Strap line

It started when two canine scientists decide to become pen pals in an era of digital media...

Sunday, 25 October 2015

Understanding the Tame Fox: The Hunt for the Genetic Mechanisms of Fearfulness

Jessica Hekman meets friendly fox.
Guest post by: Dr Jessica Perry Hekman DVM MS

Hi Mia and Julie,

One of the things I love most about dogs is how friendly they are. What's the biggest difference between a dog and a wolf? The dog probably wants to come say hi to you. The wolf is scared of you, and may demonstrate that fear through aggression if you get too close. 

But not ALL dogs are friendly, right? If “friendliness” versus “fear” was on a spectrum, most dogs would be on the “friendly” end, but some would be down towards the “fear” end with the wolves. This is what I study: what is going on in the brain of fearful dogs to make them scared of things? What are the mechanisms that are different? I'm interested in the wiring in the brain, hormone differences, neurotransmitter differences, and more.

Right now, the way I'm studying fear in dogs is by studying fear in foxes. I know you know about the famous population of tame foxes at the Institute of Cytology and Genetics in Novosibirsk, Russia. I got to go meet these foxes recently! 

They have been bred for tameness for more than forty generations, and are a lot like dogs in their friendliness and lack of fear. In the lab where I work, we compare these foxes to another group of foxes that have been selected for fearfulness/aggression. Through comparing these very different groups, we to try to understand the biological mechanisms behind their personality differences. Foxes are evolutionarily close to dogs, and because these foxes have carefully controlled genetics and environments, they are easier to study than pet dogs are. (I still hope to transition to working directly with pet dogs some day, though.)


Hekman with friendly fox.
Over the years, much has been learned about the tame foxes: their stress response is very different from that of the aggressive foxes [1]; they have different levels of various neurotransmitters in their brains [1], and they even have differences in brain structure [2]. In the lab where I work, Kukekova Lab at the University of Illinois at Urbana-Champaign, we study the genetic differences between these groups of foxes. We hope that finding differences in their genes will help us learn more about mechanistic differences in their brains. Our lab recently published a new paper in PLoSONE [3] about some of our findings.

We are still very much at the stage of just trying to figure out where in the enormous genome (3.3 billion nucleotides!) the personality differences between tame and aggressive foxes come from. (By the way, various efforts looking at personality differences in humans are at the same stage.)


Wikipedia
Our lab's approach in this paper was to look at the entire genomes of 40 foxes, 20 tame and 20 aggressive. First we found the places in their genomes in which at least one fox was different at the level of just one nucleotide. A single nucleotide differences is known as a single nucleotide polymorphism, or SNP. Unfortunately, while a SNP might be a pointer to an important difference, most SNPs mean absolutely nothing. The problem is telling which is which. And we found thousands of them – more than 100,000 of them, which we filtered down to 8,437 of them that we actually wanted to use. So how would we figure out which ones were pointing at real and important differences in the tame fox genome?

To answer this question, we looked for differences not just between individual foxes, but between the group of 20 tame foxes and the group of 20 aggressive foxes. With 8,437 SNPs you had better bet we used computers for this. It was a surprise to me when I got into modern genomics just how much of the work deals with complex computer algorithms to process the massive amounts of data we're dealing with!


Hekman with fox at the Institute of Cytology and Genetics.
We found several areas of the fox genome in which tame foxes tended to have one version of some SNPs, while aggressive foxes tended to have another version. To understand this, it can help to think of the genome as a big instruction manual, a book called “How to Build a Fox.” Mostly the instruction manuals for tame and aggressive foxes would be the same, but in a few cases single letters would be different. 

Imagine that the tame foxes all had, “The quick brown fox jumps over the lazy dog” in their version of chapter two, but the aggressive foxes all had, “The quick brown fox jumps over the lazy doc” in their version. Basically, by finding places where the tame foxes all had one version, and the aggressive foxes all had a different version, we were finding places in the genome where we hoped to find important differences, changes that help cause the tame fox personality phenotype. We found 28 regions like this, but focused on three of them as the most interesting.

Next we looked at the genes that these changes might affect, because finding gene differences was the point of the whole exercise. Remember, we still don't know most of what most of the genes out there do! This is really dark side of the moon stuff, and everyone is still guessing about what's going on in the genome, human or fox. But here are some interesting genes we found. I'm including a “wild hypothesis” with some of them. These hypotheses are probably wrong, but I hope they help to show why these genes are so interesting to us...


Hand sniffing.

