On the antiquity of dogs in America

Always on the look out for evidence on an early peopling of America I stumbled upon a recent paper by Witt et al. (2015) [1] which looks into the origin of Native American dogs through the genetic analysis of the remains of ancient and contemporary dogs.

The paper is very interesting and is based on the usual orthodox assumptions on the peopling of America: that is, that the ancestors of native Americans reached the New World 15,000 years ago. And this in my "unorthodox" opinion leads to conclusions which the authors find "unexpected" but actually are the outcome of those erroneous orthodox assumptions.

Science builds on the successive discoveries of scientists, different branches borrow the findings from others and this results in the continuous growth of human knowledge.

That is why the scientist studying dogs takes the data from humans at its face value and does not question the figure of 15 ky for the initial peopling of America. It would be quite silly if each scientist had to validate the data borrowed from other fields of knowledge.

But if knowledge is incomplete or mistaken, then the conclusions are also incomplete or mistaken. I remember reading as a teenager about the efforts of a Victorian scientist, Lord Kelvin, to calculate the age of the Sun. He used the knowledge of his days and found it to be not older than 50 - 100 My. (Read his paper), which seemed to be in contradiction with the age of the Earth.

But of course nineteenth century science did not know about nuclear fusion reactions and Kelvin based his calculations on correct and sound chemical reactions, and instead of an age of +5 billion years, he obtained a figure two orders of magnitude smaller than the actual value.

A similar situation arose when Hubble calculated the age of the Universe it was only 1.5 Bn years old, younge than the Earth. It took another discovery (Baade and the Population II stars) to fix the mistake.

Having said this, let's look at Witt et al.'s paper and its assumptions and conclusions:

EBSP and its shortcomings

Witt and his team calculated the Extended Bayesian Skyline Plot (EBSP) for humans and dogs in America. And while humans showd the classic "S" or sigmoid curve in the time ⁄ popultion size graph, dogs did not.

The dog population plot is a line with an almost nil gradient, in other words: their population appeared suddenly 9 ky ago and did not grow over time, it remained stable.

The expected graph for a group peopling a new territory is the one observed in humans (an "S"): initially the migratory population is small, and it gradually enters a period of growth when the new territory's resources are exploited; this allow an increase in natality over mortality rates. All members of the new population have abundant food (a critical resource for hunter - gatherer populations), also predators are rare or not accustomed to the new expanding population.

This growth period is followed by a transition phase when either mortality grows or natality falls due to the decrease in resources or contact with new diseases or a limitation to expansion in a finite territory (deserts, mountain ranges, limited landmass -islands).

As growth slows it finally reaches a period of plateau, stability: there is no more growth. Population stabilizies. Birth and death rates are equal. Resources are scarce and parasites & disease are taking their toll. There is an equilibrium: more births means less resources and therefore more deaths. Which pushes the population downwards for a while, but this frees some resources and this leads to less deaths and more births pushing population up a little. The outcome of these ups and downs is stability.

In other words, the carrying capacity of the environment has been reached.

But what did Witt et al. find? Human population evolved as per the "S" curve but dogs did not. This is very odd. They attempt to explain things as follows (bold highlighting is mine):

"The EBSP of the dogs shows a relatively stable population from the present to the time of the most recent common ancestor, ~9000 years BP..."
"When considering both the human and dog EBSPs, the population stability found in the dog plot is unexpected, as one would expect the dog population to increase in size over time, as the human population did. Intriguingly, this unexpected finding might indicate that humans were controlling dog matings and effectively breeding them, or that the population of dogs in the Americas quickly reached long-term carrying capacity. The stable population of dogs had a median effective population size of ~1000 female individuals at the time of the first dog burial in the Americas. This is consistent with other estimates of dog effective population size that suggest a global population of roughly 10,000 female dogs at the same time (Thalmann et al., 2013).
Additionally, the coalescence date estimate for our dataset of American dogs is also surprising, given that by ~9000 years BP dogs should have been established across the Americas as they are thought to have arrived in the Americas ~15,000 years BP with humans. However, it is interesting that from 9000 years BP to the present, the population of dogs in the Americas roughly mirrors that of humans, in that both are stable for long periods of time." [1]

See their graph below. The blue curves (A) and (B) are for DOGS, the red ones (C) and (D) are for HUMANS. The human plot displays the "S" curve (well it is inverted because of the X axis position) the dogs is flat.

