Drs. Elaine Ostrander and Heidi Parker — Unleashing the Dog Genome
Centuries of selective breeding have given rise to a staggering variety of dog breeds, each with its own traits and behaviors. But shallow gene pools have also put some breeds at higher risk for disease. Dr. Elaine Ostrander runs the Dog Genome Project at the National Human Genome Research Institute. Her team includes Dr. Heidi Parker. Together, they are digging for clues to understand how genes code for dogs' diversity and disease. Clues that might also inform the health of their two-legged caregivers.
Learn more about Dr. Ostrander, Dr. Parker and the Dog Genome Project at https://irp.nih.gov/pi/elaine-ostrander.
>> Diego (narration): A friend of mine recently adopted a dog.
[Friend says, “come here, Pop.” Dog collar jingles]
The dog’s name is Poppy. She’s a golden doodle with floppy ears and the most endearing little twinkle in her eyes that just barely peaks through her curly fringe. Walking Poppy around town, it’s curious to see how different she looks from other pooches.
[Faint sound of street traffic and footsteps on the sidewalk]
On one block, a burly Pitbull, twice the size of Poppy struts its velvety grey coat and has a playful grin from ear to ear.
Around the corner, a dapper huskie in its fluffy black and white suit perks up its pointed ears after spotting us with its piercing blue eyes.
Suffice it to say, the wide variety of dogs is pretty staggering.
>> Dr. Ostrander: There are over 300 breeds across the world, and they display an astounding range of variation.
>> Diego (narration): That’s Dr. Elaine Ostrander. She runs the Dog Genome Project at the National Human Genome Research Institute. Her team fetches DNA samples, health histories and pedigrees from pet dogs to figure out the genetics that account for so much variation and give different breeds their characteristic traits. Things like body structure, head shape, leg length and even specific behaviors.
>> Dr. Parker: Also, we look at disease.
>> Diego (narration): Dr. Heidi Parker is a staff scientist in Dr. Ostrander’s lab. She studies how diseases like bladder cancer are (essentially) encrypted in dogs’ genetic code and why some diseases occur more often in certain breeds than to others.
>> Dr. Parker: We’ve identified regions of the genome that are involved in different cancers, and how they develop. And we're trying to work out now exactly how that works
>> Diego (narration): Because particular diseases and distinct physical features are concentrated in one or a small number of related breeds, patterns of inheritance are easier to figure out in dogs. So their genome is the perfect place to dig for clues about what controls if and how something will be passed down. Clues that might not only explain the genetics of dogs but help another species that exhibits similar variation and suffers from many of the same illnesses.
>> Dr. Ostrander: dogs and humans get all the same diseases. By studying canine health, we learn things about human health as well.
>> Diego (narration): For this episode, I sat down with Dr. Ostrander and Dr. Parker to learn how man’s best friend doesn’t just make for loyal company, but helps us understand more about or history, our development, and our health.
>> Diego (interview): My first question is I’m sure one you get often, but I can’t help but ask. Dogs come in so many different shapes and sizes. I mean, a Greyhound is at least three times the size of like a Shih Tzu, and has a much shorter fur, and yet they’re considered the same species. From those different traits, you'd think that they'd have pretty different genetics. It almost seems like it’d be comparing apples to oranges. So how do you get insight or derive information about the genetics of the species as a whole, much less a completely different one, like humans.
>> Dr. Ostrander: Yeah, so it is important to remember that all dog breeds are members of the same species Canis lupus familiaris—all breeds from the giant Great Dane and Irish Wolfhound and Scottish Deerhound down to the Pekinese and the Chihuahua and the Toy Poodle, all members of the same species. And modern breeds have for the most part only been around since Victorian times. So, all those breeds you see running around in the dog park, you know, the Labs and the Spaniels and the Poodles, those are pretty recent, I mean, you know, a couple 100 years. And dogs, as a whole, were domesticated from wolves, maybe 15,000-20,000 years ago. So evolutionarily, that's a teeny, teeny, teeny, teeny part of a drop in the bucket. So, what that means is that if you want to look at the genetics that controls body size between a Great Dane and a Chihuahua, what evolution has had to do is hit a small number of times [SMACKING SOUND] and hit hard, right? Evolution hasn't had the luxury that it's had in humans to make lots and lots of little changes and move along incrementally in response to human survival and success in an environment. We've selected for traits we want, a lot of which are aesthetic. We've done it very deliberately. And we've done it to create animals that have a particular appearance and a particular behavior. So, the number of genomic hits that are required to differentiate that Great Dane and that Toy Poodle aren't going to be millions. They're not going to be tens of thousands, right? They're not even going to be thousands and probably not even going to be hundreds. They're going to be a small number of really hard, strong hits. And that's why you can study different breeds, come out with a vocabulary or an outline of genes responsible for a phenotype or a trait and then look at that in humans.
