PDA

View Full Version : Did haplogroup I only develop after the Ice Age?



epp
01-23-2016, 12:50 PM
Published data indicate that haplogroup I's most recent common ancestor was located in North West Europe, and that its population rapidly grew and diversified fairly shortly afterwards in this same location. Reproductive success like this would usually be expected to occur in promising locations, so this does not seem to be intuitively consistent with the theory that I initially expanded during the Ice Age, when much of its habitat was extremely cold and covered by glaciers. Nor does it seems consistent with these people originating pre-Ice Age and basing themselves in perma-frozen North West Europe for thousands of years until the ice sheets melted. Is it more likely that I's most recent common ancestor arrived and thrived in Europe soon after the Ice Age, just as the North West of the continent was becoming highly habitable and when it was largely empty, with little competition for resources?

Jean M
01-23-2016, 01:29 PM
Is it more likely that I's most recent common ancestor arrived and thrived in Europe soon after the Ice Age..?

No. Europe was not entirely deserted during the Last Glacial Maximum. There were pockets of people living in southern refuges. It was mainly these people who were able to expand to re-populate the more northerly parts of Europe. I don't say that absolutely no-one arrived from outside Europe as the climate improved. A few people probably came from the Levant along the coast to Greece. Others we know arrived from the Asian steppe. But a man of Y-DNA haplogroup I2a was buried in a cave in Switzerland in 13,560–13,770 cal. BP. http://www.ancestraljourneys.org/palaeolithicdna.shtml

So it would appear that Y-DNA haplogroup I had arrived in Europe in or by the Gravettian period. The latest estimate from Y-Full of its age is 42,400 ybp (formation), and TMRCA 27,300 ybp. That is not old enough to have arrived with the very first Homo sapiens in Europe 45,000 years ago, but it could have arisen from an IJ father in Europe before the Ice Age made northern Europe uninhabitable around 20,000 years ago.

Megalophias
01-23-2016, 04:41 PM
Published data indicate that haplogroup I's most recent common ancestor was located in North West Europe, and that its population rapidly grew and diversified fairly shortly afterwards in this same location.

Which study was this? I don't actually recall Northwest Europe ever being suggested as origin of Haplogroup I.

epp
01-23-2016, 04:41 PM
No. Europe was not entirely deserted during the Last Glacial Maximum. There were pockets of people living in southern refuges. It was mainly these people who were able to expand to re-populate the more northerly parts of Europe. I don't say that absolutely no-one arrived from outside Europe as the climate improved. A few people probably came from the Levant along the coast to Greece. Others we know arrived from the Asian steppe. But a man of Y-DNA haplogroup I2a was buried in a cave in Switzerland in 13,560–13,770 cal. BP. http://www.ancestraljourneys.org/palaeolithicdna.shtml

So it would appear that Y-DNA haplogroup I had arrived in Europe in or by the Gravettian period. The latest estimate from Y-Full of its age is 42,400 ybp (formation), and TMRCA 27,300 ybp. That is not old enough to have arrived with the very first Homo sapiens in Europe 45,000 years ago, but it could have arisen from an IJ father in Europe before the Ice Age made northern Europe uninhabitable around 20,000 years ago.

This is an interesting study. Yes, I agree that Europe may not have been entirely deserted during the Last Glacial Maximum, but I have doubts whether those present would have been haplogroup I:
1. All haplogroup I currently in possible Southern European refuge locations is of relatively recent origin. For this to be compatible with evidence from current samples, you would have to accept that haplogroup I expanded and diverged in Northern Europe, migrated en masse to Southern Europe during the Ice Age, then over ten thousand years later, all migrated back to Northern Europe again en masse, leaving no remaining trace of themselves in southern European populations. My view is that this is implausible.
2. The man buried in the Swiss cave was not proven I2a - the researcher merely described I2a as his "best guess".
3. The cave in question is not in a Southern refuge, but North of the Alps, near the source of the Northern River Rhine, where current DNA samples suggest some of the earliest haplogroup I development occurred.
4. The body in the study was carbon-dated at 9,850 BC, which is not before or during the Ice Age, but after it had finished.

In my view, the evidence seems most consistent with I arriving in North Western Europe from the East after the Ice Age had come to an end.

Jean M
01-23-2016, 05:10 PM
For this to be compatible with evidence from current samples, you would have to accept that haplogroup I expanded and diverged in Northern Europe.

What makes you think that haplogroup I expanded and diverged in Northern Europe? Modern DNA cannot tell us that.

epp
01-23-2016, 05:24 PM
Which study was this? I don't actually recall Northwest Europe ever being suggested as origin of Haplogroup I.
Just to clarify, I'm not suggesting that haplogroup I originated in North West Europe, rather that North West Europe is the likely location for the most recent common ancestor of its surviving lineages.
It would be quite surprising if no one had suggested North West Europe before, as the data freely available on the internet seems to indicate it fairly clearly - the most diverse I1 samples are found in this region, and the most diverse I2 samples are also found in this region.

epp
01-23-2016, 05:34 PM
What makes you think that haplogroup I expanded and diverged in Northern Europe? Modern DNA cannot tell us that.
In the absence of other reliable information, I am assuming that haplogroup I is most likely to have diverged where it is currently at its most diverse.

I'd be grateful if you would let me know whether you think I have correctly analysed the the study of the body in the Swiss cave.

Jean M
01-23-2016, 06:02 PM
In the absence of other reliable information, I am assuming that haplogroup I is most likely to have diverged where it is currently at its most diverse.

Geneticists used to claim that it was most diverse today in southeastern Europe, but in any case this methodology is problematic.

Yes geneticists developed the theory that that where they find the greatest genetic variance of a haplogroup is likely to be its point of origin, since the longer a lineage has been in a place, the longer it has had to accumulate variations. This seems satisfactory in broad continental outline, but later movements will have swirled the mix in ways that could mislead us if we expect variance today to exactly match that in prehistory. A present-day population could have acquired diversity from diverse waves of immigrants. We find a high variance within European haplogroups in the United States of America, but we know full well that these haplogroups do not have their origin there. So variance is most convincing when supported by other kinds of evidence.

Take archaeology. There was no-one living in Northern Europe when it was covered by ice sheets. So everyone living today in Scandinavia and the British Isles descends from people who arrived there in the Mesolithic at the earliest. (We now know that later migrations had a huge impact on the gene pool.) The pattern of diversity in Y-DNA I that we see in living people today cannot reflect an origin point in northern Europe before the LGM.

I2 was present in Scandinavia in the Mesolithic, but not the I1 that is more common there today.
I1 and I2 were both present in Neolithic Hungary.
The expansion of I1 in Scandinavia is compatible with an arrival in the Copper/Bronze Age.
The expansion of I2a1b2a in the Balkans is compatible with the arrival of Slavs in the early medieval period. There they might encounter the occasional I2a that descended from Neolithic farmers there, so creating diversity.

Jean M
01-23-2016, 06:08 PM
Here is Ken Nordtvedt's conjectural map of the spread of branches of haplogroup I, as created in 2011. I apologise in advance to Ken for using something of his that is several years out of date, but I have nothing more recent.

7418

Jean M
01-23-2016, 06:24 PM
I used to have an online page on haplogroup I, but did not have time to maintain it in the face of the deluge of data in recent years, so I took it down. However here's the gist of relevant bits:

Ancient DNA is gradually resolving the mysteries of Y-DNA Haplogroup I (L41). Its modern distribution was puzzling. On the one hand it seemed ancient in Europe. It rarely appears outside the boundaries of Europe and European colonies. So it was not a good candidate for arrival with farmers from the Near East. Nor did it seem the prime candidate for spread with the Indo-Europeans, since they travelled both west into Europe and east into the Indian Subcontinent. So the natural conclusion was that haplogroup I had been stalking around Europe since the Stone Age. On the other hand the pattern of I subclades in present-day European men look relatively recent. We see regional bunching, typical of relatively recent arrivals. What are we to make of these contradictions? The haplogroup may date deep into the distant European past, but it seems that most of the hunters and foragers who carried it have no direct descendants in the male line today.

Population patterns

In a hunter-gatherer economy, the population is usually maintained at replacement level, where that community remains within a particular territory. Women space births by weaning late. Population levels need to be low, as each hunting band needs to roam a large territory. The human population dropped dramatically world-wide during the last glacial maximum. Within Europe it fell to the point where we would today classify it as an endangered species.

Then the population expanded during the Mesolithic as people gradually reclaimed the territory that had been lost to the climate downturn. Once it had expanded enough to fill the territory at the low hunter-gatherer level, we would expect it to be stable until farming made higher levels possible. Haplogroup I1 does not show any star-burst of subclades at that time, so we can presume that people carrying it were in no hurry to take up farming. However we do see bursts of new lineages in I2 at c. 8,000 years ago = 6000 BC, as farming reached the Balkans. It appears that some I2 men were willing and able to adopt agriculture.