Well, (B) is quite flat and so are the median dashed and solid blue lines in (A). But look at those error bars (the violet shaded region in (A)), that does not give me much confidence in the Confidence Interval of that EBSP.

Could these large CI mask some incorrect assumptions? and, just how reliable is a EBSP?

Heller R, Chikhi L, Siegismund HR (2013) have their doubts about the widespread use of this method under all circumstances, and they caution:

"The intuitive, visual appeal coupled with a real risk of erroneous demographic inferences make BSPs vulnerable to misinterpretations. It should be underlined, however, that this is not a shortcoming of the methods themselves, but rather an under-appreciation of the dynamics of coalescent intervals in structured genealogies.
... Demographic inference methods based on the coalescent usually assume panmixia, i.e. the absence of population structure, although this is not a realistic assumption in many biological situations. A number of recent studies have investigated the effect of violating the panmixia assumption for inferring population size changes. These studies suggest that population structure can lead to erroneous conclusions about demographic changes in a population that in fact has remained stationary through time." [2]

EBSP for dogs and humans.Witt et al., [1]

So, it seems that Bayesian Skyline Plots have to be used with care and the population structures taken into account.

Another look at the data

Why would dogs suddenly appear 9 kya? Humans came to the New World 15 kya without dogs? Why is the human curve an "S" and the dogs' appears suddenly and with a stable non-growing population?

In my opinion what is missing in the charts is the growth and transition phases of the "S" curve. Dogs did arrive in the New World, but not 9 kya, they arrived 20 or 30 kya, or probably earlier; they probably even originated there and spread to Asia.

An early arrival date gives us room for the "S" curve to develop. Below I have extended the EBSP back in time, with two different options (yellow and organge curves) with different growth rates, just as examples.

Another proposed population chart for American Pre-Columbian Dogs. by Austin Whittall from Witt et al., [1]

Why would their most recent common ancestor date to only 9 kya? Does the fact that dogs and wolves interact and mate affect the validity of TMRCA calculations?

The interesting point in Witt et al.'s paper is that wolves and dogs are an intertwined family, that did not evolve separately like humans and chimps did. No; wolves and dogs mated repeatedly over time, exchanging genes and probably that too alters the conclusions of any EBSP.

The scientists write that the unusual ancient dog haplotypes that they found could have been the result of any of three possibilities [1]:

"Overall, the wolf and dog haplotypes seem to be mixed throughout the tree. The haplotypes from the ancient dogs that are identified as outliers in Fig. 2B are likely (i) founding haplotypes present in the dog population, (ii) haplotypes that show admixture with North American wolves, or possibly (iii) the result of a separate domestication of dogs in the Americas.
Notably, multiple outlier haplotypes are more closely related to wolves than dogs, suggesting admixture or independent domestication. One haplotype from an Alaskan dog is not closely related to modern dog or wolf haplotypes, whilst two other outlier haplotypes cluster more closely with modern dog haplotypes than wolf haplotypes. Interestingly, the latter two outliers cluster with some haplotypes that are only found in Siberia, the Americas or eastern Asia (Pang et al., 2009), possibly indicative of a Northeast Asian origin, consistent with the origin of Native Americans." [1]

Very interesting: dogs could have been separately domesticated in the New World, and some dogs are not even close to modern dogs or wolves... but we are expected to believe that they arrived "late" in America, only 9,000 years ago and that the human hand guided their breeding to keep their population stable...

This seems to contradict the explanations given by Witt et al.: (bold is mine)

"There are a total of eight outlier haplotypes shown in Fig. 2B that differ significantly from other ancient dog haplotypes, and there are multiple reasons why such outliers might be present in our dataset.
First, these haplotypes represent dogs that came to the Americas via Siberia that exhibited haplotypes different from the other founding haplotypes, and represent additional founder lineages.
Second, these dog haplotypes are shared with wolf haplotypes that have either gone extinct or have yet to be sampled and published in the literature. If they are shared with wolves, admixture or separate domestication are both viable possibilities.
Interestingly, some of these haplotypes cluster more closely with dog haplotypes, suggesting that these dogs represent different founding haplotypes – they are likely too distantly related to the other haplotypes in the Americas to have diverged from another founding haplotype, but are more closely related to contemporary dog haplotypes than wolf haplotypes. Notably, some of the modern dog mtDNA haplotypes that are most closely related to the outliers in Fig. 3, such as A31, A121, and B28, are exclusively found in Asian dogs, further suggesting that these were founding haplotypes

...
However, it is surprising that wolf and dog haplotypes are so thoroughly intermixed in Fig. 3, when in Fig. 2 the wolf haplotypes seem far more distant from the dog haplotypes." [1]

What is clear is that dogs and wolves are really mixed up, that Asian dogs share mtDNA haplotypes with American ones. The outliers are always interesting, and are always brushed aside as "exceptions" when they usually are the key to understanding the problem. What is not clear in this paper is how could dogs have originated in only 9 ky.

