April 3, 2015

19 Ka BP old mtDNA H from Cantabria

Jean Lohizun points me to this new study by the EHU-UPV paleogeneticist team, which reviews the ancient DNA evidence from the Cantabrian strip (or "fringe" as they call it) and, most importantly announces the oldest confirmed mtDNA H to date: belonging to an individual from El Mirón cave (Ramales, Cantabria, bordering the Basque Country), which is assigned to earliest Magdalenian culture and dated to 19,000 BP.

Update (Apr 24): the genetic findings of El Mirón were first published (in Spanish) in: M. Hervella et al., El ADN mitocondrial de los cazadores-recolectores de la región cantábrica: nueva evidencia de la cueva de El Mirón, Revista Española de Antropología Física - Vol. 35 (2014). I could not find an online reference but I have a copy of the article thanks again to Jean.

Concepción de la Rúa et al., Ancient DNA in the Cantabrian fringe populations: A mtDNA study from Prehistory to Late Antiquity. Quaternary International, 2015. Pay per viewLINK [doi:10.1016/j.quaint.2015.01.035]


The present mtDNA study on human remains of fourteen archaeological sites from Cantabria, Basque Country and Navarra provided a diachronic overview from Paleolithic–Mesolithic to Late Antiquity period of some communities settled in the Cantabrian fringe. Ancient DNA studies in European human remains indicated a genetic discontinuity between the hunter–gatherers and later populations. However, some of the mtDNA lineages found in the Cantabrian fringe in Paleolithic–Mesolithic times persist in present-day populations.

The mtDNA variability observed in hunter–gatherers and farmers in Europe denoted a complex pattern for the Neolithic transition, occurring along several different routes into and across Europe. The mtDNA lineages found in the Cantabrian fringe indicated that the dispersion of Neolithic farmers had a different genetic impact in this area with respect to Central and Mediterranean regions of Europe. The differences in mtDNA variability were also apparent after the Neolithic, as shown by the genetic distance between the Chalcolithic populations from the Cantabrian fringe and the Bell Beaker Culture (BBC) populations of Central Europe. 

It must be mentioned that they seem to have forgotten the thesis of Marie Lacan[fr], which reported mtDNA H in Linatzeta cave (Basque Country, Epipaleolithic) and Franchthi cave (Greece, Meso-Neolithic transition), among other more recent aDNA sequences. See here for my English language synthesis.

Otherwise they list all the sequences considered in their paper in table 1:

Annotations by Maju: red: El Mirón (new sequence), orange: other Paleolithic or Epipaleolithic sequences

After including Linatzeta (Lacan 2011), we get the following frequencies in the Eastern Cantabrian sub-region (Basque Country + Cantabria) for pre-Neolithic times:
  • H: 4/6=2/3=67%
    • H-CRS (H1 surely): 33%
    • H6: 16%
    • H*: 16%
  • U5: 2/6=1/3=33%
    • U5*: 16%
    • U5b1: 16%
Notice anyhow that if we choose to draw a line between Cantabria and the Basque Country, then we get a sharp contrast: 
  • Cantabria: 100% H
  • Basque Country: 67% U5 + 33% H* (or even an illusory 100% U5 if we use only the table above, ignoring Lacan's data)
Does this apparent sharp contrast make any sense? Well, one possible interpretation comes from carefully taking into account the data we have on Iberian Solutrean and Magdalenian, which indicates that:
  1. South Iberian (from Valencia to Portugal) Solutrean is heavily influenced by the Gravettian substrate (otherwise "pure Solutrean" is restricted to two caves), configuring a unique facies sometimes called Gravetto-Solutrean.
  2. South Iberian Gravetto-Solutrean, probably in its way to Portugal, strongly affected the Upper Paleolithic of Northwest Africa, being a decisive force in the Iberomaurusian or Oranian genesis. Backflows can't be discarded because of the innovation of tanged and winged arrow points, which may have been inspired by North African Aterian technology.
  3. The Portuguese branch of this Gravetto-Solutrean was the actual source (via Salamanca) of Asturian Solutrean, unlike what happened in Cantabria and the Basque Country, directly influenced by Aquitaine. 
  4. In the subsequent Magdalenian there might have been an expansion eastward of the Asturian population, because the facies divide moves to the east (so we have a Cantabro-Astur facies and a Basque facies).
In addition to that, it may be worth considering the issue of North African genetic influence in the Western Third of the Iberian Peninsula, which incidentally and irregularly includes Cantabria but not the Basque Country. Also Chandler et al. 2005 reported high frequencies of mtDNA H (and low of U) in Epipaleolithic Portugal.