So my inclination is to look for the ancestors of today's I-men in successful hunter-gatherer cultures, which had a good chance of leaving descendants. In the days when all mankind lived by hunting and gathering, all could be considered equally successful if they managed to survive in competition with other predators. This might include other human hunting bands, but fellow humans were not initially the main competition. Man had to be clever enough to out-do lions and bears and not end up at the wrong end of the food chain. Once farming entered the picture, hunters were in direct competition with people who could outbreed them and inexorably take over the territory. Successful hunting cultures at that point were few and far between. Characteristically they occupied a highly fruitful hunting or fishing niche, that could scarcely be bettered at that stage by turning it over to farming. People in such a niche could hold off any incoming farmers who thought otherwise, and choose to adopt whatever seemed useful from farming neighbours at their own pace.

Jean M
01-23-2016, 06:55 PM
The man buried in the Swiss cave was not proven I2a - the researcher merely described I2a as his "best guess"..

Not the researcher. The haplogroups were determined by Yfitter. The “maximum likelihood haplogroup” (12a in this case) is described by Yfitter as being the best guess haplogroup while the “confidence haplogroup” (I2 in this case) is described as the conservative guess haplogroup. This is Yfitter: http://arxiv.org/abs/1407.7988

I doubt whether I2 would be seriously wrong.

Megalophias
01-23-2016, 07:39 PM
Just to clarify, I'm not suggesting that haplogroup I originated in North West Europe, rather that North West Europe is the likely location for the most recent common ancestor of its surviving lineages.
Yeah, that is what I mean as well.


It would be quite surprising if no one had suggested North West Europe before, as the data freely available on the internet seems to indicate it fairly clearly - the most diverse I1 samples are found in this region, and the most diverse I2 samples are also found in this region.
Honestly I don't know much about haplogroup I diversity. I find it very difficult to compare diversity of anything in Europe, especially Northwest Europe, to other parts of the world, simply because so much of what is known is from private testing, which is massively biased geographically and lacks information about frequency and representativeness. I mostly rely on academic studies for that reason, but of course that has limitations as well.

I have never come across an I*(xI1, I2) in Northwestern Europe, but it exists in the Caucasus. And I have heard before that the greatest diversity of I1 is in Central Europe, though I wouldn't know myself. I would be interested to know where the most basal clades of I are to be found, actually.

lgmayka
01-23-2016, 09:12 PM
the most diverse I1 samples are found in this region, and the most diverse I2 samples are also found in this region.
The TMRCA of all known living I1 is only 4700 years ago (http://yfull.com/tree/I1/), according to YFull. In other words, one man living 4700 years ago patrilineally sired all the millions of I1 men living today. Because I1 expanded so rapidly at that time, both demographically and geographically, modern DNA results alone cannot possibly tell us where that one man lived.

I2's haplotree presents a very different picture (http://yfull.com/tree/I2/). One could plot the current locations of all its various branches in an attempt to guess the tracks of its demographic and geographic growth across the millennia. But even in this case, I think we need much more data--especially from the rare offshoots which often provide the best clues.

Gravetto-Danubian
01-23-2016, 09:56 PM
.

I have never come across an I*(xI1, I2) in Northwestern Europe, but it exists in the Caucasus.

I believe all the "" I*" previously found in the Caucasus is now known to be I2c- and one specific sub-branch of 3 otherwise found in Europe.

parasar
01-23-2016, 10:47 PM
Not the researcher. The haplogroups were determined by Yfitter. The “maximum likelihood haplogroup” (12a in this case) is described by Yfitter as being the best guess haplogroup while the “confidence haplogroup” (I2 in this case) is described as the conservative guess haplogroup. This is Yfitter: http://arxiv.org/abs/1407.7988

I doubt whether I2 would be seriously wrong.

Y SNP calls per Genetiker:
https://genetiker.wordpress.com/y-snp-calls-for-the-grotte-du-bichon-hunter-gatherer/
"The calls show that the Grotte du Bichon hunter-gatherer belonged to Y haplogroup pre-I2a1a2a1-L1287"

epp
01-23-2016, 11:56 PM
Geneticists used to claim that it was most diverse today in southeastern Europe, but in any case this methodology is problematic.

Yes geneticists developed the theory that that where they find the greatest genetic variance of a haplogroup is likely to be its point of origin, since the longer a lineage has been in a place, the longer it has had to accumulate variations. This seems satisfactory in broad continental outline, but later movements will have swirled the mix in ways that could mislead us if we expect variance today to exactly match that in prehistory. A present-day population could have acquired diversity from diverse waves of immigrants. We find a high variance within European haplogroups in the United States of America, but we know full well that these haplogroups do not have their origin there. So variance is most convincing when supported by other kinds of evidence.

Take archaeology. There was no-one living in Northern Europe when it was covered by ice sheets. So everyone living today in Scandinavia and the British Isles descends from people who arrived there in the Mesolithic at the earliest. (We now know that later migrations had a huge impact on the gene pool.) The pattern of diversity in Y-DNA I that we see in living people today cannot reflect an origin point in northern Europe before the LGM.

I2 was present in Scandinavia in the Mesolithic, but not the I1 that is more common there today.
I1 and I2 were both present in Neolithic Hungary.
The expansion of I1 in Scandinavia is compatible with an arrival in the Copper/Bronze Age.
The expansion of I2a1b2a in the Balkans is compatible with the arrival of Slavs in the early medieval period. There they might encounter the occasional I2a that descended from Neolithic farmers there, so creating diversity.




Yes geneticists developed the theory that that where they find the greatest genetic variance of a haplogroup is likely to be its point of origin, since the longer a lineage has been in a place, the longer it has had to accumulate variations. This seems satisfactory in broad continental outline, but later movements will have swirled the mix in ways that could mislead us if we expect variance today to exactly match that in prehistory. A present-day population could have acquired diversity from diverse waves of immigrants.



I2 was present in Scandinavia in the Mesolithic, but not the I1 that is more common there today.

From thousands of samples published on the internet, these are the estimated points of origin for the most diverse haplogroup I subclades, based on their degree of internal diversity - I* (pre- I1) - Britain/Norway; I2 (pre- I2b/I2c) - Germany; I2a2 - Britain/Low Countries; I2a1b - Britain; I2a1a (pre-I2a1a1) - Britain/NW France; I2a1a2 - Germany.

I am not saying that haplogroup I definitely originated in these areas, but merely ask - does this pattern indicate that these haplogroup I lineages are more likely to be ancestral to these areas where they are all now found, or ancestral to Southern Europe, the Caucasus, the Balkans or some other region not appearing anywhere at all in this mix?

To me, the theory that these earliest subclades each evolved somewhere else, then all relocated to North West Europe "in diverse waves of immigration" without leaving a trace in any of their points of origin, seems less plausible than the theory that they evolved approximately in the region where they are now all found.

This is simply based on DNA data, not on conjecture (such as 'conjectural maps'), nor on any uncertain premise, such as that I1 was "not present in Scandinavia in the Mesolithic" (we cannot know that).

The Swiss sample previously mentioned suggests that there was an early haplogroup I individual (whose DNA was more similar to I2a than to other surviving lineages) living North West of the Alps not too long after the Ice Age. This does not seem at all inconsistent with the premise proposed in the thread.

Gravetto-Danubian
01-24-2016, 12:06 AM
I am not saying that haplogroup I definitely originated in these areas, but merely ask - does this pattern indicate that these haplogroup I lineages are more likely to be ancestral to these areas where they are all now found, or ancestral to Southern Europe, the Caucasus, the Balkans or some other region not appearing anywhere at all in this mix?

To me, the theory that these earliest subclades each evolved somewhere else, then all relocated to North West Europe "in diverse waves of immigration" without leaving a trace in any of their points of origin, seems less plausible than the theory that they evolved approximately in the region where they are now all found.

This is simply based on DNA data, not on conjecture (such as 'conjectural maps'), nor on any uncertain premise, such as that I1 was "not present in Scandinavia in the Mesolithic" (we cannot know that).

The Swiss sample previously mentioned suggests that there was an early haplogroup I individual (whose DNA was more similar to I2a than to other surviving lineages) living North West of the Alps not too long after the Ice Age. This does not seem at all inconsistent with the premise proposed in the thread.

They cannot have 'originated' in NW EUrope, because when they did - NW Europe was under Ice.
It might be the case that simply that they better survived in northern Europe than they did in the south. This isn't too difficult to fathom.

The Bichon sample is believed to have been found within an Azilian context, which is post-Magdalenian, which spread from southern France.


From thousands of samples published on the internet, these are the estimated points of origin for the most diverse haplogroup I subclades, based on their degree of internal diversity - I* (pre- I1) - Britain/Norway; I2 (pre- I2b/I2c) - Germany; I2a2 - Britain/Low Countries; I2a1b - Britain; I2a1a (pre-I2a1a1) - Britain/NW France; I2a1a2 - Germany.

No genuine I* samples have been found - so you might be looking at old (outdated) data. And as Jean explained, looking at basal diversity is often wrong. Just look at previous ideas which placed R1b in Iberia, or Turkey becuase of supposed 'basal diversity'.
You need to look at collateral evidence, aDNA, and properly resolved modern data.

