Maria Hjørungdal Bigset

Wildlife Ecology

Research article: Thomas Jansen, Peter Forster, Marsha A. Levine, Hardy Oelke, Matthew Hurles, Colin Renfrew, Jürgen Weber, and Klaus Olek (2002). Mitochondrial DNA and the origins of the domestic horse

Review and criticism of the researh article ”Mitochondrial DNA and the origins of the domestic horse”

This article is about research done on the domestication of the horse. The goal of the research was to find out whether the horse was domesticated from one or several popultations of wild horses. The article claimes that this is important knowledge because the domestication of the horse was significant for the exploitation of the Eurasian steppe, hence important to prehistoric archaeology. According to the article some people claim that it also was important for the development of the Indo-European language family.

The method used in this experiment was one where mitochondrial(mt) DNA samples were collected from 25 oriental and European horse breeds – including the American mustangs. This resulted in 318 mtDNA sequences which were added to the already existing database of horse mtDNA. The total of sequences was now 652, which is the largest database available. The article provides an explanation of why this exact method was used by pointing out previous experiments and the results of these: one experiment was that of Vilà et al. which sequenced ancient DNA from the Alaskan horse remains preserved in the permafrost. These are undeniably wild horses, “with dates ranging between 12,000 and 28,000 years(y) ago”. Since they assumed that the genetic uniformity found in the samples were typical for a wild horse population, the researchers concluded that the domesticated horse probably developed from several different wild horse populations. This because the modern horses have such a large variety of matrilines. Even though this is a likely result, this research article states that it is legitimate to question whether this population “is representative of the genetic structure of wild horses at the relevant place(s) and time(s).” After all, the last ice age, ending 11,400 years ago, may have isolated this wild horse population and reduced their mtDNA range. In addition these horses lived far away from places of domestication and even thousands of years before domestication. There were some concerns regarding mtDNA clustering as well, so all in all the conclusion of the article is that even though the things descovered by Vilà et al. are important, they are not really good enough to give an absolute explanation in the matter of the domestication of the horse and how it has developed through the years.

In the research (phylogenetic analysis of the 652 mtDNA samples) performed by Jansen et al. several things were taken into consideration: mutation rate; Estimation of number of wild mares contributing their mtDNA to the domestic horse (based on mutation rate); Comparing number and diversity of mtDNA of wild mares to the wild Przewalski and Alaskan mtDNA types.

When selecting which horses to use for the sequencing, they let the breeders of the different breeds choose because they were able to document the ancestry of each horse for at least five generations. Also, only horses with documented geographic origin of their maternal ancestry that corresponded to the origin of the breed were used. This was to exclude modern inbreeding as far as possible. For the geographical analysis only 331 of the 652 horses were used.

mtDNA control region sequences were used. These were available from the GenBank or from publications. The result was the 652 sequences which were used in the experiment. The GenBank accession numbers and references are listed in the article, but will not be mentioned here. A map is also provided with information on where in the world the different samples were collected from. This map shows that most of the samples were collected in Europe, only a few from Asia and Africa. It is concluded that the samples cover Europe, Morocco and Arabia quite thoroughly, but not so much eastern Asia.

The article provides a detailed description of the methods used, which will not really be emphasized here. However it can be mentioned that they in the phylogenetic analysis shortened the sequences to accommodate already published sequences.

Mutation rate was estimated by comparing the mtDNA tree of the modern equids with the osteological tree of modern and fossil equids. What they found was that the mountain zebra split off first followed by the asses, then the Damara and Grant zebras, the hemiones thereafter and finally the horse(including the Przewalski horse). Looking at these results together with archaeological findings, the maximum and minimum mutation rates could be determined. However the scientists themselves does not seem to think that the estimation of mutation rates is accurate enough.

After this the researchers could start the job of estimating the population size. They considered the earliest date for horse domestication(9400 BC) together with the mutation rate in a quite complicated process. What they found was a minimum estimate. There are five reasons why it was a minimum estimate: the first is that the method eliminated “the possibility that several mares could have the same mtDNA type while others would have died out”. The second was that “the fastest mutation rate estimate was chosen”. The third that “frequent (…) mtDNA types were collapsed at the same proportion as singleton types”. The fourth that “9400 BC is probably an unreasonably early date for the beginning of horse domestication”- the date is probably somewhere between 9400 BC and 2000 BC. The fifth reason is that their “sample does not include all currently existing DNA types”.

