search this blog

Saturday, March 4, 2017

Something unexpected from Mesolithic Sardinia

More surprises from pre-Neolithic Southern Europe courtesy of a new paper at Scientific Reports. Emphasis is mine:

Abstract: Little is known about the genetic prehistory of Sardinia because of the scarcity of pre-Neolithic human remains. From a genetic perspective, modern Sardinians are known as genetic outliers in Europe, showing unusually high levels of internal diversity and a close relationship to early European Neolithic farmers. However, how far this peculiar genetic structure extends and how it originated was to date impossible to test. Here we present the first and oldest complete mitochondrial sequences from Sardinia, dated back to 10,000 yBP. These two individuals, while confirming a Mesolithic occupation of the island, belong to rare mtDNA lineages, which have never been found before in Mesolithic samples and that are currently present at low frequencies not only in Sardinia, but in the whole Europe. Preliminary Approximate Bayesian Computations, restricted by biased reference samples for Mesolithic Sardinia (the two typed samples) and Neolithic Europe (limited to central and north European sequences), suggest that the first inhabitants of the island have had a small or negligible contribution to the present-day Sardinian population, which mainly derives its genetic diversity from continental migration into the island by Neolithic times.


The CAR-H8 sample belongs to haplogroup I3, hence representing, to the best of our knowledge, the first pre-Neolithic sample carrying the haplogroup I. Studies based on complete mitogenomes have previously reported haplogroup I in ancient samples from Iran (individual I674, haplogroup I1c) and Levant (individual I1679, haplogroup I), dated to 5,105 ± 35 yBP and 8,850–8,750 yBP, respectively [39]. It was also found in two late Neolithic individuals from Germany, both belonging to haplogroup I3a and dated to around 4,000 yBP [50] but not in previous periods in Europe. Nowadays, this haplogroup is uncommon; its frequency is about 2% in modern Sardinians, 3% across Europe, and raises at maximum 6% in Northern European countries [51]. This is the first time that haplogroup I is found in a Mesolithic individual in Europe and the fact that we recovered this haplogroup in a sample of only two sequences may mean that it was present at higher frequencies in pre-Neolithic Sardinians or, in general, in the population that first settled in the island. The other sample (CAR-H7) belongs to the haplogroup J2b1. The haplogroup J has already been found in late hunter-gatherer European populations, with a frequency of about 4% 32]. The current frequency of the haplogroup J is higher than that of the haplogroup I, variable in Europe from 1.7% (Caucasus) to 15% (Wales), and representing the 13% of the total modern Sardinians mitochondrial sequences.

Modi, A. et al. Complete mitochondrial sequences from Mesolithic Sardinia. Sci. Rep. 7, 42869; doi: 10.1038/srep42869 (2017).

See also...

Mitogenome diversity in Sardinians


Roy King said...

Thanks for the posting! Apparently Mesolithic Sardinia and, previously, Mesolithic Greece have very different mtDNA form the rest of more northern Mesolithic Europe. The Sardinia site is close to the sea, and perhaps engaged in fishing-foraging as a dominant food resource activity. I'd put money on Mesolithic Sardinia more resembling Anatolian/Near Eastern Neolithic individuals than say other European Neolithic sites in autosomal DNA. Mysteries abound!

Rob said...

@ Roy

Do you know if this team will undertake full genome analysis also ?

Roy King said...


I don't know but it would be critical for them to do that.

Annie Mouse said...

I was interested in the Welsh connection particularly as there has been talk of a bronze age Balkans ->Sardinia -> Wales copper connection.

However I think the mitochondrial J in Wales is mostly J1 (from Familytree DNA project).

So far as I know there are no Paleolithic or mesolithic Js other than this one so far.

In the neolithic there are a few unspecified Js but the ones that have been further analysed are all J1. First neolithic J2 is 4k years later from nearby France.

Anyway the theory that J2b1 came over to Europe in the neolithic from the Near East along with farmers is very firmly killed stone dead. The neolithic in Sardinia is not supposed to arrive for another 2k years (Cardial 6000BCE).

And also we have not found ANY J2b in the neolithic near east. Although the large percentage in the modern Kalash and possibly Mansi (one paper said 10%, 1 said 0%) is certainly food for thought.

I think water-borne mesolithic mediterraneans sloshed around the waterways including east up to Volga in time to participate in Bronze age steppe flows. Maybe.

Matt said...

Anatolia Neolithic models on PCA (to some extent) as like 20:80 WHG:Levant Neolithic. Then Early Neolithic Europe models like 28:72 WHG:Levant Neolithic. Then there's about another 10-15 WHG to get to something like the MN samples we have, so 38:62 or 43:57.

So that implies about means 8% admixture in Early Neolithic and then 10-15% admixture more for MN, assuming they're mixing with pure WHG.

But if they're mixing with populations that were, say, similar to 50:50 WHG:Levant Neolithic who already existed in South Europe, you'd have to double the admixture. That's less plausible for MN in West and Central Europe (Iberia and Germany) where we know Loschbour, etc existed and La Brana in Northern Spain, but maybe happened if the EN picked up their admixture through Greece and Italy on the way to Iberia and Germany, and those populations were somewhere on a WHG-Anatolia Neolithic cline already.

Gioiello said...

Of course whar the authors say about the discontinuity between these hpts of 11000 years ago and later European mitochondrions should be taken with many grains of salt:
1) It is true that I3* has no known descendant so far, lacking also some mutations that we don't know if due to the test or because back mutations (T204C and A13780G at N1a level, T199C at N1aa1'2 level, G16129A at I level, G207A at I2'3 level), but anyway it demonstrates that hg. I was in Sardinia before any other I3 found so far.
2) J2a1 is deeply rooted in Europe also to-day.

Gioiello said...

Also the J2b1 sample, with many private mutations (195C, 3654T, 6053T, 6959T, 8953G, 9071T, 10957G, 16193C!), not only finds three of these mutations in two samples of to-day, that may have been in homoplasy with it:

8. JQ797953(Italy) Pala J2b1 22-MAY-2012
A73G C150T T152C T195C A263G C295T 309.1C 315.1C T489C A750G A1438G A2706G T4216C A4769G C5633T C7028T C7476T A8860G A9852G G10172A A10398G A11251G G11719A A12612G G13708A C14766T G15257A A15326G A15442G C15452A G15812A C16069T T16126C C16167T C16193T T16311C
20. KP317041(UAR) Fernandes J2b1 15-MAR-2015
A73G C150T T152C A263G C295T 315.1C T489C C522- A523- A750G A1438G A2706G T4216C A4769G C5633T C7028T C7476T A8860G C9593T G10172A A10398G A11251G G11719A A12612G G13708A A14587G C14766T A15326G C15452A G15812A C16069T T16093C! T16126C

but makes us doubt that the ages of Behar 2012b [J2b1 9281.8 +/- 3107.5] may be right, if already 11000 years ago there was a sample so differentiated from the presumed age.