Megalophias
01-24-2016, 12:38 AM
I believe all the "" I*" previously found in the Caucasus is now known to be I2c- and one specific sub-branch of 3 otherwise found in Europe.
The recent study of Dagestan reports a Lak who is I-P19(xI1-P30, I2-P215). Di Cristofaro also found a Hazara with I-M258(xI1-M253, I2-M438) in Afghanistan, now that I check.

epp
01-24-2016, 01:03 AM
They cannot have 'originated' in NW EUrope, because when they did - NW Europe was under Ice.
The premise of the thread is that their most recent common ancestor might have lived after the Ice Age. I am looking at this period, not when NW Europe was under ice.


No genuine I* samples have been found - so you might be looking at old (outdated) data. And as Jean explained, looking at basal diversity is often wrong. Just look at previous ideas which placed R1b in Iberia, or Turkey becuase of supposed 'basal diversity'.
You need to look at collateral evidence, aDNA, and properly resolved modern data.
By I* (pre-I1) points of origin, I simply mean I1 samples whose ancestors would have been I* at the earliest stages of haplogroup I's development (as the I1 mutation seems to have occurred more recently than the I2 mutation).
If, as you say, basal diversity is often wrong, this doesn't mean that it has to be wrong in this case. In fact, like any other source of evidence, basal diversity is going to be right more often than wrong. Since basal diversity predicts a similar outcome (NW European ancestry) over each of the six sub-groups identified above, it would have to be coincidentally wrong six-fold to be wrong in general (inherently implausible, unless there is other evidence to the contrary - my interest in starting this thread is to find out if there is any reliable evidence to the contrary).
Why do you seem to imply that my underlying data (published by FTDNA) is not "properly resolved"? Is it because of misunderstanding what I meant by "I* (pre-I1)"?

Megalophias
01-24-2016, 01:10 AM
Why do you seem to mply that my underlying data (published by FTDNA) is not "properly resolved"? Is it because of misunderstanding what I meant by "I* (pre-I1)"?
How much of your FTDNA data is from Northwestern Europe vs from anywhere else?

epp
01-24-2016, 01:16 AM
The recent study of Dagestan reports a Lak who is I-P19(xI1-P30, I2-P215). Di Cristofaro also found a Hazara with I-M258(xI1-M253, I2-M438) in Afghanistan, now that I check.

Interesting. Gravetto-Danubian says that "no genuine I* samples have been found". Do you have the detailed data for these people? This would seem to be consistent with the premise that all other haplogroup I individuals (I1 and I2) whose data have been published descend from a common ancestor who migrated to North West Europe from the East.

Gravetto-Danubian
01-24-2016, 01:20 AM
The recent study of Dagestan reports a Lak who is I-P19(xI1-P30, I2-P215). Di Cristofaro also found a Hazara with I-M258(xI1-M253, I2-M438) in Afghanistan, now that I check.

Yes its very interesting. Grugni found IJ* in Iran, the Hazara I* and the Dagestani you mentioned. But its hard to say with just 4 or 5 SNPs tested, at best. They need full sequencing, or at least testing of more phyloequivalent SNPs.
BTW which Dagestani paper are you referring to ? Cagiagli 2009 ?

epp
01-24-2016, 01:27 AM
How much of your FTDNA data is from Northwestern Europe vs from anywhere else?

My data is FTDNA's entire published dataset. If you are inferring that North Western Europe is probably over-represented in FTDNA's data, that is probably correct, but the same methodology does not similarly predict a North Western European origin, for instance, for any of haplogroups C, E, F, G, H, J, N, O, P, Q, R, R1a, R1b or R2 - despite a European prevalence of some of these haplogroups.

epp
01-24-2016, 01:30 AM
Yes its very interesting. Grugni found IJ* in Iran, the Hazara I* and the Dagestani you mentioned. But its hard to say with just 4 or 5 SNPs tested, at best. They need full sequencing, or at least testing of more phyloequivalent SNPs.
BTW which Dagestani paper are you referring to ? Cagiagli 2009 ?

I agree that it's hard to infer anything without seeing the full data. That's why I've only used full data samples from a large reliable dataset.

Megalophias
01-24-2016, 01:32 AM
"Coevolution of genes and languages and high levels of population structure among the highland populations of Daghestan" was the latest paper, which tested for an upstream I2 marker.

No full sequences, of course. I hope in the future we will see many large full sequence studies like Francalacci, Hallast, Karmin, Yang, and the latest paper by Barbieri et al, targeted at various regions, including the Caucasus (but above all India!)

Megalophias
01-24-2016, 01:35 AM
My data is FTDNA's entire published dataset. If you are inferring that North Western Europe is probably over-represented in FTDNA's data, that is probably correct, but the same methodology does not similarly predict a North Western European origin, for instance, for any of haplogroups C, E, F, G, H, J, N, O, P, Q, R, R1a, R1b or R2 - despite a European prevalence of some of these haplogroups.

And have you tried for all those haplogroups? I wouldn't be surprised if you could predict a European origin for N and Q, at least.

Gravetto-Danubian
01-24-2016, 01:40 AM
My data is FTDNA's entire published dataset. If you are inferring that North Western Europe is probably over-represented in FTDNA's data, that is probably correct, but the same methodology does not similarly predict a North Western European origin, for instance, for any of haplogroups C, E, F, G, H, J, N, O, P, Q, R, R1a, R1b or R2 - despite a European prevalence of some of these haplogroups.

Where did you find the I*, or 'pre-I1' ?

epp
01-24-2016, 10:43 AM
And have you tried for all those haplogroups? I wouldn't be surprised if you could predict a European origin for N and Q, at least.
Yes. N & Q samples didn't indicate European ancestry, but some tests come out with clearer, more definitive answers than others - haplogroup I seems to show one of the clearest pictures.

epp
01-24-2016, 10:52 AM
Where did you find the I*, or 'pre-I1' ?

Sorry, I don't think I've explained myself clearly enough. By this, I mean current I1 samples. As I was analysing the position at a date early in haplogroup I's development (before the I1 mutation is estimated to have occurred), I was just identifying I1's ancestor at that time as being I*.

Gravetto-Danubian
01-24-2016, 11:22 AM
Sorry, I don't think I've explained myself clearly enough. By this, I mean current I1 samples. As I was analysing the position at a date early in haplogroup I's development (before the I1 mutation is estimated to have occurred), I was just identifying I1's ancestor at that time as being I*.

I understand (I think). So going back to your original statement
From thousands of samples published on the internet, these are the estimated points of origin for the most diverse haplogroup I subclades, based on their degree of internal diversity - I* (pre- I1) - Britain/Norway; I2 (pre- I2b/I2c) - Germany; I2a2 - Britain/Low Countries; I2a1b - Britain; I2a1a (pre-I2a1a1) - Britain/NW France; I2a1a2 - Germany.


I don't think that's quite getting at the crux of the dispersion of the respectively mentioned sub- groups . For example, I doubt that I1 first expanded from/ originated in Britain/ Norway. In fact, is bet it expanded from the Carpathian region/ EE during the late Neolithic, and possibly had secondary expansion in bronze age Scandinavia, and was later taken to England with Vikings and Saxons

epp
01-24-2016, 01:08 PM
I doubt that I1 first expanded from/ originated in Britain/ Norway. In fact, is bet it expanded from the Carpathian region/ EE during the late Neolithic
I am interested to see any data you have that might indicate that.

I have found the Hazara sample, which is accompanied by more detailed data. It is only one sample, so it's hard to read too much into it, and the researchers don't seem to have tested for I1, so it could be I* or I1. In fact, the data shows a close similarity to I1, suggesting a relatively recent common ancestor.

The Hazara sample diverges from the North Sea's I1 at a similar point to an estimated migration of I2c from North West to South East Europe. So, if the limited data in this study is reliable, it perhaps suggests the sample (either I* or I1) descended from a participant in a general migration (including I2c) eastwards. Like yours, my estimate for this migration would be late Neolithic and my estimated point of origin would be Germany (not too far from the Western Carpathians), so we don't differ much at all.

Of course, by this point, the data suggests that I2 (especially I2a) had already been developing for a number of millennia in North Western Europe.

lgmayka
01-24-2016, 03:17 PM
Like yours, my estimate for this migration would be late Neolithic and my estimated point of origin would be Germany (not too far from the Western Carpathians), so we don't differ much at all.
Perhaps this is one source of the confusion. I myself do not think of Germany as part of Northwestern Europe, but apparently most other people do (https://en.wikipedia.org/wiki/Northwestern_Europe).

gravetti
01-24-2016, 06:19 PM
Jean M: Ancient DNA is gradually resolving the mysteries of Y-DNA Haplogroup I (L41). Its modern distribution was puzzling. On the one hand it seemed ancient in Europe. It rarely appears outside the boundaries of Europe and European colonies.

http://www.nature.com/jhg/journal/v55/n5/full/jhg201030a.html

I-M170, a European-specific haplogroup, was considered as being of Balkan origin.29, 38 I-M170 lineages subsequently dispersed toward Caucasus region and Central Europe.49, 54 In our samples, I-M170 was observed exclusively in Tataer population at 33.3% frequency.