Now the mtDNA sequences from the domestic horses, Przewalski horses, 1000- to 2000- year old samples found in southern Sweden and Estonia and 12,000 to 28,000 year old Alaskan permafrost remains were pooled and the unrooted evolutionary network was reconstructed. There is a figure showing this tree in the article (fig.2) and it is quite complex. The root of the network was identified by using the most closely related outgroups of equids, namely the Equus grevyi, Equus kiang, and Equus hemionus.

What was found then was that the Alaskan fossil equids fits within the Eurasian mtDNA variation. Previously it was thought that the Alaskan mtDNA was a sister branch to the Eurasian’s. By doing some more adjustments they also found that the latest possible date for the first modern caballine horse is 300,000 years ago.

Another discovery was that of 17 very frequent mtDNA types which are “old enough to have developed a star-like branching structure”. This told the researchers “that the ancestral type of each has had greater breeding success compared with other ancestral mtDNA types in the network which are poorly developed or even extinct.” A fact that is historically interesting.

The scientists found that certain clusters of mtDNA was geographically restricted to different areas, and therefore also were linked to breeds. For instance there was a cluster “belonging” in Central Europe, the British Isles, and Scandinavia (incl. Iceland). This include the Exmoor, Fjord, Icelandic and Scottish highland ponies – all of which are northern European.

There was one cluster restricted to the Przewalski’s horses, including three mtDNA types. All of these were closely related and they were not found in any other breed. The scientists thus conclude that the Przwalski horse is not one of the ancestors of the domestic horse.

By using the mtDNA mutation rate (which they state, a number of times in the article, are not accurate enough) they determined the minimum number of mares domesticated from the wild. The result was that “77 successfully breeding mares is a minimum estimate for the number of wild mares ever domesticated”. These 77 ancestral types were nearly as varied as the whole horse mtDNA pool of today. They state that this is a whole lot more diverse than the monophyletic Alaskan wild horse DNA or the Przewalski wild horse mtDNA.

In the end the article proposes some questions about their findings and they give some possible explanations. For instance they propose an explanation on why the Przewalski horse was not used in the domestication of the horse and on why their mtDNA has so little diversity. They also ask if it is likely that the horse was domesticated independently by different human communities in different geographical areas. Their answer is that the beginning of domestication started in one area and then a diffusion of the knowledge happened.

Finally the article declares that to learn more about the ancient horse domestication a better calibration of the mutation rate would be necessary.

My personal view on this research (and on the article) is that it seems to be very thoroughly carried out. The scientists are aware of the weaknesses to their research and are very good at pointing this out. In my opinion this raises the credibility of the article much more than a very categorical article would have.

When I desided to read this article it was initially because I wanted to learn more about the Przewalski horse. Even though this article did not present too much information in this horse, I learned a couple of interesting things; Namely that the Przewalski horse seemingly is not one of the ancestors to the domesticated horse we know today! Even on the web-site of the “Foundation for the Preservation and Protection of the Przewalski horse”(FPPPH) it says that it is one of the ancestors of the tame horse [1] .”Our domestic horses have a number of ancestors, and since the nineteenth century the forest tarpans and the steppe tarpans died out, the Przewalski horse is the only remaining ancestor left.” [2]

Since this research article seems serious, and the methods they have used seems good, I am willing to believe that they are correct in their findings and that the FPPPH might want to update their information a little bit.

Another thing I learned is that the Przewalski horse has remained untamed for a reason. It has actually resisted domestication. When domestication was tried the wild horse got problems “such as pacing, aggression, impotence and infanticide” as stated in the article. This is confirmed in the “Particulars about the Przewalski horse”. 

I have not used a lot of sources when trying to confirm/invalidate some of the information in this research article because I figured that the FPPPH must be one of the best sources to correct information about this horse type, even though they seem to be a bit outdated when it comes to genetics.

[1] “Particulars about the Przewalski horse” by Jan Bouman, 1986 (has been updated), http://www.treemail.nl/takh, on this web-site it is possible to download a booklet on the Przewalski horse.

[1] Same as above (p. 5 of 33)