I guess that other interpretations are possible such a more subtle cline or patchy distribution but I would not discard this hypothesis, which in essence proposes that Solutrean and Magdalenian were in general dominated by U5 but this did not affect (at least not very intensely) most of Iberia, nor surely other regions like Italy or Eastern Europe, where we see haplogroups that are not U5 (Sunghir's and Karelian H, Italian mysterious HV, etc.)

This implies that the main redistribution of mtDNA H in Europe, that part organized around H1 (which also includes H3 and various H*) actually happened mostly in the Neolithic from areas like Portugal. However we know nearly nothing about the Atlantic pre-Neolithic DNA North of the Bidasoa River (some of which could also be H, particularly R*-CRS reported in Britain) so multiple sources are possible. The huge blank of data corresponding to the Western French State and also Atlantic Islands, etc. is crying for a comprehensive sampling, and not just for mtDNA. 

Neolithic Basque mtDNA is unlike what is found in Continental Europe

The authors pay limited attention to the issues relative to Paleolithic and dedicate most of the paper to analyze the Basque ancient mtDNA in contrast to other comparable data from elsewhere in Europe. This is synthesized in fig. 2:

Fig. 2. Multidimensional Scaling analysis (MDS), based on a Fst genetic matrix calculated from the frequency distribution of the mtDNA haplogroups of different populations [Neolithics (green), Chalcolithics (purple), Late Antiquity (red), present-day Near East and northern Caucasus (orange) and Europeans (black)]. Abbreviations for present-day populations in Europe: Eastern Mediterranean (MdE), Central Mediterranean (MdC), Western Mediterranean (MdW), Northeast Europe (NE), NortheCentral Europe (NC), Northwest Europe (NW), Southeast Europe (SE) and Alps (ALP). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.).

Notice that all the ancient mtDNA from the so-called "Cantabrian Fringe" in this graph is from the Basque Country (Navarre included), often outside of the Cantabrian strip and rather from the Upper Ebro basin. Notice also that the all the Chalcolithic data belongs to military contexts from the Ebro Valley and are probably therefore not representative of the overall Basque region, although they may represent well the Upper Ebro in that period (more influenced apparently by Mediterranean inputs of Cardial affinity).

Also I already discussed all this (and more) quite in depth in my dedicated entry of 2013

The authors argue that Germany's Bell Beaker samples (BBC) are not quite similar to the Basque ancient and modern pools, what is true if you are nit-picky enough, but they share the same common tendency in PC1 towards modernity. In contrast all other Neolithic samples are clearly non-modern European and must have been largely suffered by replacement in the Chalcolithic or later periods.

PS- I forgot to mention that apparently Paabo and co. had already sequenced this very same specimen in 2013, yet they have not published anything for unknown reasons, some think that ideological ones. It's of course possible that they do have good reasons but two years is a long time to await publication really, we the people, who pay their salaries and budget with our extreme economic pains, expect reliability from our well-paid researchers.


  1. I assume this is the Red Lady attributed to the Magdalenian

    First, I believe this would by far be the oldest attested H within 40-200 generations of H's estimated origin. Although I don't believe this super-haplogroup originated in Iberia, you could make that case given this result, and again we have a controversial result from Hervella and Izagirre with low resolution rCRS from HVR, and they ride off shooting pistols in the air.

    Taken at face value, I think it would mean that some of the earliest twigs of H spread to Iberia via Gravettian expansion and the haplogroup was broken into two east-west chunks, and the relative rise of H in modern populations is owed to some unusual process involving relatively isolated populations at the extreme margins of Europe suddenly rolling in to an imploding Late Neolithic.

    They need to get higher resolution than just rCRS from HVRs. There may be something else going on here like other unknown sister branches in the R tree that are confusing the ID. It could also be indeed H from Gravettian expansion but an oddball lineage like H86 with a very low or no modern frequency.

    1. "I assume this is the Red Lady attributed to the Magdalenian"

      I believe so, however I never heard of "the red lady" before today so for me it's just another ancient sample.

      ... "the oldest attested H within 40-200 generations of H's estimated origin."

      Only if you choose to believe in that pseudoscientific superstition called "molecular clock", particularly in some more popular versions of it, I personally prefer to use logic and archaeological references and, from that viewpoint, there are basically three time frames for arrival of any lineages to Europe from West Asia: (1) Earliest UP (50-40 Ka BP), (2) early Gravettian (c. 32 Ka BP) and (3) Neolithic (c. 8-7 Ka BP or also the later Vinca wave c. 5 Ka BP for lineages restricted to the Balcans and Italy).