Gravetto-Danubian
01-24-2016, 09:40 PM
I have found the Hazara sample, which is accompanied by more detailed data. It is only one sample, so it's hard to read too much into it, and the researchers don't seem to have tested for I1, so it could be I* or I1. In fact, the data shows a close similarity to I1, suggesting a relatively recent common ancestor.

I'm not following. All we know is that the Hazara chap is I*- as in negative for 4 or 5 major defining SNPs for I2a1, I2a2, and I1. he could still be one of those, or I2c, if tested further (more SNPs, or FGS). Whatever the case, it is not directly relevant as to where the TMRCA of modern I1 originated from.

So, I'm not sure which 'data' you're seeing proves your theory on 'a relatively recent common ancestor' for I1 in Germany. It could be. But we have I1 from Neolithic Hungary. There is modest I1 in Balkans today. At the moment, we don't know if Balkan I1 descends from Neolithic people, Iron Age Germanics, both, or neither.


The Hazara sample diverges from the North Sea's I1 at a similar point to an estimated migration of I2c from North West to South East Europe. So, if the limited data in this study is reliable, it perhaps suggests the sample (either I* or I1) descended from a participant in a general migration (including I2c) eastwards. Like yours, my estimate for this migration would be late Neolithic and my estimated point of origin would be Germany (not too far from the Western Carpathians), so we don't differ much at all.

Which data ? When you quote 'data' you need to give concrete figures.


Of course, by this point, the data suggests that I2 (especially I2a) had already been developing for a number of millennia in North Western Europe.

Ditto.

I'd put it to you that history of haplogroup I & post-glacial Europe is much more complex than what you're seeing, and it wasn't simply a case of I drifting out from NW Europe

eastara
01-24-2016, 10:20 PM
Jean M: Ancient DNA is gradually resolving the mysteries of Y-DNA Haplogroup I (L41). Its modern distribution was puzzling. On the one hand it seemed ancient in Europe. It rarely appears outside the boundaries of Europe and European colonies.

http://www.nature.com/jhg/journal/v55/n5/full/jhg201030a.html

I-M170, a European-specific haplogroup, was considered as being of Balkan origin.29, 38 I-M170 lineages subsequently dispersed toward Caucasus region and Central Europe.49, 54 In our samples, I-M170 was observed exclusively in Tataer population at 33.3% frequency.

Very often Weterners make wrong conclusion without knowing the true origin of some population. The so called Chinese Tataer are in fact Russian Volga Tatars, who settled in China only recently. No surprise they have I1, I2a Dinaric and I2b1.
https://en.wikipedia.org/wiki/Chinese_Tatars
The same could be the case with the Hazara, it is known that some Tartars live in their area.

Remember the old theory haplogroup I originated in Moldova, as it has the most diversity of both haplogroups I1 and I2?
However Moldova is the classic genetic "sink" as at least 90% of the current population are recent migrant from 18 century on. When Russia took over Bessarabia it expelled most of the existing Crimean Tartar and Nogay(who were also recent migrant from Central Asia) population to the Ottoman Empire and Caucasus and resettled it with Russians and Ukrainians, as with Vlachs, Bulgarians, Gagauzes etc from the Balkans.

Megalophias
01-24-2016, 10:50 PM
I have found the Hazara sample, which is accompanied by more detailed data. It is only one sample, so it's hard to read too much into it, and the researchers don't seem to have tested for I1, so it could be I* or I1. In fact, the data shows a close similarity to I1, suggesting a relatively recent common ancestor.
They did test him for M253 (see the supp info). However, if the haplotype resembles I1, it could well be some pre-I1 branch. (G-D, they also tested for M438, so it can't be I2c.)


Yes. N & Q samples didn't indicate European ancestry, but some tests come out with clearer, more definitive answers than others - haplogroup I seems to show one of the clearest pictures.
I just had a look at FTDNA Haplogroup N project and noticed that there are more N samples from England than from all of East Asia - even though the number of Chinese men with N alone exceeds the entire male population of the UK. :P

I am curious though, what exactly is your methodology? How are you calculating and weighting the haplotype variance?

Dorkymon
01-24-2016, 10:54 PM
Remember the old theory haplogroup I originated in Moldova, as it has the most diversity of both haplogroups I1 and I2?
However Moldova is the classic genetic "sink" as at least 90% of the current population are recent migrant from 18 century on. When Russia took over Bessarabia it expelled most of the existing Crimean Tartar and Nogay(who were also recent migrant from Central Asia) population to the Ottoman Empire and Caucasus and resettled it with Russians and Ukrainians, as with Vlachs, Bulgarians, Gagauzes etc from the Balkans.

Hello there, most of the current population of Moldova is not composed of recent migrants. Only Russians, Gagauz and some Ukrainians settled here since the 18th century. Romanians who form the majority of the population have been living there from at least the 14th century, when the principality of Moldavia has been established and chroniclers started keeping records.

epp
01-25-2016, 12:21 AM
I'm not following. All we know is that the Hazara chap is I*- as in negative for 4 or 5 major defining SNPs for I2a1, I2a2, and I1. he could still be one of those, or I2c, if tested further (more SNPs, or FGS). Whatever the case, it is not directly relevant as to where the TMRCA of modern I1 originated from.
I couldn't find any I1 test on the Cristofaro study; perhaps I missed it. The study does, however, contain a chart which excludes the sample from M438, which would seem to rule it out as I2c. In any case, even a cursory look at the STRs reveals the sample that is more recently related to I1 individuals than I2. This does makes it indirectly (if not directly) relevant to where the TMRCA of modern I1 originated.


So, I'm not sure which 'data' you're seeing proves your theory on 'a relatively recent common ancestor' for I1 in Germany. It could be. But we have I1 from Neolithic Hungary. There is modest I1 in Balkans today. At the moment, we don't know if Balkan I1 descends from Neolithic people, Iron Age Germanics, both, or neither.
To clarify, I don't have a theory, and am merely noting the results of analysed published data - in this case, the STRs of the sample published in the study. I am not much interested in Balkan I1, which is all very similar and also similar to some North West European I1 haplotypes, indicating a very recent date of common origin in the Balkans, and therefore suggesting that it is probably not the location at which I1 split from I*.


Which data ? When you quote 'data' you need to give concrete figures.
Again, I wonder whether you are misinterpreting my intent. I am not looking for peer review or trying to prove anything. Anyone can take the same FTDNA databases that I have looked at and repeat the STR analysis for themselves. As I have mentioned in an earlier post, I am interested to see if there is any DNA evidence that is inconsistent with the result of my own data analysis exercise, suggesting that all major subclades of haplogroup I have most recent common ancestors located in North Western Europe after the Ice Age.


I'd put it to you that history of haplogroup I & post-glacial Europe is much more complex than what you're seeing, and it wasn't simply a case of I drifting out from NW Europe Yes, I am sure the history of haplogroup I is complex. However, what I am looking at is not complex per se. I2 split from I* (pre-I1) in one individual. I am interested in when and where this single event was likely to have taken place. The results of my data analysis suggest a 'most likely' answer to this question. I am wondering whether there is any other DNA data, rather than conjecture, that suggests a different answer.

eastara
01-25-2016, 12:24 AM
I was talking about Bessarabia, or Eastern Moldova. However, Western Moldova was only settled with Romanians (Vlachs) since late Middle Ages and not since the Mesolithics. There are some theories the real origin of the Romanian (Vlachs) is in fact the South West Balkans, they are not the direct descendants of the Dacians, Roman, etc, as in the official history books, which in facts shows in their DNA. These are the Romanians who used to live in the old Wallachia, South and East of the Carpathian mountains, not in Transylvania, Bukovina, etc ethnically mixed territories.

Gravetto-Danubian
01-25-2016, 12:31 AM
I couldn't find any I1 test on the Cristofaro study; perhaps I missed it. The study does, however, contain a chart which excludes the sample from M438, which would seem to rule it out as I2c. In any case, even a cursory look at the STRs reveals the sample that is more recently related to I1 individuals than I2. This does makes it indirectly (if not directly) relevant to where the TMRCA of modern I1 originated.


To clarify, I don't have a theory, and am merely noting the results of analysed published data - in this case, the STRs of the sample published in the study. I am not much interested in Balkan I1, which is all very similar and also similar to some North West European I1 haplotypes, indicating a very recent date of common origin in the Balkans, and therefore suggesting that it is probably not the location at which I1 split from I*.


Again, I wonder whether you are misinterpreting my intent. I am not looking for peer review or trying to prove anything. Anyone can take the same FTDNA databases that I have looked at and repeat the STR analysis for themselves. As I have mentioned in an earlier post, I am interested to see if there is any DNA evidence that is inconsistent with the result of my own data analysis exercise, suggesting that all major subclades of haplogroup I have most recent common ancestors located in North Western Europe after the Ice Age.