>> Diego (interview): So, because evolution had to act fast, selected traits had to be controlled by a small number of gene that could account for a lot of variation, and it’s easier to study…
>> Dr. Ostrander: Yeah, exactly. Yeah.
>> Dr. Parker: And what's really interesting when we're talking about things like diseases in dogs, is that very often, because of these little pockets of selection that we have all over the world, those diseases may only be in a few of those, and not necessarily spread across the entire population. And so, we can kind of narrow it down into these really-- you know, what we consider really large families of dogs and say, why are they getting this disease? And it may be a really small number of genes, because it's just a really big family that is having, you know, whatever the disorder is that you're looking at.
>> Diego (interview): Right. There's a lot of a lot of sameness within those communities so it's easier to pick out things that make individuals different—disease in some cases.
>> Dr. Ostrander: Yeah.
>> Diego (interview): Well, you’ve been able to analyze some of those variations in characteristics, like body size, body structure, aging even by collecting samples from dog owners. What has the science told you there? Are there like specific genes that are particularly interesting?
>> Dr. Ostrander: I've been, in the last few years, really interested in morphology and in archaeology and ancient DNA. So, I want to know why, you know, the Great Dane or the Irish Wolfhound or the Scottish Deerhound has really long legs and the Basset Hound has really short legs. And those are questions we have answered in large part through Heidi's work. I want to know why some breeds have pushed-in faces, and some have elongated faces. I'm interested in understanding anything that relates to other species on the planet that can give us insight into other species, including of course humans. And as our genome sequencing abilities have become more sophisticated, we can really learn a lot about ourselves.
>> Dr. Parker: Yeah, so there's always something interesting coming up. So, like Elaine talked about: morphology. With dogs, we've mapped so far like 14 different genes that account for a huge amount of that variety in things you're talking about like a Greyhound versus the Shih Tzu, you know, very huge differences like that. So, that's really exciting to look at. Also, we look at disease. I study bladder cancer in dogs. And we found out a couple of years ago that the bladder tumors in dogs have this mutation that shows up in, you know, 8% or 9% of all human tumors, regardless of the type. So, here's this dog tumor that's really developing exactly the same way as some human tumors are. So, that's always exciting, when you start finding, you know, exact sort of duplicates and how they can behave the same way in two completely different species.
>> Dr. Ostrander: And I think we'd be remiss if we didn't talk about behavior, because behavior was the impetus when I started the Dog Genome Project. Oh, I've got to find out why herding dogs herd and pointers point. So, we do have people in the lab who are looking at the relationship of breed behaviors, and human behaviors. We've been particularly interested in ADHD in subsets of breeds as part of their natural behavior that, you know, breeders have selected on or what they display as part of the breed standard and how that compares to human ADHD, and where's the similarity in the underlying genetics.
And, you know, you can take a step back and you can just watch dogs do their thing. And you can develop hypotheses. You can say, huh, is that a herding dog, that Border Collie, OCD? Or does it have some particular learning ability that's really unique? Does it have some cognitive skills that are really unique? Does it have vision that far surpasses other dog breeds? You know, a herding dog will start herding if you put sheep in front of it or whatever, they're just going to do it. I watched a Border Collie on my street try and herd a great big, huge, pure white Labradoodle. It just dropped into the herding position, gave it the eye, and then crouched along, you know, it just thought it was a sheep, I guess. So, we want to understand those sorts of things as well.
>> Diego (interview): Interesting. I imagine that it would be a lot harder to map the genetics of behaviors, seeing that they’re more difficult to measure or defined. Especially compared to physical traits like weight or height.
