Ancient Y-DNA discussion: no J2a in pre-Bronze Europe?

UPDATE 10 November 2014: The publication of DNA data from Neolithic to Iron Age Hungary did make a revision of this article necessary, as the sample BR2 (ca. 3,200 years old) was found to be positive for J2a-M67-CTS3261 (Kandell T.). His autosomal analysis lets suspect that he descends mainly from Western ancestors with no near connections to Anatolia Caucasus or the Near East,  so if we use 8 generations as limit to detect substantial admixture and add another 2 generations because of the probable position in Continental Europe, we can assume (+ ca. 10 x 28y) that ca. 3,500 years ago (1,500 BC) J2a was present in Europe. This new discovery makes the previous provocative title “no J2a in pre-antiquity Europe” obsolete and some thoughts needed revision. Mainly conclusion 6. was added and 7 revised; also the hypothesis 7. was revised. Still this finding is compatible with (late?) Bronze Age migrations of J2a to Europe. Further very old and high resolution aY data (like Ust’-Ishim 45k and Kostenki-14 37k) will help to optimize the age calibrations of the Y-Haplogroups and together with enough ancient and modern HQ Y-sequences to further revise this article.

Update 13 December 2014: Image “Neolithic and later Y-Hg migrations to Europe”
Update 15 February 2015: Haak et al. preprint aDNA, Y-DNA analysis Krefter. R1 migration needs revision. YFull age estimates recalibration.
Update 7 March 2015: New NGS-Y-samples of Estonian Biocentre for J2 include GS000035477-DID Georgian which allows further phylogeny definition sub of CTS3261. The Haplogroup structure of BR2 by single SNP definition, the most terminal SNPs still shared with HG01302 Puerto Rican: M67>Z7671>CTS3261>Y11202>Z30688>Z30685
Update 10 October 2015: inclusion of Neolithic Hungarian J2 samples: 5030-4840 BC (ALE14) and 5000-4300 BC? (FEB3) in Szecsenyi-Nagy 2015.

This speculative discussion is focused on the Y-Haplogroup J2a, but the ancient DNA in combination with known history and the modern distribution suggests to include other Haplogroups (Hg) for a better analysis. The post is intended to provide a basis to discuss contrasting hypotheses. This post was motivated by reading the latest (preprint) paper on ancient (Central) European DNA by Szécsényi-Nagy et al.:

Surprisingly, Y chromosome haplogroups, such as E1b1b1 (M35), E1b1b1a1 (M78), E1b1b1b2a (M123), J2 (M172), J1 (M267), and R1b1a2 (M269), which were claimed to be associated with the Neolithic expansion, have not been found so far in the 6th millennium BC of the Carpathian Basin and Central Europe.

Ancient European Y-diversity evidence

Ancient European Y, GMap by RM Stevens

Looking at other existing ancient Y chromosome studies (collections by Dienekes, Eupedia and a GMap) the following conclusions can be made:

1. I2 is most probably the “oldest surviving major European Y-Haplogroups” with presence at least since Mesolithic (hunter gatherers) and today are important in Northern and Balkanic Europe. C1a-V20 should be another hunter gatherer lineage which nowadays almost disappeared. The origin of I1 could be CE Europe.
   
2. R1b L278 is in 5600 BC Samara (Volga E Europe), 5120 BC M415 NE Spain, 2890 BC Z2105 Samara Volga, 2260 BC P312 in Harz CN Germany and P312 with U106 dominates Western Europe
3. R1a was found in 5250 BC Karelia (NE Europe, M459) and M417 2,410 BC in Saalekreis CN Germany and till today is dominant in Eastern Europe.
4. G2a is well confirmed as very present Neolithic (5125 BC in Harz Germany) in Europe and somewhere along the history was marginalized in Europe (Sardinia, Caucasus)
5. T1a-M70 is in 5135 BC Saxony-Anhalt CN Germany but likely did not expand
6. J2 is in 4935 BC Hungary (most likely subgroup J2b2a1-L283 which has high diversity and frequency in modern Balkans).
7. E1b-V13 has one presence in Neolithic Spain (5,000 BC); together with the strong presence in the modern Balkans this seems to be the second Neolithic Hg
8. J2a-M67-Z30685 shows up in 1,200 BC remains from nowadays Hungary – analysis of Ted Kendall
9. Prior to the end of the European Migration Period (~800 CE) so far no further modern European Haplogroups like E (xV13), J1, J2b, L, N, Q were found in Europe.
10. European J2a sub-lineages predominantly have low diversity and J2a haplogroups suspected to be not older then 6,000 years almost all have a distribution going from Middle East both to Europe, West- and South Asia. The age of the oldest European J2a-sublineage – suspected by me as J2a-Z6056 (xM68) – needs to be estimated by NGS-analysis.
11. N1-LLY22g was found in Medieval Hungarian samples (1,000 CE) together with the Tat C allele associated to Finno-Ugric populations, where this Hg is still dominant.