Yes, I am sure the history of haplogroup I is complex. However, what I am looking at is not complex per se. I2 split from I* (pre-I1) in one individual. I am interested in when and where this single event was likely to have taken place. The results of my data analysis suggest a 'most likely' answer to this question. I am wondering whether there is any other DNA data, rather than conjecture, that suggests a different answer.

Well there is Bichon sample - I2a1, 12000 BC. It is from Azillian culture - an epimagdalenian derivative- which incontrovertibly expanded from southwest Europe.

STR analyses aren't a particularly useful tool for what you're attempting to achieve - although it was thought they were in 2003. I'm also unsure why you're conflating I* with pre-I1 ? It doesn't make sense to me . Have you looked a basic haplogroup I tree - I* would be pre-I2 and pre- I1.

epp
01-25-2016, 12:36 AM
I just had a look at FTDNA Haplogroup N project and noticed that there are more N samples from England than from all of East Asia - even though the number of Chinese men with N alone exceeds the entire male population of the UK. :P

I am curious though, what exactly is your methodology? How are you calculating and weighting the haplotype variance?
Even though there are plenty of N samples from England, an analysis of the data still suggests an Asian common ancestor for these N samples. A large volume of English data that is similar to other English data does not affect the result. It is fairly standard methodology, calculating variance in STRs, aged according to average rates of mutation of STRs, and weighted by assessed reliability. I have found it produces answers that are generally very consistent with each other when there is a large enough sample size.

epp
01-25-2016, 12:59 AM
Well there is Bichon sample - I2a1, 12000 BC. It is from Azillian culture - an epimagdalenian derivative- which incontrovertibly expanded from southwest Europe.

STR analyses aren't a particularly useful tool for what you're attempting to achieve - although it was thought they were in 2003. I'm also unsure why you're conflating I* with pre-I1 ? It doesn't make sense to me . Have you looked a basic haplogroup I tree ?

The Bichon sample was carbon-dated at 9,850 BC - this was after Ice Age, as indeed was 12000 BC, if either of these dates from only one study are reliable.

Culture is not particularly informative, in my view. It can be transferred piecemeal, in fragments and at random rates, and is certainly not passed down intact from ancestor to descendant. I play golf, but it doesn't mean my ancestors are from Scotland.

I am conflating I* with pre-I1 because I* was the direct ancestor of I1. I'm a bit unclear why this doesn't make sense to you.

Kale
01-25-2016, 01:24 AM
The whole deal with I* is rather strange because it is the only haplogroup really in which the two major branches I1 and I2 coalesced at such different times. According to Yfull, I coalesces at 27300 years, I2 coalesces at 21700 years...then for a whopping 17000 years we have pre-I1 trotting around wherever and only at 4700 years ago coalesces into the TMRCA of I1 today.

If you are looking for the origin of I, I don't think looking at I1 is going to do you much good.

bix
01-25-2016, 02:27 AM
"Culture is not particularly informative, in my view. It can be transferred piecemeal, in fragments and at random rates, and is certainly not passed down intact from ancestor to descendant. I play golf, but it doesn't mean my ancestors are from Scotland."

The technological or lythic package could be something a little more involved than culture alone and I'd venture to say it was a bit more life and death than a game of golf. I mean, the tools and skills had to have been handed down with great effort and care. It's not easy killing a very large, injured, pissed off animal with arrows tipped with microblades or stone tipped spears, I imagine. You just didn't walk into a hunter gather group like you were applying for a job--or maybe our man in Bichon Cave did. Didn't end well for him or the bear.

Gravetto-Danubian
01-25-2016, 02:38 AM
The whole deal with I* is rather strange because it is the only haplogroup really in which the two major branches I1 and I2 coalesced at such different times. According to Yfull, I coalesces at 27300 years, I2 coalesces at 21700 years...then for a whopping 17000 years we have pre-I1 trotting around wherever and only at 4700 years ago coalesces into the TMRCA of I1 today.

If you are looking for the origin of I, I don't think looking at I1 is going to do you much good.

I have wandered if the major bottlenecking seen in I1 is because it originated somewhere in EE or the Balkans, due to being swamped by later arrivals (farmers (G), and pastoralists (R1)) - that and the lack of it being found in western Europe so far.
One lineage survived through the Neolithic perhaps as part of LBK in northern Europe, from where it later expanded during the Nordic Bronze Age

Megalophias
01-25-2016, 04:46 AM
The Bichon sample was carbon-dated at 9,850 BC - this was after Ice Age, as indeed was 12000 BC, if either of these dates from only one study are reliable.
Bichon was 13,560–13,770 cal BP, which is ~11 700 BC. This is during the Final Pleistocene, still within the Würm glacial "Ice Age", but after the Last Glacial Maximum and during the Bølling-Allerød warm period. Or possibly during the Older Dryas cold period within the warm period. :biggrin1: Anyway, it was still plenty cold but after the LGM, glaciers were retreating, populations could expand, which I believe is what you have in mind.


Even though there are plenty of N samples from England, an analysis of the data still suggests an Asian common ancestor for these N samples. A large volume of English data that is similar to other English data does not affect the result. It is fairly standard methodology, calculating variance in STRs, aged according to average rates of mutation of STRs, and weighted by assessed reliability. I have found it produces answers that are generally very consistent with each other when there is a large enough sample size.
I don't know how you'd get an Asian common ancestor when there are hardly any Asian samples, there are European branches of even N1a, and N2 is known only from Europe, but okay. Your method seems reasonable, even if I don't buy your conclusion.

Gravetto-Danubian
01-25-2016, 05:15 AM
Bichon was 13,560–13,770 cal BP, which is ~11 700 BC. This is during the Final Pleistocene, still within the Würm glacial "Ice Age", but after the Last Glacial Maximum and during the Bølling-Allerød warm period. Or possibly during the Older Dryas cold period within the warm period. :biggrin1:.

And Northern Europe was (a) still being re-colonised c. 13 ky BP (b) and was again de-colonised almost entirely during the Younger-Dryas cold snap (12 ky BP)

Conclusion : hg I could not have originally dispersed from Northern Europe . But this is not to say that some lineages we see today didn't secondarily arrive south from north of the Alps- case in point : I2a1b (my folks).

B)

Dorkymon
01-25-2016, 09:05 AM
I was talking about Bessarabia, or Eastern Moldova. However, Western Moldova was only settled with Romanians (Vlachs) since late Middle Ages and not since the Mesolithics. There are some theories the real origin of the Romanian (Vlachs) is in fact the South West Balkans, they are not the direct descendants of the Dacians, Roman, etc, as in the official history books, which in facts shows in their DNA. These are the Romanians who used to live in the old Wallachia, South and East of the Carpathian mountains, not in Transylvania, Bukovina, etc ethnically mixed territories.

You were actually talking about Bessarabia, which is currently a part of Ukraine, also known as Budjak. Eastern Moldova has been populated since the 14th century by Vlachs, who were invited to settle from Transylvania initially. There are castles and towns dating from that period.

And with regards to the origins, I would stick with the scientifically accepted theories, instead of speculations that are fueled by someone's national ambitions. If anything, y-dna alone is supportive of the Daco-Roman origin (I2a1b + R1a + R1b + J2 ~ 75% of paternal lineages).

epp
01-25-2016, 09:47 AM
The whole deal with I* is rather strange because it is the only haplogroup really in which the two major branches I1 and I2 coalesced at such different times. According to Yfull, I coalesces at 27300 years, I2 coalesces at 21700 years...then for a whopping 17000 years we have pre-I1 trotting around wherever and only at 4700 years ago coalesces into the TMRCA of I1 today.

If you are looking for the origin of I, I don't think looking at I1 is going to do you much good.
I1 is one piece of a two-piece (I1 and I2) jigsaw (or possible three-piece, if the Hazara sample is indeed I*). However, I would agree that I1 is the smaller of the two pieces.

epp
01-25-2016, 09:49 AM
"Culture is not particularly informative, in my view. It can be transferred piecemeal, in fragments and at random rates, and is certainly not passed down intact from ancestor to descendant. I play golf, but it doesn't mean my ancestors are from Scotland."