Ancient Dogs

Which brings us to the antiquity of dogs. The oldest "dog" is the famous Goyet dog. A skull found in Belgium which as dated at 31.7 -36 kya, it seems to be that of a prehistoric dog. But it is not ancestral to modern dogs or contemporary wolves, it is a line that died out, an ancient domestication event [3]. Alternatively it may belong to a new gray wolf population that has not yet been described.

Belgium is far from America, but Altai (home to the mysterious Denisovans, home to Neanderthals and crossroads of Asia and Siberia) has produced the remains of another ancient "dog". Ovodov et al., (2011) [4] report the finding of a 33,000 year old dog skull found at Razboinichya in Altai.

They did some bone comparisons and found that the skull is within the range of Neolithic and later dogs, that its teeth are large and wolve-like, they concluded that "this specimen may represent a dog in the very early stages of domestication, i.e. an “incipient” dog, rather than an aberrant wolf." [4].

But they assume that the Altai dog is a dead end because "this canid material pre-dates the LGM and additional putative dogs are not found until thousands of years later, in the late Glacial – early Holocene (ca. 14,000–11,500 cal BP), we conclude that the lineage represented by the incipient dog from Razboinichya Cave did not survive the LGM" [4].

I am surprised that they consider this dog a dead end. The lack of evidence is not evidence of lack. These dogs could have moved on, into America or evolved into later dogs. The lack of bones may simply mean that they have not yet been found.

Druzhkova et al., (2013) [5] took another look at the Altai dog and conclude that it is a "Primitive dog" and they included a comparison with different canids in their paper:

"In order to further evaluate these close relationships we analysed the pairwise genetic distance, assuming a Kimura 2 Parameter substitution model (Figure 4) and found that haplotypes clustering in Clade A had the smallest genetic distance (2–5 differences on an alignment of 413 nucleotides) followed by the second smallest distances existing in comparisons to pre-Columbian dogs (1–9 differences), and the largest distance to contemporary wolves (4–26 differences). Moreover, when comparing distances derived from all dog haplotypes to that from all wolves, we found a significantly smaller genetic distance of the Altai specimen to modern dogs...[5]

There is a figure in their paper which is startling: this 33,000 year old Altai dog is placed in a branch of the tree with Pre-Columbian dogs and Beringian Wolves, some of which branch off before the Altai dog does! The Clade A dogs which has almost all contemporary dog sequences and "contains the majority of dog haplotypes (45 out of 72) including diverse breeds such as Tibetian Mastiff, Newfoundland, Chinese Crested, Cocker Spaniel or Siberian Husky" [4] is on that same "branch", but is slightly older.

Altai dog and other canids. From Drushkova et al., [5]

There is another Pre-Columbian dog (Clade B) as an outlier on a branch with Old and New World Wolves, isolated from all other dogs.

Druzhkova et al. conclude that: "...this preliminary analysis affirms the conclusion that the Altai specimen is likely an ancient dog with a shallow divergence from ancient wolves. These results suggest a more ancient history of the dog outside the Middle East or East Asia, previously suggested as centres of dog origin. [5]

Ancient American dogs with Asian roots

van Asch et al., (2013) [6] corroborate the Asian ties of American dogs and find that a haplotype (D28) which was believed to be 'Ancient America-specific' was also detected among East Asian dogs and therefore "all ancient American sequences except one (D40; possibly the result of dog–wolf crossbreeding) can now be linked to haplotypes present in East Asia or Siberia. We also note a large proportion of unique haplotypes among the ancient [American] samples." [6]

The "uniqueness" of American dogs appears time and time again in these papers as does the "dog-wolf crossbreeding" episode. These events, in my opinion, point towards a very ancient origin for American dogs, opposit to what Witt et al. suggest. [1]. Furthermore, van Asch state that "the American dog population was not formed through a severe genetic bottleneck, but that several lineages were brought to America from Asia in the pre-Columbian era, probably in several waves of migration." [6] , which contradicts a recent 9kya origin for these dogs.