GRIN2B: this gene codes for a receptor for one of the major neurotransmitters in the brain, glutamate. Glutamate is involved in learning and memory. Wild hypothesis: maybe tame foxes are less afraid because of a difference in how they learn about what to fear.

GABARAPL1 (GEC1): this gene is involved with opioids, molecules in the brain that help us feel good. Wild hypothesis: maybe tame foxes are more friendly because social interactions feel different (better) to them.

COUP-TFII (NR2F2): this gene is important during embryonic brain development, especially in the amygdala, a part of the brain that tells us when to be afraid. This gene also influences expression of oxytocin, a neurotransmitter which functions in social bonding.

These genes are extremely interesting, but even more than that, this work helped our lab implicate specific regions of the genome in the differences between tame and aggressive foxes. That list of regions will prove invaluable as we do more work in the future, using different tools to examine the tame fox genome and seeing which tools point at the same regions.

Tame fox kisses to you both!

Jessica


Dr. Jessica Perry Hekman graduated Tufts Cummings School of Veterinary Medicine in 2012, where in addition to her DVM she received an MS for work on stress behavior and cortisol levels in hospitalized dogs. She completed a shelter medicine internship at the University of Florida in 2013, and is now a PhD student at the University of Illinois at Urbana-Champaign, where she works at Kukekova Lab. Her research interests are in the biological mechanisms behind fearfulness in dogs. You can follow her at @dogzombieblog.

Images copyright Jessica Hekman.

[1] Trut, Lyudmila N., I. Z. Plyusnina, and I. N. Oskina. "An experiment on fox domestication and debatable issues of evolution of the dog." Russian Journal of Genetics 40.6 (2004): 644-655. 


[2] Huang, Shihhui, et al. "Selection for tameness, a key behavioral trait of domestication, increases adult hippocampal neurogenesis in foxes." Hippocampus (March 2015). 


[3] Johnson, Jennifer L. et al. "Genotyping-By-Sequencing (GBS) Detects Genetic Structure and Confirms Behavioral QTL in Tame and Aggressive Foxes (Vulpes vulpes)." PLoSONE (2015).


© 2015 Jessica Perry Hekman | Do You Believe in Dog?

Tuesday, 13 October 2015

What can a DNA test tell you about your mixed breed dog?

The results are IN!
Rudy's DNA test results have come back.


Let's see what a DNA test can tell us about this mixed breed dog.

What you thought he was

We asked you all to place your bets on what mix of breeds he might contains, and boy did you come to the party! Here's what the poll results say YOU think he is:


The 'Other' category included suggestions of: Collie, English Foxhound, Irish Setter, Galgo, Super cute curly tail hound (!), Glamour dog(!), Borzoi, Petit Basset Griffon Vendeen, Pomeranian and German Shepherd.


What the DNA test says he is

Not surprisingly, Rudy has been identified as having come from a line of mixed breed dogs. 

We know that he was picked up as a stray in a very rural/bush area as a four month old puppy. In that location, it's likely he was bred to be a hunting dog, and was bred from a line of dogs very similar to him. Both of Rudy's parent have been identified as being mixed breeds. This has meant the DNA test is not just a clear cut simple cross between two breeds, but a bit murkier to decipher. 

What we expected

We always knew there would be sighthound in the mix! Fortunately, one grandparent on each side (e.g. his mother's mother and his father's father) have been identified as being a single breed. So the test says that Rudy is identified as one-quarter Scottish Deerhound and one-quarter Greyhound. This makes sense and fits with his physical appearance.


(Photo: source)
(Photo: Sue Muir)

Part of Rudy's DNA test results report


What we weren't expecting

Because of Rudy's muddled up mixed breed lineage, the DNA test results offer us a further five breeds that have been identified as "the 5 next best breed matches which appeared in the analysis of your dog's DNA. One or more of these breeds could have contributed to the genetic makeup of the ancestors indicated by the mixed breed icon. The breeds are listed by the relative strength of each result in our analysis with the most likely at the top of the list". This is definitely where the fun starts!

*Drumroll*

With the highest 'relative strength' (undefined and unclear if this is supposed to be % of total dog, or % of the 50% unaccounted for, or some other strength) of 10.86 (again, 10.86 units of what, or out to a total available number of ##, is unclear) - is... 

The Dobermann! I find this plausible. Dobermanns are certainly not unusual in Australia and I can see inclusion in a line of hunting-purpose dogs making sense to someone at some point, way back when. Rudy also has a wonderful mate who's a Dobe, so now they're totes cousins.


