The Mitochondrial Eve

[brogers]

The concept of a 'mitochondrial eve' is a concept that has been around in human terms for some time. The idea is that we as humans all trace faithfully back to one female, or the 7 Daughters of Eve if you have read Bryan Sykes interesting novel on human mitochondria, and it is borne about by one of the principal achievements of mitochondrial genetics, which is the classification of all modern Europeans into seven groups, the mitochondrial haplogroups. Each haplogroup is defined by a set of characteristic mutations on the mitochondrial genome, and can be traced along a person's maternal line (your mother, to her mother, to her mother, etc) right back to a specific prehistoric woman. In his book, Sykes refers to these women as "clan mothers" (subsequent research has identified 29 of these and this is just in Europe) and all these women in turn shared a common maternal ancestor, the Mitochondrial Eve.

In terms of horses we have had similar attempts at classifying the mitochondrial lineage, although it must be said that the early protagonists, the likes of Goos, Franzel, Lowe and Bobinski, who faithfully arranged the thoroughbred in terms of female families using numbers, had no idea that their arrangement had genetic merit and certainly didn't do their work on that basis. The complete mitochondrial DNA sequence of the horse was first discussed by Xu & Arnason (Gene; 1994), and there has subsequently been a lot of papers discussing the phylogeny of the horse including some interesting papers including Jansen (Proc. Natl Acad. Sci 2002) and Ceislak (PlosOne; 2010). In regards to the thoroughbred, in terms of history Bower (Biol. Letters 2010) has written some interesting papers but commercially, in some sense, the most relevant papers for the thoroughbred industry have been that of Hill (Int. Soc for Animal Genetics; 2002), which first discussed errors in the stud book in terms of the Lowe/Bobinski/Shirai numbering, and then Harrison (Mitochondrion, 2006) who looked outside the d-loop and towards athletic potential (mitochondria after all is important in cellular respiration). There is a paper by Binns, et al entitled "Thoroughbred mitochondrial DNA reveals closer than expected links between maternal genetic history and pedigree records" that has been submitted to the Equine Veterinary Journal that looks particularly interesting in this vein.

It seems now, that following the work of Jansen and Ceislak, a research team headed by Alessandro Achilli, a researcher in the department of cellular and environmental biology at the Universita di Perugia in Italy, has analyzed equine mitochondrial DNA to find its own Mitochondrial Eve. The paper, which was released today and is free to download, outlines an interesting history of wild horses (equus ferus) living throughout Europe and Asia during the Paleolithic period. The authors note that many of the different mitochondrial haplotypes, and thus differing breeds, probably didn’t survive the peak of the last glacial period, from 26,500 to 20,000 years ago and another later period that covered Europe in ice. The study revealed 18 major haplogroups (A–R) across all breeds (so strictly speaking the concept that there are 50+ female families in terms of Shirai numbers in the thoroughbred is not correct), and the root of the tribe corresponding to an Ancestral Mare Mitogenome born approximately 130–160 thousand years ago. All haplogroups were detected in modern horses (E. caballus) from Asia, but the F haplogoup was only found in E. przewalskii—the only remaining wild horse.

http://www.pnas.org/content/early/2012/ ... &view=FitH

[Pan Zareta]

The study revealed 18 major haplogroups (A–R) across all breeds (so strictly speaking the concept that there are 50+ female families in terms of Shirai numbers in the thoroughbred is not correct)

While I agree it's extremely unlikely that each of the numbered (by Shirai et al.) TB families represents a unique mitchondrial lineage or even unique founder matriarch ~1650-1750, the fact that Achilli et al. identified only 18 different haplogroups has little or no bearing on estimates of their actual number. Spread among those 18 haplogroups they reported 81 different haplotypes few, if any, of which are apt to share a common ancestress within the last ~1000 years, much less the last 362 years. Assuming this report has identified all haplotype diversity within the domestic horse (which they do not claim and is unlikely), the number 81 becomes a potential limit on the number of mitochondrial lineages. It can be further refined by counting the number of haplotypes reported in haplogroups known to occur in the TB.

In practice, I believe Harrison reported 33 mtDNA haplotypes at the 2011 symposium - ? Whatever the number, his is probably the best estimate at present of mitochondrial lineages present in the breed. The actual number of founder matriarchs ~1650-1750 most likely exceeds 33 but is considerably less than the sum of all (British & colonial) numbered families. (This distinction between # of mt lineages and # of founder matriarchs is no doubt meaningless from a performance standpoint, but it's of considerable interest to breed historians. )

Imo, the most important point made by this report is the utter inadequacy of using a single hypervariable section of d-loop alone, as was done in two published studies of TB mtDNA (Hill et al. 2002, Bower et al. 2010), to characterize mtDNA haplotypes. Harrison & Turrion-Gomez acknowledged this shortcoming in their study reported 2006 which as you note looked outside the d-loop, incorporating ~80% of the coding region in their TB mtDNA analysis. That's certainly an improvement upon use of d-loop alone. But having compared their published sequences with those of Achilli et al. I must say that the 2006 report, even as amended 2009, seems rather unlikely to have defined all or even most of the performance-critical variants in ATP6, ND4, & ND5.

[brogers]

I have actually had correspondence with Dr Achilli. He agreed that the use of one HVS (~400bp) which Hill and Bower both used was inadequate, ands could lead to errors, but that the use of two HVS (~800bp) of the d-loop would be significantly better. Ideally you would be doing the lot, but that is still quite expensive in commercial terms to do and it really only just classifies the horses into families.

He added that given the close stud book and the fact that the thoroughbred has been selected for performance most likely meant that there was not going to be variants within the mtDNA that promoted performance per se, rather that the mitochondrial encoding nuclear based genes would have variants within them that could be specific to the haplogroup.