A recent paper by Freedman et al., (2014) [7] comes up with some dates that differ from those given above (i.e. the +30 ky for the prehistoric dogs of Altai or Belgium):

"Our divergence time estimates imply that dogs and wolves diverged 14.9 thousand years ago (kya) with 13.9–15.9 kya Bayesian 95% credible interval (CI), assuming an average mutation rate per generation of µ = 1×10-8 and three years per generation (Figure 5A). Divergence times between wolf populations were tightly clustered at 13.4 kya (11.7–15.1 kya), and divergence between dogs was estimated to have occurred slightly more recently, at 12.8 kya (11.8–13.7). [7]

In other words, dogs only evolved 12.8 kya... so did the Early Americans (15 kya according to Orthodoxy) enter the New World without dogs? and, Witt et al., were right, dogs reached the New World 9 kya (and very likely distant parts of the Old World at the same time)? It seems unlikely. The culprit for a recent divergence date may lie in the generaton lenght or the "molecular clock" ticks assumed by the authors.

This paper was trying to find out if dogs arose from multiple domestication events (i.e. the Australian dingo from Asian wolves and the basenji in Africa from some west Eurasian wolf). Instead they found that dogs are very closely related to each other (may I add, the outlying Pre-Columbian dog is an exception) and arose from a now extinct ancestor of modern wolves. They also frolicked in the bushes with wolves during the initial domestication period, thus genes were swapped and this makes their exact history difficult to understand.

Closing comments

We have reviewed several recent papers that propose: dogs appeared 30 kya and became extinct. American dogs appeared 9 kya, suddenly, as a stable population. American dogs are closely related to East Asian ones. Dogs as a group appeared 12.8 kya. Domestication took place once but wolves admixed early on in the process with dogs. Domestication may have taken place separately in America....

Confusing? yes indeed.

What is clear is that Pre-Columbian American dogs cluster together with the 30 kya Altai dog and East Beringian wolves. That the only outlier among dogs is a Pre-Columbian dog (Clade B) in [5]. That all other dogs are close to, but not in the same branch as Altai, Pre-Columbian dogs and Berigian wolves.

What if dogs originated earlier, in isolation in the New World, and then moved West, across Beringia into Eurasia. A band of Neanderthal or some other archaich human reached America, domesticated New World wolves and created the first dog. Dogs dispersed form there to the rest of the Old World.

Sources

[1] Kelsey E. Witt, et al., (2015) DNA analysis of ancient dogs of the Americas: Identifying possible founding haplotypes and reconstructing population histories. doi:10.1016/j.jhevol.2014.10.012. Journal of Human Evolution Volume 79, February 2015, Pages 105–118
[2] Heller R, Chikhi L, Siegismund HR., (2013) The Confounding Effect of Population Structure on Bayesian Skyline Plot Inferences of Demographic History. PLoS ONE 8(5): e62992. doi:10.1371/journal.pone.0062992
[3] O. Thalmann et al., (2013). Complete mitochondrial genomes of ancient canids suggest a European origin of domestic dogs. Science, doi:10.1126/science.1243650
[4] Ovodov ND, Crockford SJ, Kuzmin YV, Higham TFG, Hodgins GWL, et al. (2011) A 33,000-Year-Old Incipient Dog from the Altai Mountains of Siberia: Evidence of the Earliest Domestication Disrupted by the Last Glacial Maximum . PLoS ONE 6(7): e22821. doi:10.1371/journal.pone.0022821
[5] Druzhkova AS, Thalmann O, Trifonov VA, Leonard JA, Vorobieva NV, et al. (2013) Ancient DNA Analysis Affirms the Canid from Altai as a Primitive Dog PLoS ONE 8(3): e57754. doi:10.1371/journal.pone.0057754
[6] Barbara van Asch, Ai-bing Zhang, Mattias C. R. Oskarsson, Cornelya F. C. Klütsch, António Amorim, Peter Savolainen (2013), Pre-Columbian origins of Native American dog breeds, with only limited replacement by European dogs, confirmed by mtDNA analysis. DOI: 10.1098/rspb.2013.1142Published 10 July 2013 PubMed 23843389 The Royal Society
[7] Freedman AH, Gronau I, Schweizer RM, Ortega-Del Vecchyo D, Han E, et al. (2014) Genome Sequencing Highlights the Dynamic Early History of Dogs. PLoS Genet 10(1): e1004016. doi:10.1371/journal.pgen.1004016

---

Patagonian Monsters - Cryptozoology, Myths & legends in Patagonia Copyright 2009-2014 by Austin Whittall ©