The peopling of Europe has happened in successive waves of migrations from East to West. Homo Sapiens left Africa to settle in the Middle East, then advanced into Europe around 40,000 years ago (probably intermingling with the indigenous Neanderthals). These early Paleolithic Europeans, known as Cro-Magnons, have left their genes in the modern population. Their Y-chromosomal DNA was almost certainly haplogroup I (and perhaps also to the older haplogroup F).
The advent of agriculture in the Fertile Crescent and the domestication of cows, goats, and pigs in the South Caucasus and North Mesopotamia region prompted a new sedentary lifestyle. These early farmers expanded east towards modern Iran and Afghanistan, and west towards Europe. Based on the current knowledge, notably the recent ancient DNA tests from Neolithic France and Germany, it seems that at least two distinct groups of people left the Middle East for Europe.
The first wave of agriculturalists most likely originated in the Caucasus region and made its way through Anatolia, northern Greece, then expanded north along to Danube basin, and south along the coasts of Italy, Sardinia, North Africa, southern France and Iberia. These people belonged almost exclusively to haplogroup G2a.
The second wave might have happened in the late Neolithic/Chalcolithic or in the Bronze Age. Based on the haplogroups it brought to Europe, J2, J1, E1b1b and T, chances are that this group came from the Levant and/or Mesopotamia. Perhaps two groups merged in Anatolia, Greece and Italy giving rise to the ancient classical civilizations of Greece, Etruria and Rome.
Modern Europe is dominated by two paternal haplogroups, R1a and R1b, which are associated with the Bronze Age conquest of Europe and South Asia by horse-riding pastoralists from the Pontic-Caspian steppes. They are thought to be the original speakers of modern Indo-European languages, from English, French and Greek to Russian, Iranian and Hindi.
The haplogroup R bias
As I have explained here, the exceptionally fast replacement of all paternal lineages by R1a and R1b lineages in Europe, Central Asia and South Asia, is probably due to several factors (heavy losses in battles among males in conquered populations, polygamy among Indo-European rulers, higher genetic predisposition of haplogroup R to conceive boys, etc.).
Unfortunately, this unfair bias in favour of haplogroup R makes it much more difficult to estimate the percentage of native Paleolithic vs Middle-Eastern Neolithic blood among modern Europeans. What is sure is that haplogroup R1a and R1b both have disproportionately high percentages compared to their real genetic contributions on other chromosomes than Y. So much is obvious from mtDNA studies, which do not show an obvious surplus of Indo-European genes, and indeed indicate a strong continuity from the Paleolithic and Neolithic.
This is why I have worked on re-calculating the percentages of Y-DNA after removing haplogroups R1a and R1b. The results are startlingly in line with autosomal studies separating northern and southern Europeans.
I have added the new percentages of I1, I2a and I2b together to obtain a percentage of Paleolithic admixture. Likewise, I have made the total for the Middle Eastern admixture (G2a, J1, J2, E1b1b and T).