The technological or lythic package could be something a little more involved than culture alone and I'd venture to say it was a bit more life and death than a game of golf. I mean, the tools and skills had to have been handed down with great effort and care. It's not easy killing a very large, injured, pissed off animal with arrows tipped with microblades or stone tipped spears, I imagine. You just didn't walk into a hunter gather group like you were applying for a job--or maybe our man in Bichon Cave did. Didn't end well for him or the bear.
Christianity is a less facetious example of a highly intricate cultural shift that tells you little about biological ancestry. Few in Christianised Europe are or were descended from Jews.

epp
01-25-2016, 09:54 AM
I have wandered if the major bottlenecking seen in I1 is because it originated somewhere in EE or the Balkans, due to being swamped by later arrivals (farmers (G), and pastoralists (R1)) - that and the lack of it being found in western Europe so far.
One lineage survived through the Neolithic perhaps as part of LBK in northern Europe, from where it later expanded during the Nordic Bronze Age
Possibly, but a Balkan origin for I1 seems to lack evidence. In my view, a more likely explanation is that predominantly NW European I1 descended from roughly the same region both where it is currently prevalent and where it and its brother clade I2 appear to have had a most recent common ancestor.
However, a Balkan I* ancestor of this common ancestor of I1 and I2 is entirely consistent with my data.

epp
01-25-2016, 09:58 AM
[QUOTE=Megalophias;136093]Bichon was 13,560–13,770 cal BP, which is ~11 700 BC.QUOTE]
Possibly, although I understand that it was also secondarily dated at 9,800 BC. Not that I would place huge reliance on the results of a single study or sample.

Gravetto-Danubian
01-25-2016, 10:23 AM
Christianity is a less facetious example of a highly intricate cultural shift that tells you little about biological ancestry. Few in Christianised Europe are or were descended from Jews.

Ha ha. True but its a false analogy. The adoption of a religion cannot be compared to a life-line chaîne opératoire

Gravetto-Danubian
01-25-2016, 10:31 AM
Possibly, but a Balkan origin for I1 seems to lack evidence. In my view, a more likely explanation is that predominantly NW European I1 descended from roughly the same region both where it is currently prevalent and where it and its brother clade I2 appear to have had a most recent common ancestor.
However, a Balkan I* ancestor of this common ancestor of I1 and I2 is entirely consistent with my data.

We don;t know where I1 arose. But this is the current balance of evidence

- it is dated to have 'formed' 27 kya. That's pre-LGM. That means I1 was somewhere well south of "the same region both where it is currently prevalent " becuase 25-14 kya, that region was not habited human beings.

epp
01-25-2016, 10:55 AM
We don;t know where I1 arose. But this is the current balance of evidence

- it is dated to have 'formed' 27 kya. That's pre-LGM. That means I1 was somewhere well south of "the same region both where it is currently prevalent " becuase 25-14 kya, that region was not habited human beings.
But where does that 27kya date come from? On what evidence is it based? If true and if the Hazara sample were I*, my STR calculations would proportionately estimate the formation age of I2 to be pre-40,000 BC; this seems implausible. A Neolithic origin for I1 is more consistent with my own figures.

Gravetto-Danubian
01-25-2016, 11:00 AM
But where does that 27kya date come from? On what evidence is it based? If true and if the Hazara sample were I*, my STR calculations would proportionately estimate the formation age of I2 to be pre-40,000 BC; this seems implausible. A Neolithic origin for I1 is more consistent with my own figures.

Here (http://www.yfull.com/tree/I1/)

You can also have a look at the Karmin paper.

Its all based on WGS, which has been shown to be suprior than the STR -based estimates.

lgmayka
01-25-2016, 04:21 PM
But where does that 27kya date come from? On what evidence is it based?
He is referring to what YFull calls the "formed" date--the date at which the clade living today diverged from its siblings.

In other words, YFull is telling us that the I1 living today began to diverge from its siblings over 27,000 years ago; yet the common patrilineal ancestor of today's I1 lived only 4700 years ago. A single patrilineage stretches across the entire 22,000 years in between those dates! We cannot possibly know from modern DNA alone where that lone patrilineage wandered throughout so many millennia. It may have wandered north and south, east and west, multiple times in that period.

Another point of view is that this scenario is simply unbelievable, and that we will eventually find intermediate offshoots somewhere--presumably in some region that is still poorly sampled.

After all, I-L161 (http://yfull.com/tree/I-L161.1/) looked quite Western European until we found a 6900-year-old offshoot, I-Y13331 (http://yfull.com/tree/I-Y13331/), in Poland and Bulgaria.

Megalophias
01-25-2016, 04:53 PM
Possibly, although I understand that it was also secondarily dated at 9,800 BC. Not that I would place huge reliance on the results of a single study or sample.
The date quoted from the study was direct date on the skeleton itself. There was another study in 2008 (summary here (http://www.academia.edu/1214438/La_grotte_du_Bichon_un_site_pr%C3%A9historique_des _montagnes_neuch%C3%A2teloises)), which obtained 14 radiocarbon dates, including a bunch on wood, 2 on the bear, and 2 on the human skeleton. That study dates the occupation to the Allerød, "12th millennium before our era" (I couldn't get access to the whole thing). The youngest radiocarbon date quoted from that study comes to 10 763 - 11 021 cal BC using Calib 7.1. So definitely Pleistocene, though I don't see what difference that makes.


But where does that 27kya date come from? On what evidence is it based? If true and if the Hazara sample were I*, my STR calculations would proportionately estimate the formation age of I2 to be pre-40,000 BC; this seems implausible. A Neolithic origin for I1 is more consistent with my own figures.
OK, first let us be clear on terminology: the date of 27 (25-30) thousand years for the origin of I is for when I1 diverged from I2. The origin of I2a, that is the time when I2a split from I2c, is placed by Y-Full at about 22 (20-23) thousand years ago, and the origin of I1, the split of four different primary branches, at 4700 (3900-5500) years ago. These are all from Y-Full, but results from Karmin et al are very similar.

The dates are based on the usual SNP counting method, that is counting the number of SNPs in whole genome sequenced samples and multiplying by a mutation rate, allowing for the necessary confidence intervals. The mutation rate used by Karmin is based on securely dated ancient DNA, that of Y-Full is an average of ancient DNA and modern pedigree-based dates; the purely fossil rate is a little slower, but not by much. It is of course possible that the results are wrong, because mutation rates *can* vary. But they would have to be very far wrong for I to originate after the LGM.

What dates are you getting for I, I1, and I2 using STRs - I mean their TMRCAs - and how are you selecting the STR mutation rates? If you put I1 in the Neolithic then that is quite close to the SNP based result, actually a little older even, so it's strange that your date for I itself should be much younger.

epp
01-25-2016, 08:51 PM
Here (http://www.yfull.com/tree/I1/)

You can also have a look at the Karmin paper.

Its all based on WGS, which has been shown to be suprior than the STR -based estimates.
I've looked briefly at yfull. It's an interesting development, although from my perspective, I have some reservations - principally:
1. The results rely more on concluded evidence from a few 'experts', and testing appears less easy for others to replicate and authenticate from published data.
2. The terminology is misleading - that a subclade is said to have been 'formed' at the date of it's parent clade's TMRCA gives the impression that its defining mutation occurred at that date; such misleading terminology is suggestive of possible bias (here, in favour of making the subclade's mutation appear older than it is). In this instance, stating that I1 was 'formed' 27ybp masks the lack of evidence that the I1 mutation occurred at any time before its 4700ybp TMRCA.
3. If age is based upon the number of defining mutations, I1 has 298 SNPs according to yfull, but only 24 according to the ISOGG. Presumably, the other 274 are what ISOGG would call 'experimental' or 'under investigation', so conclusions would be partly speculative even if there were no doubts about the methodology.

I would be grateful if you would let me know how and in which way this analysis has been demonstrated to be superior to STRs in estimating TMRCAs.

epp
01-25-2016, 09:08 PM
What dates are you getting for I, I1, and I2 using STRs - I mean their TMRCAs - and how are you selecting the STR mutation rates? If you put I1 in the Neolithic then that is quite close to the SNP based result, actually a little older even, so it's strange that your date for I itself should be much younger.
My TMRCAs are about 60% of yfull's, on average.
I don't put I1 in the Neolithic. Previously, I estimated the earliest possible age for I1 as late Neolithic, based on its calculated split point with the Hazaran I* sample identified in an earlier post. Of course, when it split with the Hazaran sample, it could still have been I*, with I1 per se only properly arising much later.

Megalophias
01-25-2016, 09:35 PM
I've looked briefly at yfull. It's an interesting development, although from my perspective, I have some reservations - principally:
1. The results rely more on concluded evidence from a few 'experts', and testing appears less easy for others to replicate and authenticate from published data.
It isn't based on "experts", it's based mostly on private samples that people have asked them to analyze, plus quite a lot of publicly available genome sequences. But it is whole genomes, huge files, not a small set of STR values you can put on one page. However, you can certainly get the files (the public ones, you'd obviously have to ask individuals for their private ones) and run the analysis yourself. There are a few academic studies with detailed SNP results too, namely Hallast et al "The Y-chromosomal tree bursts into leaf" and Karmin et al "A recent bottleneck of Y-chromosome diversity", you can download the information and figure out TMRCAs for yourself from those.


2. The terminology is misleading - that a subclade is said to have been 'formed' at the date of it's parent clade's TMRCA gives the impression that its defining mutation occurred at that date; such misleading terminology is suggestive of possible bias (here, in favour of making the subclade's mutation appear older than it is). In this instance, stating that I1 was 'formed' 27ybp masks the lack of evidence that the I1 mutation occurred at any time before its 4700ybp TMRCA.
You're being paranoid. They aren't trying to hide anything, there is a link to their methodology right on the page, the TMRCA is right next to the 'formed' age. There is no single I1 mutation, there are hundreds, and obviously we can't tell what order they occurred in (well, we know a little from ancient DNA but not much).


