Tuesday, November 27, 2012

Mixed Breed Dog DNA Tests

Are mixed breed dog DNA tests accurate?  

The short answer is yes.

But will they tell you what you want to know? Probably not. But then again maybe.

If your mixed breed pooch has a purebred parent or grandparent you'll be happy with the test.  If she's a mish-mash of this and that, you'll be scratching your head.  Here's why.  (Read the detailed version of my explanation, DNA Tests, published in The Bark (Nov/Dec 2012). 

It's complicated.

Dogs are not like us (or any other mammal apparently).  They're made up of many packages of traits rather than a whole bunch of individual traits.  The pointing/feathery fur/happy feet/floppy ear package goes into, lets say 30% of breeds, whereas the funny face/elegant tail/find-a-varmint-in-a-hole/sloppy-kiss bundle goes into, for instance, 40%. And both packages go into making up numerous breeds. (Yes, I'm making this up, but just to illustrate a point. Geneticists know about the packages but haven't as yet figured out what's in each one.)

Back to your dilemma. Once the lab tracks the mongrel DNA back past a purebred parent or grandparent,  it's likely only able to tell you what packages of traits are involved.  What it's saying is "Sparky is all mongrel, but he's got some trait packages similar to the Husky and the Border Collie."  Then you go, "That's impossible. Sparky looks like a Corgi/Beagle mix."

Here's the other weird thing about dogs. All modern dogs, no matter what they look like, are so closely related that if you go back a few generations, your pooch is closely related to every breed.   (More details about this in my article in the 2012 winter issue of The Bark.)

Here are a couple of examples.

Chance (pictured left enjoying his pond) a ten year old mixed breed dog who's lived with me for six years was my DNA guinea pig.  In 2009 I had his ancestry tested with the MMI Canine Heritage Breed Test. At that time  MMI looked at 96 markers and tracked them to 38 identifiable breeds.  Chance's breed composition analysis results: no purebred parent, one purebred grandparent that was a Siberian Husky, and in his ancestry - Borzoi, Doberman, and Border Collie.

The results make sense.
He takes after his Husky grandpa
This is what my boy looks like
As far as what's in the mix- Borzois were mixed with Huskys to increase speed.  Dobies are a relatively new breed.  Working Border Collies don't breed true, meaning as long as they do the job they can appear very different from one another.  According to the company, new breeds don't "breed true" in that their DNA doesn't always cluster as the breed that they are.  Rather it can indicate breeds that were used to make the new breed.   That's a vague answer, but it's enough of an answer to sway my opinion to say that this test is accurate. But I wanted to know more.

In April, 2012, I again tested his ancestry, this time using the MARS Wisdom Panel.  Mars looked at 321 markers and tracked them back to 185 breeds in their database. This is their report:  No purebred parent.  One grandparent that was a cross between a German Shepherd and a mongrel.  The other grandparent was a cross between a Japanese Chin and a Standard Schnauzer.   In the mix- German Pinscher, Pembroke Welsh Corgi, Italian Greyhound and Leonberger.  Hmmmm.  I'm losing confidence in the tests.  I'll do one more.  Two out of three?  (A fool and her money soon part).

In May, 2012, I tested his DNA with the Canine Heritage test again because now they had 120 breeds in the data base. (Update: Mars Wisdom Panel purchased Canine Heritage in July 2012 so Canine Heritage is no longer available)  I could have paid $25 to upgrade his 2009 test.  But to be fair in my test of the tests experiment, I decided to send the DNA in as if he had never been tested. The test results: No purebred parent, grandparent or great grandparent.  But his distant relatives include Alaskan Malamute, Pembroke Welsh Corgi, and a very small smattering of English Setter and Cocker Spaniel. OK. What's going on here?

What's going on is that the labs doing the DNA analysis have made huge advancements in the last five years and are able to dig a lot deeper into a dog's ancestral DNA.  Unfortunately the additional information is only adding to our confusion.

How can dogs looks so different than what they actually are?

Would you believe a DNA test that said this dog was a purebred Whippet?  She is.
Bully Whippet

And although she is a purebred, this is a good illustration of how a mixed breed dog might look completely different than what the DNA results indicate. and how it can happen in a very short period of time. In this case, it has to do with a mutation on one gene.  If a Whippet has one copy she will be very fast.  If she has two copies she will be over-muscled.  (I'm over simplifying a very complicated subject. If you want to know more, read the New York Times article.)