>> Dr. Ostrander: Yes. Of course, that is more complex. That is not two genes, right? And that's where the community, I think, has always struggled with coming up with behavior assays. And even for what we recognize are anomalous behaviors, like, there is a lot of OCD in the Bull Terrier. Like they would chase their tail hour after hour after hour until they literally bore holes on the carpet. It was like almost the diagnostic was to walk into their living room and go, yup, the dog has OCD, look at the carpet. And for a while, we really did try and work on that. We had wonderful collaborators from Tufts. We collected samples. There were people here at NIH that were helping us. But what it really came down to were degrees of OCD. And even though dogs who had it manifested certain traits, it was still very hard to quantitate. And so, we took a step back from those anomalous behaviors, and went back to where we started, which was these breed standard behaviors; what makes a herder herd, what makes a pointer point. And I think that's where we're making more progress now.
>> Dr. Parker: The other interesting thing about mapping behaviors and why it's a little more challenging; it's not only the phenotyping, but behaviors are most likely the early designation of sort of breed or breed types. This would have dated back long before people started saying, “oh, well now, I need to breed this short-haired with black and white spots and floppy ears,” you know. First, they started saying, ‘I need a breed who's going to guard my livestock, I need somebody who's going to guard my home, I need something that's going to go hunting with me.” And then the other stuff, the sort of the morphology came after that.
>> Diego (interview): Huh, you mentioned that consistencies in behaviors could have been the early designation of breeds, but what qualifies a breed today?
>> Dr. Ostrander: Well, so there's a genetic definition and there's a political definition and, you know, all kinds of things, right? So--
>> Dr. Parker: So, a lot of times though like with new breeds, it's really based on people. People form a group and they start following a particular set of guidelines and stick within those guidelines to create something that breeds true, where generation over generation, you get the same kind of dog. You don't get, you know, the weird pop-up or whatever, but that everybody stays the same and there's a conscious effort to do that--
>> Dr. Ostrander: Let me just add to that by saying: they have to breed true through multiple generations for the things that are part of the breed standard. If you look up the American Kennel Club standards, they are exquisitely precise, you know, down to how many inches the legs should be, if there should be one curl over the back, or a corkscrew, or a straight tail, or two curls over the back, fur length, I mean, all of this stuff. So whatever you pick out is your breed standard that you care about, that's what you have to be true for. Other stuff, you don't. And so, if you don't care what color the coat is, well, then it doesn't really matter what color the coat is, right? But if you're going to say, “nope, all these dogs have to have a black coat or all these dogs have to have a silver coat,” then you're going to have to make sure that happens. So, you have to agree on a breed standard. That's important.
>> Diego (interview): That sounds pretty arbitrary to me. I mean, someone somewhere just decides that’s just the way it is?
>> Dr. Ostrander: Yeah.
>> Dr. Parker: Yeah.
>> Dr. Ostrander: Pretty much.
>> Dr. Parker: Yeah, sometimes it's just someone, sometimes it's maybe a committee. And that's why sometimes they change. And that's why new breeds are developed, too. It's like, well, we've decided that that's not the way we want our breed to look, we want it to be different. And so, a sort of brand-new breed pops up.
>> Dr. Ostrander: But it is true that for some breeds there is a behavior standard as well. And for the working dogs, boy, they've really got to do it or, you know, they're not going to be bred because people's livelihood and their success depends on this.
>> Diego (interview): Yeah and that was especially true when dogs were first being domesticated. It’s clear that the evolution of dogs has been heavily dictated by humans. Would it be too far to say that without humans, modern dogs wouldn’t exist?
>> Dr. Ostrander: So certainly, many modern breeds wouldn't exist. If we go back and we look in the Arctic 10,000 years ago, where dogs had evolved, I guess you would say, without much intervention or introgression from other breeds, then as humans in the area became interested in trade of metals, of glass, someone’s got to be able to haul the stuff. So now there’s selection and suddenly we see introgression of animals from Eurasia, telling us that humans are bringing other breeds over, other populations over. And these dogs have a purpose. They have to be transport animals. So, with early humans, yes, we would still have a lot of dog breeds, because our survival depended on them. As we became more pastoral and we needed herding dogs and we needed guard dogs, absolutely. In the absence of any human, is there a reason to have anything but wolves? I don't know. Heidi, what do you think?