Hypotheses Y-immigration from Late Neolithic to antiquity

Neolithic and later Y-Hg migrations to Europe, CC-BY ChrisR

Because the focus is on J2a based on the evidence the hypotheses are build from West Eurasian rise of animal domestication (ca. 8,000 BC) to the introduction of metallurgy to Europe (3,200 BC) to the begin of the Migration Period (376 CE).

1. The absence of J2a in Neolithic and early Bronze age European samples in my opinion makes a presence of this Hg in the Eastern Mediterranean area (including the Levant) and probably also in the whole Fertile Crescent before the animal domestication (and maybe also some millennium later) unlikely. Rather G, E and some maybe seldom or extinct Hg should have dominated the area. And this is the thought mostly diverging from known hypotheses of the origin and Neolithic presence of J2 (and also J1). This thought is valid probably for almost all other Hg’s not found as well: E (xV13), J1, L, Q, T making their expansion to Europe in the Bronze age or later likely.
2. The most discussed and unclear migration and expansion in Europe concerns R1b and also R1a. Lets follow the majority opinion that R1b-M343 did originate in Central Asia, migrated to the Black Sea area (R1b-M269), did came to Europe over the Lower Danube area (R1b-L23) and expanded in Central Europe (R1b-L11) to Western Europe. R1a is believed to had its origin in Northern Central Asia (R1a-M420) and expanded to Europe over the steppe North of Caspian and Black Sea (R1a-M417) and there expanded to the West and East.
3. If R1b and R1a could expand so late and extraordinarily in Europe, there is no reason to believe that similar events did not also happen with other Hg’s in the Old World in the same relatively short time.
4. If the early R1b diversity, frequency and presumed migrations are combined, IMO a connection with J2a could be made, if J2a was really absent from Fertile Crescent before the animal domestication. Lets assume:
   a) R1b came from Central Asia to the Iranian Plateau and somewhere there met existing J2 sub-lineages. A later additional or alternative explanation of migrations to the West could be the invasion of R1a-Z93 lineages (Aryans?).
   b) Later a migration to the Northern Zagros area occurred; possibly J2 sub-lineages were pushed by R1b (or/and R1a).
   c) Later maybe sometime before the Bronze Age (3,300 BC) most of J2a went to expand in Mesopotamia, (part of) J2b went to the Southern Balkans and R1b to the Northern Balkans (possibly over the Northern Black Sea area).
5. I believe there is reason to suspect R1b-L23 (approx.) played a major role in the distribution of Bronze knowledge to the Balkans and Central Europe combined with technological advantage and dominance over previous Europeans living in wood-free Neolithic settlements. Up to this time there was plenty of space and opportunities for newcomers in Europe. I suspect also J2b and J2a had early access to Metallurgy technology to their military advantage over neighbors and possible rapid expansion in the last 5,000 years.
6. If all the other Hypotheses are correct G2a possibly with some other G and E subgroups are the oldest surviving Levantine Haplogroups.
7. The main question arises combining all the Hypotheses: could J2a not reach the Eastern Mediterranean area before the early Bronze Age, but before the Iron Age (1,200 BC)? The less speculative question is: was J2a already in Europe in Late Neolithic?

Origin of J2a-L26/Page55/S57

Speculative origin J2a-L26, CC-BY ChrisR

As J2 (approx. 30,000 years) and J2a (approx. 21,000 years) are very old and it is difficult to track down the origin for them, here the focus is on a younger lineage including the majority of all J2a lineages: J2a-L26 (approx. 17,500 years).

1. The expansion of this lineage could roughly coincide with the ice decay after the end of the LGM. During the LGM an interesting proposed environmental refugia is the Arabo-Persian Gulf Oasis in the now flooded Persian Gulf.
2. The J2a-L26 sub-lineages probably were already pretty diversified (M67 approx. 13,500 years, L24 approx. 14,000 years, etc.) when agriculture and animal domestication was invented and used widely (14,000-10,000 years ago).
3. Combined with the Hypothesis that J2a was not present in Neolithic Levant and the Y-STR-diversity analysis for J2a-L24/M530 and J2a-M67(xM92) of Grugni et al 2012 we can deduce that J2a at least up to early Neolithic had its home east of the Zagros mountains.
4. The J2a-PF5116 groups (incl. M67) probably went westwards earlier and most likely in the Northern Zagros region (Metallurgy time). Cross migrations also trough South Zagros possibly contributed to the complex J2a-L26 phylogeny with no clear ancestral distribution.
5. Summary: it should be save to say J2a-L26 was somewhere in or near the Iranian plateau. His Y-ancestors probably came from the South (Gulf Oasis) or another refugium nearby.

Further Evidence needed to validate the Hypotheses

1. High-res public NG-Sequencing of the most diverse sub-clades of all the discussed Hg’s combined with refined age estimates based in the Y-diversity
2. Many additional ancient Y-DNA from the Neolithic to Antiquity mainly in the Mediterranean and Middle East area sequenced with better then low-coverage.
3. The oldest European (pre) I remains with good age dating, sequenced with maximum possible coverage to have a good Y-age calibration point.

Notification of any further evidence missed here (points 1-3) is appreciated.
I would like to thank Ted Kandell for his postings and discussion with me in the FB Haplogroup G group.