New York Times
So then, how many generations would it take for the little dog gene (the IGF1 mutation is the gene that, when switched on, makes little dogs little) to be trumped by a more dominant size variant to create a dog that is no longer little. Apparently not very many.

This 75 pound beast of a girl, Sumo, has a grandparent that was a Chihuahua mix.  The results were so counter intuitive to what she looks like that the lab (Canine Heritage) did the test twice.  Same results.

Packages of traits add to the mystery
As I mentioned above, to further complicate things, dogs get their traits in a collection of packages.  So if one trait in the package is trumped by another version of a gene during the transition from purebred to mutt, like the domino effect all the traits in the package may change markedly as well.  It's like the default setting in your computer.  (I get to mix metaphors on my blog, because after all, it's my blog)

That's probably what's going on here.  A reader sent me the DNA results of Daisy's Wisdom Panel test.  Intuition says she is a Lab mix, but DNA says something else.


Daisy's parents and grandparents were all mutts, with fairly strong DNA signals from Cocker Spaniel, Australian Cattle Dog, German Shepherd and Dachshund.  Other breeds in the mix, with weaker signals, include Toy Fox Terrier 18.6%, Irish Water Spaniel 17.2%, Rat Terrier 3.6%, Finnish Spitz 3.1%.

Not a Labrador Retriever anywhere in her family tree. How can this be?    My guess is that there are a whole bunch of linked traits that geneticists haven't identified yet.  When one trait is trumped by another, lots of stuff changes.  
Think of Daisy as the quintessential default setting.

On the other hand, sometimes a Lab is just a Lab.  My Lollie's results came back showing that she had a grandparent that was a purebred Lab.  Her mongrel relatives were part German Shepherd and part Rottie.  She looks like a Lab/Rottie mix, with the face of an angel. (Results on angel DNA pending.)

Here's another example, in my own back yard.
My Izzy was born in late 2010, weighs 25 pounds and looks like a Chihuahua/Whippet cross in a Golden Retriever costume.  I tested Izzy's DNA with the Canine Heritage Mixed Breed DNA test and the Wisdom Panel Mixed Breed DNA kit.

The Wisdom Panel results indicated that one parent was a Miniature Pinscher/Shih Tzu mix. On that side of the family her grandparents were too, as well as her great grandparents. This means that one parent was a "Pin -Tzu" designer hybrid which is an intentional cross between the Miniature Pinscher and Shih Tzu.  Her other parent was a mongrel of indeterminate ancestry. On that side of the family, all her ancestors were mixed breeds of unknown origin.  That's a Pin Tzu above, and Izzy to the right.  So based on Izzy's looks, I would say this test is accurate.

The Wisdom Panel test indicated she is the offspring of a carefully bred
hybrid dog who fell in love with a mongrelly mongrel.
The results of the Canine Heritage test indicated that both parents were mixes, which confirms what the first test described if you agree that a hybrid designer dog is a mix.  Like the first test, it indicated a purebred Shih Tzu grandparent, but it didn't mention any Miniature Pinscher ancestry.

Then again, sometimes it's totally confusing.  A reader sent this comment and photo.


 "I just got the results of my dog today, and am scratching my head.  I guess anything's possible, but I REALLY expected Catahoula in there somewhere! ...The DNA test came back as Kuvasz/Chow mix with some mixed breeds in there. I'm kind of baffled, because I'm not even sure if those dogs have the blue merle gene!I sent Wisdom Panel a question asking if Catahoula could be close to any of those breed markers or not. Wonder what they'll say." 

MARS responded promptly and comprehensively. To read what MARS said, click here

And sometimes a dog is just a dog.
Cholee's results indicate that on one side she has all mixed breed ancestry. On the other side she has grandparents and great grandparents who were Dachshunds, along with a great grandparent who was a Collie.

Her person wrote, "It's just fascinating how much we assume about a dog's appearance, and also how that impacts what we expect behaviorally from that animal.  We thought that Cholee was part Beagle, maybe Lab, maybe Rhodesian (mostly based on personality), but she is no none of those ... which makes sense.  She seems to be a good example of the default dog - 35 pounds and tan."