>> Dr. Parker: Well, you have populations like the dingoes in Australia, right, that are a, you know, wild dog population. And it's possible that you would have had something like that in other parts of the world. They would be competing with the other canids in the region. So, it's hard to say without any humans at all, you know, because dogs seem to be obviously a population that in general is attracted to humans, human settlements, things like that. It's hard to say, right? Would there be this wild dog population that would be kind of like a, you know, a jackal or a coyote or something like? It certainly wouldn't look like it does today. You would have something much more on that wild appearance. Probably no floppy ears and short legs.
>> Dr. Ostrander: And their behavior would be so different. So, we were able to collaborate with an investigator from Princeton, and one from Oregon on just the coolest study. And what they were doing was they wanted to see what was differentially selected in modern dogs versus wolves. So, what are variants you see in all modern dogs, regardless of breed, regardless of where they develop, regardless of when, that you never see in the grey wolf? And one of the things that they came up with was rearrangements in a region that in humans we call the Williams Baron region. And we call it that because the syndrome that's associated with it has as one of its features, hyper-sociability. So, kids who have this, yes, they have developmental delays, yes, they have other issues, but they are very, very hyper-social. That's one of the hallmarks of it. And then you look back at dogs and wolves, well gosh, if you're going to domesticate dogs, and they're going to live with us, and they're going to be our companions, and they're going to live in our homes, and they're going to be with our kids, we need to communicate with them. We need to tell them, you know, stand here to guard these sheep against wolves. Or we need you to herd so we need you to go left and then right, and then over this hill, and so on and so forth. Or if you're going to hunt with us, you got to be quiet, and you need to run this way when the duck is falling. And, you know, all those communication skills, dogs had to be able to understand this. And so, the fact that that region is under selection as part of domestication is just incredibly exciting. And now, with genomic technologies, we can find those genes and those genomic rearrangements. And that's really the first example of that.
>> Diego (interview): That's pretty cool how the genealogy of dogs opens a window into human history. But getting back to the window it opens into our biology, many of the medical conditions that affect humans also affect dogs, you know, everything from epilepsy, to diabetes, to heart disease, and to cancer. Bernese Mountain Dogs, for example, have a high incidence of sarcoma, which is a cancer that affects the bones or soft tissue. Can you tell me about what your research is looking at, regarding those genetic risk factors?
>> Dr. Parker: Yeah, we look at a number of different breeds that have a high incidence of cancer, so the Bernese Mountain dogs with histiocytic sarcoma, also the Flat-coated Retrievers have a high incidence of the same disease. By comparing dogs that get cancer later in life with those that don't, we try to identify these differences across the genome. So, we have certainly identified regions of the genome that are involved in these different cancers, and how they develop. And we're trying to work out now exactly how that works because usually they don't knock out a gene, they don't remove an entire protein from the animal. They're just affecting how different genes are expressed and what time they're expressed to develop these tumors. So, Bernese Mountain Dogs, I think about 25% of the dogs are expected to get histiocytic sarcoma within their lifespan. That's a lot of individuals. And if you're thinking about genes, where it takes a few regions of the genome that are mutated in order to build to this cancer, that means probably everybody in the population has one or two, you know, maybe not all of the components, but some of them in order to get those mixtures at that high level. So then you start worrying, are we looking at things that everybody has? And how can we find out if it's in every dog in the breed, regardless of if they have the cancer right now? So then we start looking, maybe can we compare them to a very closely related dog? Their cousins, the Greater Swiss Mountain dogs—very closely related, look a lot like Bernese Mountain dogs, but have shorter hair, they're a little bit bigger—they almost never get histiocytic sarcoma. So, then we start looking; can we do some comparisons? What are the differences there? And that's true for almost every breed of dog, you can identify a close cousin, and then look at, do they have the same disease, because that helps us? Or do they not have the same disease because we can also use them as, you know, sort of case control comparisons?
