Nature Genetics 39(11): 1318-1320.
Mosher DS, P Quignon, CD Bustamante, NB Sutter, CS Mellersh, et al. 2007 A mutation in the myostatin gene increases muscle mass and enhances racing performance in heterozygote dogs. PLoS Genetics 3: 779-786. (
pdf )
How breeding the best to the best can be worse