Lookit.  The bottom line is that dogs are very weird, and a lot is still unknown about how they are put together. But we dog lovers knew this all along, because there is nothing else in the world that comes close to a dog.  So how could it be easy to make one? 

Send me your dog's photo along with her DNA results and I'll post them on my blog.  Email me - jlbrac@earthlink.net

Saturday, November 24, 2012

Dog Ears - Shapes and Sizes

Statistical analysis of dog ear shapes
h2meyer photography
I've done a lot of obscure research about dogs during the last 20 years, most of it while writing my dissertation.  Quite honestly, I've been collecting and analyzing data that few people really want to know.  But if you're one of the few, you might find this interesting, or at least mildly amusing.
Ear shape of 374 breeds of dogs
Of the 374 breeds described in The Atlas of Dog Breeds of the World

232 have drop ears 

78 have erect ears 

 17 breeds have rose ears

18 have semi drop ears

Ear flaps are traditionally docked in 18 breeds
(If left as nature intended, the ear shape would vary, as it's a trait not selected for)

And 10 breeds have ears anyway they want to be.

Another obscure ear shape fact: When Europeans first explored the new world they noted in their journals that all of the Native American breeds had prict ears and vocalized by howling, not barking.

If you want to know more about dog ear flaps, or just cool stuff about dogs, here are some sources:

Allen, G. M. (1920).  Dogs of the American Aborigines.  Bulletin of the Museum of Comparative Zoology,  Harvard College, 43 (9).

Brackman, J., (2001).  Engineering Pedigreed Dogs:  A Semiotic Perspective.  Semiotica Journal, 133-1/4, 141-155.  (Doctor Barkman's dissertation)


Wilcox, B., and Walkowicz, C. (1989).  The Atlas of Dog Breeds of the World  (5th ed. rev.). Neptune City, NJ:  TFH Pub. Inc. 

Sunday, November 18, 2012

Lincoln's Dog

Abe Lincoln and his dog were devoted companions. Where Lincoln went in Springfield, the dog followed. Fido was never scolded. He was always allowed in the house when he scratched the door, and into the dining room at mealtimes where he was fed from the table.   Apparently he was so spoiled that he slept on his own horsehair sofa. To read more about Lincoln's dog, Fido, read the Abraham Lincoln Research Site.

Photos: Henry Horner Collection at the Abraham Lincoln Presidential Library and Museum

Thursday, November 15, 2012

Rabies Awareness

Don't be complacent about rabies.

Rabies is a relentless killing machine that exploits the very thing we love most about our dogs, their sociability with people. 

You might be thinking that rabies is controlled in the U.S., and therefore your dog doesn't need to be vaccinated.  But you'd be wrong.

It's true that the number of human deaths in the U.S. has declined to only 2 or 3 a year. But hundreds of companion animals are put down every year because of suspected exposure to rabid wildlife, only because they weren't vaccinated.

According to the Center for Disease Control, domestic animals accounted for almost 10% of the 6000 confirmed cases of rabies in animals last year.
Rabies cases in U.S. in 2008. About 10% were pets.

If your dog isn't vaccinated, it's a good day to do it.

For More Information

A very good book about a very bad disease:  Rabid by Bill Wasik and Monica Murphy (2012)

Read my review in The Bark magazine

Friday, November 9, 2012

Military Dogs

Veterans Day is held on the anniversary of the end of WWI to honor US veterans. This post is to remember all the dogs that served in the military - as mascots, scouts, trackers, bomb sniffers, rescue dogs and companions.

Civil War perhaps.  Arrow points to the dog.

Sallie was a companion of the 11th Pennsylvania Infantry.
She is memorialized on Oak Ridge at Gettysburg
National Park.

General Custer with one of his many dogs.
Library of Congress

The site is the Custer memorial.

Sergeant Stubby was the most decorated dog of WWI.

Rags was a WWI hero.
Library of Congress

Marine in Korea
U. S. Naval Institute

U.S. Naval Institute


Beside the links above, these are good sites to see photos and learn more about dogs in the military:
US Army Quartermaster Foundation War Dogs 1942 - Present
U.S. Naval Institute's Sea Dog Page. Fantastic photos.
I posted a photo of Civil War era Mastiffs.