3. If age is based upon the number of defining mutations, I1 has 298 SNPs according to yfull, but only 24 according to the ISOGG. Presumably, the other 274 are what ISOGG would call 'experimental' or 'under investigation', so conclusions would be partly speculative even if there were no doubts about the methodology.
ISOGG is incredibly behind on everything, not surprising since there is loads of new information constantly coming out. You can look at the reliable, academic studies I referred to above for independent SNP sets.

This is not watertight, we don't know for sure that the SNP counting dates are right, though they probably are in the right ballpark anyway. So far they fit with ancient DNA but we need more ancient full genomes to anchor the tree. Old STR methods came up with contradictory results, so they get a bad rap, but newer methods are supposed to be better.

Megalophias
01-25-2016, 09:50 PM
My TMRCAs are about 60% of yfull's, on average.
I don't put I1 in the Neolithic. Previously, I estimated the earliest possible age for I1 as late Neolithic, based on its calculated split point with the Hazaran I* sample identified in an earlier post. Of course, when it split with the Hazaran sample, it could still have been I*, with I1 per se only properly arising much later.

Hmm. That would be impossible for really old clades, but I don't suppose you are aiming to calculate those. Have you calculated the age of clades that we have ancient DNA evidence for, to test whether your dates fall below the lower bound? R1b-P312 for example.

epp
01-25-2016, 11:28 PM
Old STR methods came up with contradictory results, so they get a bad rap, but newer methods are supposed to be better.
Results that occasionally look a bit odd suggests they are genuine. When a paranoid person like me sees subclades that are calculated to be younger than their parent clades in every circumstance, I become less confident, especially when the dates on which they are said to be "formed" precede the dates on which their mutations actually occurred; Your comment that these methods "are supposed to be" better unfortunately doesn't reassure me sufficiently! Although having said that, I am still open to their conclusions being near the mark.

epp
01-25-2016, 11:49 PM
Have you calculated the age of clades that we have ancient DNA evidence for, to test whether your dates fall below the lower bound? R1b-P312 for example.
P312 comes out late Bronze Age, if that tells you anything.

Even if we accept yfull's much larger TMRCA estimates, this still only identifies six haplogroup I subclades in existence before the Ice Age was in retreat - I*, I2*, I2a*, I2a1*, I2a1a* and I2a2*; and in FTDNA's databases, there is not a single current sample within any of these subclades. Even these newer methods seem to show the development of haplogroup I's surviving lineages occurring post-Ice Age. And irrespective of the dating differences, the greatest diversity within each of the major subclades is found in North West Europe.

eastara
01-26-2016, 12:12 AM
You were actually talking about Bessarabia, which is currently a part of Ukraine, also known as Budjak. Eastern Moldova has been populated since the 14th century by Vlachs, who were invited to settle from Transylvania initially. There are castles and towns dating from that period.

And with regards to the origins, I would stick with the scientifically accepted theories, instead of speculations that are fueled by someone's national ambitions. If anything, y-dna alone is supportive of the Daco-Roman origin (I2a1b + R1a + R1b + J2 ~ 75% of paternal lineages).

No, Bessarabia is a big part of current Moldova as you can see here:
https://en.wikipedia.org/wiki/Bessarabia

However, Moldova and Dobrudja were first the hinterland of the Old Bulgars, Romanians came there around 1000 later.
Although many Bulgarians would like to be descending largely of the the Old Bulgars, DNA testing shows different - if they were indeed a Cental Asian Turkic or even Iranian tribe, nowadays we have very little of that genetically.

Megalophias
01-26-2016, 01:10 AM
Results that occasionally look a bit odd suggests they are genuine. When a paranoid person like me sees subclades that are calculated to be younger than their parent clades in every circumstance, I become less confident, especially when the dates on which they are said to be "formed" precede the dates on which their mutations actually occurred.
Of course the "formed" date precedes when the mutations occurred, it is the date at which the branch diverged from the parent haplogroup, which is by definition the TMRCA of the parent clade. "Divergence" date might be a better term, but if you bother to click on the link at the bottom of the page the FAQ explains what all the terms mean and how everything is calculated. The subclades are never older because they bump up the age of the parent clade to the age of its oldest subclade when that happens (personally I'd prefer they left it as is, but whatever). However, if you click the little "info" button beside each entry, it will tell you the number of SNPs in each sample in the clade, the calculated age for each sample in the clade, and the calculated age of the clade without correction, and it will tell you if they did increase the age to match the subclade. It's all there, man! They really aren't hiding anything.


Your comment that these methods "are supposed to be" better unfortunately doesn't reassure me sufficiently!
I was saying that new *STR* methods are supposed to be better. Meaning I assume *your* method doesn't suck like the old ones. :biggrin1:


P312 comes out late Bronze Age, if that tells you anything.
There are ancient samples from the Early Bronze Age (2026-1885 BC) with mutations *downstream* of P312. There is also a U152+ Bell Beaker sample but it isn't directly dated, though presumably it is from the 2nd half of the 3rd millennium BC. I think there is directly dated from Bell Beaker sample from around 2200-2300 BC but it is only P312+.


Even if we accept yfull's much larger TMRCA estimates, this still only identifies six haplogroup I subclades in existence before the Ice Age was in retreat - I*, I2*, I2a*, I2a1*, I2a1a* and I2a2*; and in FTDNA's databases, there is not a single current sample within any of these subclades. Even these newer methods seem to show the development of haplogroup I's surviving lineages occurring post-Ice Age. And irrespective of the dating differences, the greatest diversity within each of the major subclades is found in North West Europe.
Sure, it would make sense for major subclades of I2 to have formed during the expansion of western European populations after the end of the LGM.

Gravetto-Danubian
01-26-2016, 01:16 AM
P312 comes out late Bronze Age, if that tells you anything.

Even if we accept yfull's much larger TMRCA estimates, this still only identifies six haplogroup I subclades in existence before the Ice Age was in retreat - I*, I2*, I2a*, I2a1*, I2a1a* and I2a2*; and in FTDNA's databases, there is not a single current sample within any of these subclades. Even these newer methods seem to show the development of haplogroup I's surviving lineages occurring post-Ice Age. And irrespective of the dating differences, the greatest diversity within each of the major subclades is found in North West Europe.

I think your efforts a laudable, but one gets the feeling that you choose to ignore certain parts of the evidence

Moreover, you take at face value that the Hazara chap is actually I*. He might not be, because testing was done by RFLP -based SNP testing, which could produce erroneous results - misreads. IMO, all such I* results are to be taken with caution unless reproduced with WGS. They might turn out to nestle well within an already described haplogroup, or a new, minor one. That is to say, I doubt we are really seeing a case of genuine I* (but happy to be proven wrong).

In fact, having looked at the STR patterns of the individual into 2 different haplogroup predictors (base on 28 STR haplotypes)- they suggested the individual could be haplogroup G2 (Cullen: G2 60%, F 25%, I1 6%; Athey: 100% G2 probability)


7465

gravetti
01-26-2016, 08:23 AM
eastara:"Although many Bulgarians would like to be descending largely of the the Old Bulgars, DNA testing shows different ..."
Is there any Old Bulgar aDNA? Could you please send me the link?

Gravetto-Danubian
01-26-2016, 08:50 AM
eastara:"Although many Bulgarians would like to be descending largely of the the Old Bulgars, DNA testing shows different ..."
Is there any Old Bulgar aDNA? Could you please send me the link?

A commonly believed myth by some is that the Bulgars "disappeared" whilst extinct Thracians somehow passed on their genes

gravetti
01-26-2016, 11:18 AM
A commonly believed myth by some is that the Bulgars "disappeared" whilst extinct Thracians somehow passed on their genes
Interesting.The origins of the Thracians remain obscure, in the absence of written historical records.

Dorkymon
01-26-2016, 12:47 PM
No, Bessarabia is a big part of current Moldova as you can see here:
https://en.wikipedia.org/wiki/Bessarabia

However, Moldova and Dobrudja were first the hinterland of the Old Bulgars, Romanians came there around 1000 later.
Although many Bulgarians would like to be descending largely of the the Old Bulgars, DNA testing shows different - if they were indeed a Cental Asian Turkic or even Iranian tribe, nowadays we have very little of that genetically.