>> Dr. Ostrander: So our lab has spent a lot of time on Bernese Mountain dogs and histiocytic sarcoma. We've just published on flat-coats and histiocytic sarcoma. In the past, we've-- and I guess we're actually bringing this project back, squamous cell carcinoma of the digits. So this is an interesting cancer. We started getting calls from poodle owners, Standard Poodle owners, saying, “my dog has, you know, cancer of one toe or two toes or three toes.” And the only treatment is to take the toe off, which is horrible for a dog. I mean, that's just horrible. There's the breeder point of view, you can't breed those dogs because, you know, something's wrong with them. But just as a dog owner and your dog is a member of your family, what a horrible thing to have to put your dog through.
>> Diego (interview): Yeah.
>> Dr. Ostrander: So they said, “can you figure out what's going on?” And I said, “OK, sure. Send samples.” And so from all over, we were inundated with samples of Standard Poodles with one, two, three, four, you know, toes that have this. And we looked at obvious things. Was this a melanoma? And we worked with UCSF to check that out. And no, no, no, it's not. And then a postdoc in the lab came to me and she said, “you know, we're not getting all colors of Standard Poodles. We're basically getting the darker color Standard Poodles. We're not seeing this disease in the lighter colors, the whites and the creams.”
And so that was an interesting epidemiologic observation. And it turned out to be really important in figuring out what was going on because the gene that was mutated is one of these kind of cross-over genes that we know is important in cancer, but we also know is important in coat color. So, it turns out, in this case, it is a pretty complicated story. But we're getting more samples of other breeds from collaborators now in Italy, so we can tease out more of the functional and biological way that this acts.
>> Diego (interview): Well carrying out these studies to improve the health of our furry companions is commendable in and of itself, but as I understand it, they could also provide a benefit to us directly. So how are these studies being translated to humans?
>> Dr. Parker: Yeah, and I don't know the exact numbers. And, Elaine, you might know this, but oftentimes clinical trials in humans start out with just laboratory investigation, and then they do something like in mice to see how it works, and if it's tolerated, and if it affects tumor growth. Then they go into human. And between mouse and human, there's a really high failure rate of what actually works in humans versus mice. Whereas dogs are already treated with basically human drugs. Humans are kind of the test population for dogs. If it works in humans, we'll try it out in the dog. So, if you could take it the other direction and go through dog first and then into human, we could really, we think, increase the number of trials that would actually pass in humans versus that huge jump from mouse to human.
>> Dr. Ostrander: And the other thing is, dogs are a different genetic background. So, if you take a human drug that maybe works somewhat in humans, and you put it in dogs, and maybe it works a whole lot better. Well, then you can look at what's in the genetic background of dogs that might direct you to design a new drug that would then make your primary drug more effective in humans. Or let's say, the drug works sometimes in humans and sometimes not, it's a success or it's a failure. You give it to dogs, and it's universally a failure. And so then you can say, ok, what's different? What's different in the dog genome that could help us understand the failure? And maybe we can understand then why some humans are failing therapy. And everyone at NIH will know things are of course never black and white when you're treating people with disease. But if you can begin to tease out the nuances of response of, you know, developing resistance to the drug, of the associated quality of life changes associated with the drug and you can do that in dogs, it can really suggest avenues of research that are going to be directly applicable to humans. And that, to me, is the most valuable information.
>> Diego (interview): So, it seems like these comparative genomic studies might lead us to, you know, at least more reliable genetic testing and improved diagnostics and treatment in humans ultimately.
>> Dr. Parker: Well, and remember, when you're talking about dogs, you're talking about individuals who are actually visiting the doctor on a regular basis, right? They're getting diagnosis, they're getting treatment, they're getting regular checkups, things like that. So you're actually looking at how does this work in a real life setting, not in a laboratory setting where everything is controlled. But is this improving, you know, their lifespan? Is this improving their quality of life, things like that, that you really don't see in in a laboratory.
>> Dr. Ostrander: You know, Heidi is right. Humans spend phenomenal amounts of money on their pets every year. Do you remember those numbers?
>> Dr. Parker: The exact amount, no, but they're huge. But this reminds me that I have a dog right now that I think has a torn ACL and she has an appointment to visit the sports medicine doctor in a couple of months, you know, the specialist for orthopedic surgeries because she's probably going to have to have that repaired. You know, don't do that on a mouse.