      That means that, unless I'm proven otherwise, any pre-Neolithic lineage (excepted African or Siberian ones, I guess, but these are minor) arrived to Europe c. 32 Ka BP at the latest. That or it evolved already in Europe (but that would imply a major backflow to West Asia and Central Asia that is hard to explain). So IMO H is much older than your typical googled-up figure of 20-22 Ka BP.

      And it should be that way in order to explain the apparent H17'27 in Sunghir (25 Ka BP). H is at least of Gravettian chronology.

      My own work on mtDNA molecular-clock-o-logy, suggests that huge star-like haplogroups like H or M (those are the only really big ones, with more than 40 basal branches each) imply a large sudden expansion, resulting in a stable large population what has the effect of dramatically slowing down the effective mutation downstream because the already fixated lineages rarely allow a mutant to survive, just because of drift: try it yourself, anything of the rank of Ne>10 and of course Ne>100 tends to keep the genetic pool very stable (for rare mutations as is the case of mtDNA, of course there's always a chance but it's much larger in smallest populations of the kind Ne<10). I once called this the "cannibal mum" effect because fixated ancestral states tend to effectively eliminate novel mutant ones systematically.

      Said that, it's pointless to try to estimate age based on the downstream equidistance between two or more branches, as it's standardly done, much less when branches are all but of the same length (within R, U has hyper-long sub-branches, and H hyper-short ones), but we have to calibrate from top to bottom, for example equating age(L3)=125Ka and age(M)=100 Ka and extrapolating from there (it fits pretty well, oddly enough).

      For example I did something similar here: http://forwhattheywereweare.blogspot.com/2013/06/synthesis-of-early-colonization-of-asia.html

      Or earlier I recalibrated Oppenheimer's scholastic estimates to make them more realistic re. the archaeology and the Pan-Homo split: http://forwhattheywereweare.blogspot.com/2013/05/oppenheimer-2012-scholastic-oroborus-of.html

      I have other takes on this matter but find difficult to find stuff in a blog already five years old, try "molecular clock" label.

      "They need to get higher resolution than just rCRS from HVRs"

      They are using the classic method of RFLP confirmation. It doesn't seem like the EHU-UPV has the budget to do "modern stuff" like control region sequencing, as the Germans do. However it was reported in 2013 that Paabo had also sequenced this same specimen, and yet he has not published anything on the matter for unknown reasons, see: http://www.unm.edu/features/2013/straus-has-good-field-season-at-el-mir%C3%B3n-cave-in-spain.html

      Some think he's being partisan, just like Jean Manco.

  2. Maju
    A couple of your questions

    1) why is there voluminous data of mtDNA coming from Iberia, but no Y DNA data ? ( is it a lab issue ?)

    2) what would you speculate to be in the LGM pool of Y Hgs harviured in LGM iberia ?

  3. 1. I suspect it's a problem of relative backwardness (and most of it is coming from the Basque Country and more rarely Catalonia, the most advanced countries in the region) and also budget (austerianism is taking a massive toll) but it's just a supposition (I don't mingle often with academics, I don't know what's going on behind the doors). Only the East Germans (Max Plank Institute) and occasionally the Swedes do seem to have budgets (and maybe the academic culture) to make full sequencing of aDNA.

    2. I can't say. I still expect to find R1b (Euro subclade) in Paleolithic, mostly because it does not show up in Neolithic samples (and I do not believe in "molecular clock" dogmas but rather in geo-structure of haplogroups, which points to a western origin of these lineages) but only time and research will tell. I'd also imagine some E1b-M81 in West Iberia, if it is, as I suspect, a backflow from the Oranian genesis.

  4. I agree that this is a paradigm shifting important find, and your partition of the data is likewise worthy of note. 19,000 years ago is right in the thick of the LGM, not merely an Epipaleolithic pilot wave migration from the Near East.

  5. I do remember that a mtDNA study done on Taforalt remains did found that 19% of the samples belonged to the haplogroup H.

    1. It was a power point presentation by Kéfi et al. (see the references in the update HERE, but I find difficult to find it online today anymore). The results (HVS-I) were:

      · H or U (CRS, most likely all H): 11 (47.8%)
      · H (not CRS): 4 (17.4%)
      · JT: 3 (13%) (this would be JT*(xJ,xT) apparently, also found in nearby Andalusia and allegedly still found in the Arif today)
      · U6: 2 (8.7%)
      · V: 2 (8.7%)
      · L3, M or N: 1 (4.3%)


Please, be reasonably respectful when making comments. I do not tolerate in particular sexism, racism nor homophobia. Personal attacks, manipulation and trolling are also very much unwelcome here.The author reserves the right to delete any abusive comment.

Preliminary comment moderation is... ON (your comment may take some time, maybe days or weeks to appear).