Here are some selected results classified from the most Palaeolithic to the most Middle Eastern :
Over 2/3 Paleolithic European
- Iceland : 94.5% Paleolithic vs 0% Middle Eastern (+ 5.5% Mongoloid)
- Norway : 84% Paleolithic vs 7% Middle Eastern (+ 9% Mongoloid)
- Denmark : 83.5% Paleolithic vs 15.5% Middle Eastern (+ 3.5% Mongoloid)
- Croatia : 81% Paleolithic vs 19% Middle Eastern
- Ireland 74.5% Paleolithic vs 25% Middle Eastern (+ 0.5% Mongoloid)
- Scotland 74% Paleolithic vs 23.5% Middle Eastern (+ 2.5% Mongoloid)
- England : 73.5% Paleolithic vs 26% Middle Eastern (+ 1.5% Mongoloid)
- Sweden : 69.5% Paleolithic vs 10% Middle Eastern (+ 19.5% Mongoloid)
(- Brittany : 69% Paleolithic vs 31% Middle Eastern)
From 1/2 to 2/3 Paleolithic European
(- Basque country : 64% Paleolithic vs 36% Middle Eastern)
- Netherlands : 63% Paleolithic vs 34.5% Middle Eastern (+ 2% Mongoloid)
- Wales : 59% Paleolithic vs 41% Middle Eastern
- Poland : 59% Paleolithic vs 41% Middle Eastern
- Belgium : 57.5% Paleolithic vs 42% Middle Eastern (+ 0.5% Mongoloid)
- Germany : 56% Paleolithic vs 41% Middle Eastern (+ 3.5% Mongoloid)
- Czech Republic : 54.5% Paleolithic vs 41% Middle Eastern (+ 4.5% Mongoloid)
From 1/3 to 1/2 Paleolithic European
(- Sardinia : 47.5% Paleolithic vs 52.5% Middle Eastern)
- France : 45% Paleolithic vs 54.5% Middle Eastern
- Switzerland : 40.5% Paleolithic vs 56% Middle Eastern
- Austria : 40% Paleolithic vs 58% Middle Eastern (+ 2% Mongoloid)
Less than 1/3 Paleolithic European
(- North Italy : 26.5% Paleolithic vs 73.5% Middle Eastern)
- Spain : 24.5% Paleolithic vs 75.5% Middle Eastern
(- Alsace : 23.5% Paleolithic vs 76.5% Middle Eastern)
- Greece : 20% Paleolithic vs 80% Middle Eastern
(- Galicia : 19% Paleolithic vs 81% Middle Eastern)
- Portugal : 15.5% Paleolithic vs 83.5% Middle Eastern
- Italy : 13% Paleolithic vs 87% Middle Eastern
(- Crete : 12.5% Paleolithic vs 87.5% Middle Eastern)
(- Auvergne : 10.5% Paleolithic vs 89.5% Middle Eastern)
- Cyprus : 10% Paleolithic vs 90% Middle Eastern
(- Sicily: 7.5% Paleolithic vs 92.5% Middle Eastern)
Mongoloid haplogroups here are N and Q, both of Siberian origin and present mostly in Germanic or Slavic populations. The Q is Mediterranean populations being of Near-Eastern origin (Phoenician or Jewish ?), it is counted as Middle Eastern.
This way of calculating doesn't work well for Russia or Ukraine because the original population was R1a and R1b. Removing them gives very odd results (Russia = 32% Paleolithic + 25% Middle Eastern + 43% Mongoloid ; Ukraine = 32.5% Paleolithic + 52% Middle Eastern + 15.5% Mongoloid). If we keep haplogroup R it fits much better with the look of white Russians (67.5% Paleolithic + 8% Middle Eastern + 24% Mongoloid) and Ukrainians (69% Paleolithic + 24% Middle Eastern + 7% Mongoloid). This is another reason to believe that the Bronze-Age steppe people were indeed R1a and R1b. Likewise it makes more sense to keep R1a in Finland and Baltic countries to avoid excessive Mongoloid admixture that does not reflect the reality.
Note that all Scandinavia and the British Isles (except Wales) are over 2/3 Paleolithic European. All these countries are also overwhelmingly long-headed (dolichocephalic) and long-faced, just like Paleolithic skulls from northern Europe. If the Irish, for instance, who are 80% R1b, really had 80% of Pontic steppe Indo-European DNA, they would be much more brachycephalic and look very different. The R1b Y-chromosome spread in a very imbalanced way, quickly replacing most other lineages. But the core genetic pool of northern Europeans is to be found among the pre-R1b population, mostly among members of haplogroup I.
This also explains why the Frisians, from the northern Netherlands, who are over 70% R1b, don't look so different from the Swedes, who are only 20% R1b. Once R1b and R1a are removed, the core is similar : mostly I1 with some I2b, G2a, J2 and E1b1b.
The Welsh and Poles appear similar, both at 59% Paleolithic and 41% Middle Eastern. It's obvious that they look quite different, and if one looks at the haplogroups besides R1a and R1b, they are indeed very different. Wales has mostly I1 (42%) and G2a (17%), while Polish substrata is composed primarily of I2a2 (33.5%) and E1b1b (18.5%).