This web page was produced as an assignment for an undergraduate course at Davidson College.

Paper Review: The Origin of Domestic Dogs

Paper Summary

Research Question and Approach: Thalmann et al. (2013) aimed to resolve the current discrepancy between genetic data and fossil remains with respect to the time and geographic origin of dog domestication. Genetic data indicates domestication began 15,000 years ago in East Asia while the oldest fossils found thus far suggest the start of domestication in Europe more than 30,000 years ago. Because modern dogs currently demonstrate significant variation in phenotypes and geographic dispersion as a result of the domestication process, the beginning of that process is an intriguing unknown. In order to make a best estimate about the origin of the domestic dog, this group utilized mitochondrial (mt) genomes from 18 ancient canids in comparison to mtDNA of modern wolves, dogs, and coyotes for a total of 148 modern genomes. They completed a phylogenetic analysis to assess the possible ancestors of modern dogs and as well as a Bayesian analysis to evaluate the demographic history of a particular subset of dogs and canids.

Conclusions: The authors indicate that the phylogeny established by mtDNA analysis reveals conclusive evidence that the modern dog origin lies in the European canid population instead of the East Asian population. All modern dogs demonstrated relations to either ancient or modern European canids, with no indication of an East Asian predecessor. The authors suggest that dogs and wolves diverged >15,000 years ago and hypothesize that domestication resulted from close geographical interactions between hunter-gatherers and early dog ancestors.

My Opinion

This article’s statements of overall trends and conclusions are very clear. The first two figures and the body content consistently support the conclusion that dog domestication originated in Europe rather than East Asia. The last figure and its interpretation add additional context to the trend over time. The paper also provides an excellent context to the research question and explains well why previous studies have contradicted one another. At the same time, this paper could likely have been more powerful with a few additional details and clarifications.

With respect to research presentation, the figures are somewhat clear to understand, but unfortunately the main message of each figure has to be found buried in the body of the paper instead of the figure caption. Figure 2 is the least clear. The decline between 5,000 and 2,500 years ago is never explained. In addition, if parallels between human or wolf population growth are to be drawn in the discussion of the figure, this population data should ideally be included as a comparison in the figure itself. Table 1 could also have been a more powerful visual if the authors had arranged the information in groups A and B in some particular order (e.g. by classification, approximate age, region of origin, etc.). 

With respect to research content, this paper, like any other paper, is limited in length, but it feels as though some important research details are lacking. For example, context as to how ancient samples were accessed was omitted. The paper does an excellent job explaining the context of the research, the results of previous research, results, and study limitations. However, a more detailed description of methods used to compare and analyze the 166 genomes is needed for a stronger study summary. They also fail to explain the relation between the terminology of canids, wolves, and dogs. This may be self-explanatory to some researchers but likely not all. Another small, necessary improvement is resolving the discrepancy between use of the word “Pre –Colombian” (872) and “Pre-Columbian” (figure 2 caption).



Figure 1: This key figure establishes precise ancestries for wolves (orange) and modern and ancient dogs (blue). The grey specimens did not have a clear taxonomic classification. This phylogeny was established through Bayesian, maximum likelihood, and coalescence-based analyses, and bootstrap values were calculated for further statistical analysis (significance, defined as >90% bootstrap value, identified by asterisks). Interestingly, all modern dogs and most wolves fall into 4 clades. While this figure carries a lot of data, the most important piece is that each dog clade, regardless of its diversity (triangles), is more associated with European ancestry than East Asian ancestry. Clades A, C, and D have closes phylogenetic associations with ancient canids while clade B has associations with modern wolves, but all associations are to specimens of European origin. Based on these associations and time to most recent common ancestor (MRCA) derived from the phylogeny, researchers conclude that domestication originated in Europe between 32,100 and 18,000 years ago. This date range implies that dog domestication occurred prior to agriculture-based societies and instead amidst hunter-gatherer societies. Researchers also point to the grey specimens as indication of failed domestication attempts.


Table 1: This table details important sequencing and morphological characteristics of the 18 ancient specimens. Group A was identified through custom capture arrays while the mtDNA for group B was sequenced after PCR with particular biotin-enhanced primers. The table demonstrates a wide range of coverage levels as well as approximate ages of the samples. Samples were also classified morphologically as dog-like or wolf-like, with the 2 samples of ambiguous morphology (the grey specimens in figure 1) in italics. This table demonstrates the wide variety of coverage, approximate age, and geographic origin regions among the ancient specimens but ultimately confirms that the oldest specimens more related to dogs are from Europe.


Figure 2: This figure demonstrates how dog population size changed over time. The Bayesian Skygrid plot follows the population size (y-axis) of modern dog clade A and its ancient sister groups from 20,000 years ago to today. The middle line is the median effective population size (median logNe) while the light blue surrounding region shows a 95% confidence interval from a HPD (highest probability density) calculation. This figure confirms the pre-existing hypothesis of the two-phased progression of domestication. An overall trend of increase in population size from 20,000 years ago to 5,000 years ago is associated with early domestication. In contrast, the increase in size from 2,500 years ago to present is suggested to parallel and link to human population growth. This analysis demonstrates the overall trend of growth in population size as the canid was domesticated over a 20,000-year period.   


Thalmann O., et al. 2013. Complete Mitochondrial Genomes of Ancient Canids Suggest a European Origin of Domestic Dogs. Science. 342:871-874.

Betsy's Home Page

Genomics Page
Biology Home Page

Email Questions or Comments.

© Copyright 2014 Department of Biology, Davidson College, Davidson, NC 28035