Buyer beware – or at least do the homework.

Over two years ago, I wrote about the downsides of large panel disease gene tests. I was thinking about how these tests might possibly impact breeding programs and gene pools.
This month there was an important opinion piece in Nature by Lisa Moses, a research scholar at Harvard Med School and a practicing veterinarian, Steve Niemi, a veterinarian and director of the Office of Animal Resources at Harvard, and Elinor Karlsson, director of vertebrate genomics at the Broad Institute of Harvard and MIT, that highlights similar concerns. They cite weak science, lack of validation, imprecise results or interpretation, and conflicts of interest as reasons to be very cautious with the “direct to consumer” disease identification tests for dog owners.

They use an example of a pug named Petunia that apparently was tested as a carrier of a marker for DM, degenerative myelopathy, a very imprecise marker that is meaningless in some breeds and meaningful in others.  The owners opted to euthanize Petunia rather than have her live what they assumed would be a horrible life. While pugs are known for spinal issues, they are not typically due to DM, even though the marker is found in the breed. But Petunia was put to sleep anyway.

In the last year, I have been increasingly alarmed seeing ads depicting, for example, a sad mixed breed dog with a caption about being glad “mom” found out he carried a copy of the mutation for PRA (progressive retinal atrophy.) This was a highly misleading advertisement from a popular company – since carrying only one copy of the mutation for PRA means the dog will never go blind. That pup should be celebrating, but that’s not what sells. Moreover, most mutations tested by these large panels are both breed specific AND recessive, meaning most mixed breed dogs are highly UNLIKELY to carry two copies and therefore get sick.

Not unlike the disease testing, the genetic diversity information garnered by these tests is limited. The same people ensuring the validity and importance of the disease testing are also telling breeders their genetic diversity methods are superior. This is to be expected – they are selling the tests and it’s their job. However, they use the fact that they use more markers and that these are found on all chromosomes. More markers also sound good to pet owners and breeders – more sounds better than less, right? – but in fact 230,000 markers are out of 2.8 billion base pairs, a tiny fraction of .008% of the canine genome. There are on average 12,000 untested markers in between each of the markers that are tested, so this is actually a low density panel – it only seems quite thorough.

Likewise, breed identification is fun – but it’s also theoretical. There are no multi-generational studies that show that offspring resulting from multiple breed crosses that have been tested and shown to be accurately detected by these breed identifying tests. There are some tests of livestock showing accurate identification of mixes of two breeds, but there are none for dogs. There’s no one policing these tests or ensuring their accuracy.  That’s ok – because it’s only a fun thing, not a serious veterinary test. But pet owners should know their results aren’t guarantees of anything, in the same way that none of these tests can guarantee purebred status of dogs.
Why not? Because dogs are a single species, and breeds are not scientifically defined populations. Breeds are more or less related, and they often blur at the edges –  since most have common ancestors. These large panel tests detect both recent and ancient ancestry, and are dependent on their reference samples – the dogs they test as examples of a breed. There are over 220 AKC breed clubs alone. The UC Davis Veterinary Genetics Lab has tested over 500 identified canine populations!

Often the VGL Canine Diversity test we use for our analyses, which was created by veterinarians, geneticists and biostatisticians to help breeders manage their gene pools, is said (by advocates of the  large panel tests) to be “old technology.” In other words, it has been around long enough to be reliable and thoroughly tested. Your AKC DNA profile? Same method. Your FCI parentage testing? Same method. It is also the same method used to establish human paternity and convict felons through forensics. It can identify any human on earth. And it is recommended by leading conservation geneticists.

The data garnered from it is excellent for estimating genetic relatedness. There are a great many research papers published in the last decade that explore all the strengths and weaknesses of the various estimators of genetic relatedness and how best to make use of them (such as this from 2011 and this from 2015 and this from 2015 and this from 2015 and this from 2016, among many others.) With all this recent information, it has been possible to have accurate estimates of genetic relatedness, and our members see it every time they look at our software’s estimated relationships. This DNA test is easy to analyze, is affordable, quick to process, accurate even with slight changes in DNA, and can be run with less than pristine DNA samples. It’s an excellent test for its important purpose – accurately testing genetic diversity for dog breeders, planning breedings and managing their gene pool.

Lastly, all tests have strengths and weakness. There’s certainly some usefulness to large panel tests – as long as you know what to do with the information, and as long as you are willing to do your homework. In the end, we are still breeding animals and we are not able to control all possible factors. We still have to select proper, functional structure and temperament, robust good health and investigate pedigree health histories. All tests are additional to those – and can be enormously helpful, but can never take the place of taking this endeavor very seriously and making conscientious, comprehensive choices.
Natalie Green Tessier