>> Diego (interview): That's very true.
>> Dr. Ostrander: And just like humans, dogs will live with something for most of their life—hip dysplasias, elbow dysplasias, cataracts, patellar luxation, right? And so how you're going to treat them over the span of 20 years becomes really informative for thinking about comparable human diseases.
>> Diego (interview): Yeah, definitely. I mean, dogs aren’t just test subjects, they’re part of people’s families and they care for them as such.
>> Dr. Ostrander: They're absolutely part of your family. And you know, you raise an important point really, because part of the reason surveillance is so high and so much money is spent and so much education and care goes into dogs is because not only are they members of our family; it's important to remember that for some people, it's actually the only member of their family that lives with them. You know, people later in life, kids long gone, spouse gone, or partner gone, this is it. Their life really centers around their companion animals. And Heidi, she can speak to this because she travels so much to specialties on dog shows and such, but you see this all the time.
>> Dr. Parker: Yeah, they're definitely their family members and sometimes they like him more than their family, I think. But yeah, these are really close companions. And you know, some people will travel everywhere they go with their dog and they're not taking their health lightly.
>> Diego (interview): Well speaking of your participant’s humans, because without them samples from dogs would have no way of getting to you. How can people get involved? I've read that your goal is to get to 10,000 genome sequences. How many have you done so far? And what's the process for getting the samples you need?
>> Dr. Ostrander: So there's sort of-- it's like different levels of membership. So, we've amassed a collection of 35,000-40,000 DNA samples in our freezers. It's always important to say is, you know, we don't breed any dogs, we don't keep any dogs in kennels, we don't have dogs NIH. This is a citizen science project. Those 40,000 DNA samples or 35,000, you know, they come from people who over the last 30 years have contacted us and said, you know, I have a Leonberger, I have, you know, a Toy Poodle, I have a Chinook, which is a very unusual breed for the most part. And we have people in the lab who send out kits and instructions and consent forms. And people get the blood drawn or the cheek swabs and they send them back to us, and that's been going on for 30 years, almost, 28. And that's why we've been able to be successful. And of course we share those, you know, we send those out and share those or send out the data and of course, all of our sequenced data is made public as well.
>> Dr. Parker: And you can believe it, we still need more.
>> Dr. Ostrander: And that's exactly right, because we want representation of every breeding population worldwide. We started obviously with American Kennel Club breeds in the US, but we're really extended worldwide now. You know, we have collections going on in so many different places, particularly in East Asia. That's a big focus right now. And we have a project called the Americas Dog Project, where we're collecting throughout the Americas and we're underrepresented in South America, so a lot of energy is going there. And not just breeds, but populations. So, we just published a paper on Highland Wild Dogs and New Guinea Singing Dogs in Indonesia. So, we're always interested in rare and unusual breeds, always. So, you can just go to our webpage and you can send us a message and we can tell you if we'd like to get a sample or not.
>> Dr. Parker: Yeah. And yeah, as Elaine says we're collecting all these rare breeds and things that are regional that really don't show up outside of sort of the area where they originally were developed. But we actually still need a lot of very common breeds, too. It's surprising. We go through a lot of DNA. And it was almost like, oh my goodness, we're running out. We need more, you know, Chihuahuas. They're not exactly rare but, you know, we still need them.
>> Dr. Ostrander: And so, we have directed projects where we're doing whole genome sequencing, like the cancer projects, the ancient DNA projects. But we want to build a database of 10,000 dog genomes shared for everybody. Everyone can have it, no strings, no constraints, no permissions. It's all publicly available and that's it. And it's really supposed to facilitate dog genome research everywhere, all over the world, and especially for people who don't have unlimited money to do sequencing, right, which is everybody. And so, the first phase was about 3,000 dogs and that's pretty much done at this point in time. And that does include some samples that we just did here at NIH on our own that we're adding into that. And we want representation of everything that's common, as well as niche populations. So anyone who's traveled or who lives part-time in, you know, anywhere interesting or maybe even not interesting where they can sample dogs, can always go to our Dog Genome webpage. And just send us a message and, you know, hopefully your dog will fit somewhere into the Dog Genome Project because we would love to include everybody that we can.