Dobermann reportedly represented in Rudy with strength of 10.86
Now, our first surprise... coming in with strength of 10.55 - the Shetland Sheepdog! Ha! I would have been more convinced by a Rough Collie I think, but who knows, maybe Rudy's great great grandma was a house dog sheltie? It would explain those neck flares... This is nothing compared to the next couple!


Shetland sheepdog reportedly in Rudy with a strength of 10.55

Oh hi there Puli, with a strength of 6.64. In a million blue moons, I would never have picked you in this line up! There are not that many Puli breeders in Australia, and to think that one was used to contribute to a farmer's hunting line seems... odd. But then, my personal favourite is still to come....

Puli reportedly in Rudy with strength of 6.64

...but we'll save it for last. Coming in with the second lowest strength of the five mixed breed contributors identified, we have the Irish Wolfhound. I know most of you thought this was going to be a leading contributor to Rudy's make up, but whatever strength represents, 3.28 doesn't seem like much of it. So now my favourite....

Irish Wolfhound identified in Rudy with a strength of 3.28

Basset Hound. I guess Rudy gets his leg from the other side of the family, right?! With a strength of 1.63, it's the final and lowest reported strength identified in Rudy's report.

Basset Hound reportedly in Rudy with a strength of 1.63

Part of Rudy's DNA test results report


The science behind mixed breed DNA tests

So how did the results end up like this? DNA tests for mixed breed dogs vary between providers. We used the Australian Advance/Waltham test which is 'powered by Wisdom Panel', validated against Australian dog populations. This test examines the 321 markers from the DNA against a database of DNA markers for over 200 representative (NB: not comprehensive!) breeds of dogs and a computer program evaluates and returns a probable 'pedigree tree' reaching back three generations. Every possible combination the computer program arrives at is scored and the tree with the highest score is deemed most probable and presented in the report.

An important note about this kind of test is that 321 markers are not that many. Other canine research (genotyping for whole genome analysis) can use 170,000 markers. Human ethinicity testing relies on 20,000 (to determine caucasian/non-caucasian) -700,000 markers. 

321 markers provide a reduced scope of DNA marker testing, and they are comparing an unknown dog against a bank of typical groups of alleles that representatives for breeds, so the robustness of the test results should definitely be considered as a suggestion, more than an absolute truth. 

Another way to think of it

The best analogy (that may be over simplified, but I think is still useful) that I have been able to come up with to help explain this test in relation to mixed breed dogs is to think of dogs as colours. 

By this, I mean there is a broad spectrum and range, all able to be mixed together in various combinations, over time. We've applied some values to the range (such as when blue becomes green, or orange moves into red) which we can think of as breeds. 

Image: Dean Russo
So consider Rudy as being a light brown colour. The DNA test is essentially trying to determine the combination of colours that arrived at that shade of light brown. It's pretty sure there's some red and green in there, perhaps some yellow too. But because he's such a mixed up colour, it's harder to work out if there's also been orange  (or was it a certain shade of red with a different shade of yellow?), white, a darker brown or even some bright blue included, and when they might have been mixed into him. 

The computer program has presented his report with one possible combination of colours that arrived at his shade of light brown, but it's not the only possible combination to get there. And when I consider where he came from and the likely uses and sources of his forebears, I can be fairly sceptical about some of the results (I'm looking at YOU mop dog!).

For example, I would probably have believed fox hound over basset hound. Or rough collie over sheltie, and I'm not confident how well 321 markers can discriminate between low levels of these breeds in comparison to each other by using the database of 'typical representatives'. As time goes by and the databases are expanded, these tests become more reliable. For example, the test conducted this year is likely more correct than one done five years ago. If they extend the number of markers examined to 1,000 in the future, this would improve the accuracy again.

So - what is Rudy?



He is our dog. Much-loved family member, silly goose, and constant source of delight to our family. His breed heritage is not so important to us. We knew he was sighthound mix type of dog when we adopted him, and he still is. When I next get asked (as I always do!) "What IS HE?", I can now reply with a slightly more informed "He's a mix, mostly deerhound and greyhound, with little bits of a few other things in there too". He is certainly a dog.

Look forward to any comments and questions you might have,

Mia

Further reading:

van Rooy, D., Arnott, E. R., Early, J. B., McGreevy, P., & Wade, C. M. (2014). Holding back the genes: limitations of research into canine behavioural geneticsCanine Genetics and Epidemiology1(1), 7.

Hedrick, P. W., & Andersson, L. (2011). Are dogs genetically special?.Heredity106(5), 712.

McPhee, C. G. (2011). Advances in canine genetic testing—And what these tests mean for youVeterinary Medicine106(12), 608.

© 2015 Mia Cobb | Do You Believe in Dog?