You are obviously reading dubious sources if you imply that Bulgarians stretched on the whole terrtiory of Bessarabia, a Soviet term that encompasses the Eastern side of the Romanian principality of Moldavia (in Ukraine this is associated with Bugeac). Anyway, Old Bulgars have been living at the outskirts of the current Republic of Moldova for a long time, near the location of the present-day Gagauz people and spilling into Bugeac and Dobrogea, which is currently a part of Ukraine and Romania. Romanians populated Moldova in the 13-14th centuries, coming from Transylvania and not South of the Danube.

https://upload.wikimedia.org/wikipedia/commons/c/cc/Valaques-Vlachs.jpg

I'd stop here in order not to trash this thread with an offtopic. If you want to continue this discussion then feel free to message me.

gravetti
01-26-2016, 05:48 PM
Dorkymon you are obviously reading dubious sources too if you imply that Romanians populated Moldova in the 13-14th centuries, coming from Transylvania.What's your source?

https://en.wikipedia.org/wiki/First_Bulgarian_Empire

https://www.google.se/search?q=bulgarian+empire&newwindow=1&biw=1093&bih=479&tbm=isch&tbo=u&source=univ&sa=X&ved=0ahUKEwiE9OD_g8jKAhVE83IKHVonAMYQsAQIKA

Dorkymon
01-26-2016, 07:40 PM
Hungarian chronicles from that period mainly, such as Gesta Hungarorum. I wouldn't consider them dubious since they are foreign and come from a historic nemesis. Vlachs did not stick only to the current area that is inhabited by the Romanian people, but they were spread all over the Balkans, plus Poland, Slovakia and Czechia later in the middle ages. So, this is confusing and has created the conditions for a lot of conspiracy theories to develop.

https://en.wikipedia.org/wiki/Moldavia

https://en.wikipedia.org/wiki/Origin_of_the_Romanians

https://en.wikipedia.org/wiki/Romania_in_the_Early_Middle_Ages

https://en.wikipedia.org/wiki/Gesta_Hungarorum

https://en.wikipedia.org/wiki/Moravian_Wallachia

https://en.wikipedia.org/wiki/Vlach_law
Map of the North Balkans and present day Romania according to the narration in the 9th century Gesta Hungarorum:

https://upload.wikimedia.org/wikipedia/commons/thumb/6/6d/Gesta_hungarorum_map.jpg/1280px-Gesta_hungarorum_map.jpg

epp
01-26-2016, 08:58 PM
I think your efforts a laudable, but one gets the feeling that you choose to ignore certain parts of the evidence

Moreover, you take at face value that the Hazara chap is actually I*. He might not be, because testing was done by RFLP -based SNP testing, which could produce erroneous results - misreads. IMO, all such I* results are to be taken with caution unless reproduced with WGS. They might turn out to nestle well within an already described haplogroup, or a new, minor one. That is to say, I doubt we are really seeing a case of genuine I* (but happy to be proven wrong).

In fact, having looked at the STR patterns of the individual into 2 different haplogroup predictors (base on 28 STR haplotypes)- they suggested the individual could be haplogroup G2 (Cullen: G2 60%, F 25%, I1 6%; Athey: 100% G2 probability)


7465
This is why I prefer to rely on analysis based on large, reliable sources of publicly accessible data, such as FTDNA, and not on unverified data (DNA or archaeological) from individual studies.

epp
01-26-2016, 09:46 PM
Of course the "formed" date precedes when the mutations occurred, it is the date at which the branch diverged from the parent haplogroup, which is by definition the TMRCA of the parent clade. "Divergence" date might be a better term, but if you bother to click on the link at the bottom of the page the FAQ explains what all the terms mean and how everything is calculated.
That's just it. yfull defines I1's "formed" date as the date when they estimate that brother I2's ancestor separated from other I*, even though this date tells you nothing about when I1 itself was later formed from a descendant of the remaining I* population.
It doesn't even tell you when I1 "diverged" - only when an ancestor of I2's diverged from other I.

The subclades are never older because they bump up the age of the parent clade to the age of its oldest subclade when that happens (personally I'd prefer they left it as is, but whatever).
Exactly - why not bump down the subclade age to that of the parent clade? A choice to bump up, rather than to bump down or to leave as it is, indicates a possible preference for/bias towards maximising DNA age. You must recognise from being a senior member of this site that people interested in DNA have all sorts of odd biases.

However, if you click the little "info" button beside each entry, it will tell you the number of SNPs in each sample in the clade, the calculated age for each sample in the clade, and the calculated age of the clade without correction, and it will tell you if they did increase the age to match the subclade. It's all there, man! They really aren't hiding anything.
The information is hidden - behind, as you say, FAQ and Info buttons. Most will probably not bother to click on these.

Does yfull publish geographical locations of samples within its analysis so that you can check for regional patterns, as is possible with FTDNA's STR data?

lgmayka
01-26-2016, 10:15 PM
Does yfull publish geographical locations of samples within its analysis so that you can check for regional patterns, as is possible with FTDNA's STR data?
Customers have to enter that information themselves. Some do, some don't. I-Y6340 (http://yfull.com/tree/I-Y6340/) is definitely Polish.

Megalophias
01-26-2016, 10:30 PM
This is ridiculous, you are just throwing out random absurd objections. :crazy: Can we stop and get back to STR analysis?


That's just it. yfull defines I1's "formed" date as the date when they estimate that brother I2's ancestor separated from other I*, even though this date tells you nothing about when I1 itself was later formed from a descendant of the remaining I* population. It doesn't even tell you when I1 "diverged" - only when an ancestor of I2's diverged from other I.
How could they *possibly* know when I1 diverged from unknown I* that no one has sequences from? Sorcery? This is ridiculous beyond reason. It tells you when they estimate that pre-I1 diverged from pre-I2, which is all it is supposed to do. The TMRCA of I1 is right there next to it!


The information is hidden - behind, as you say, FAQ and Info buttons. Most will probably not bother to click on these.
Again this is ridiculous. If people look at information someone has kindly provided, don't understand what it means, don't bother to look up what it means, don't read the explanation that is linked for them *right frigging there*, and draw the wrong conclusions, that is certainly not the fault of the people who are presenting the information, for heaven's sake.


Does yfull publish geographical locations of samples within its analysis so that you can check for regional patterns, as is possible with FTDNA's STR data?
All of the public samples, and many of the private ones, have their origin labelled right next to them, as surely you noticed. Of course they can't publish geographical information from people who don't give them permission.

eastara
01-26-2016, 11:12 PM
Hungarian chronicles from that period mainly, such as Gesta Hungarorum. I wouldn't consider them dubious since they are foreign and come from a historic nemesis. Vlachs did not stick only to the current area that is inhabited by the Romanian people, but they were spread all over the Balkans, plus Poland, Slovakia and Czechia later in the middle ages. So, this is confusing and has created the conditions for a lot of conspiracy theories to develop.

https://en.wikipedia.org/wiki/Moldavia

https://en.wikipedia.org/wiki/Origin_of_the_Romanians

https://en.wikipedia.org/wiki/Romania_in_the_Early_Middle_Ages

https://en.wikipedia.org/wiki/Gesta_Hungarorum

https://en.wikipedia.org/wiki/Moravian_Wallachia

https://en.wikipedia.org/wiki/Vlach_law
Map of the North Balkans and present day Romania according to the narration in the 9th century Gesta Hungarorum:



Nobody denies the spread of the Vlachs, who in their march North reached South Poland an Western Ukraine. The dispute is where they first came from. Most contemporary historians think it is very unlikely the centre of migration was somewhere North of the Danube, but rather the area which is now Eastern Serbia, Macedonia and North Western Bulgaria.
It is also irrelevant when exactly they came to Moldova - 14th or 18th c. as they are incredibly late migrants there as far as the age and the origin of haplogroup I goes.

Gravetto-Danubian
01-27-2016, 01:05 AM
Nobody denies the spread of the Vlachs, who in their march North reached South Poland an Western Ukraine. The dispute is where they first came from. Most contemporary historians think it is very unlikely the centre of migration was somewhere North of the Danube, but rather the area which is now Eastern Serbia, Macedonia and North Western Bulgaria.
It is also irrelevant when exactly they came to Moldova - 14th or 18th c. as they are incredibly late migrants there as far as the age and the origin of haplogroup I goes.

Some would say that Vlachs (from the Balkans) aren't the same as Romance speakers north of the Danube, ie are not 'ancestral' to them. They’d further argue that there’s no evidence of any mass-migration of “Vlachs” from Macedonia to Moldavia.
(although they might be wrong ;))

Dorkymon
01-27-2016, 09:59 AM
Some would say that Vlachs (from the Balkans) aren't the same as Romance speakers north of the Danube, ie are not 'ancestral' to them. They’d further argue that there’s no evidence of any mass-migration of “Vlachs” from Macedonia to Moldavia.
(although they might be wrong ;))

Some of them are same and this is further supported by the fact that Romanians have a higher frequency of J2 than any of its neighbours. Logically, J2 should increase the closer a country is to Greece and Italy, but Romania seems to host an usually high proportion of J2. This is only explained by the fact that the dominant haplogroup in Aromanians (Vlachs from North Greece and Albania) is J2.


One fourth of the Vlach people (isolated communities of Romance language speakers in the Balkans) belong to J2, considerably more than the average of Macedonia and northern Greece where they live. This, combined to the fact that they speak a language descended from Latin, suggests that they could have a greater part of Roman (or at least Italian) ancestry than other ethnic groups in the Balkans. (http://www.eupedia.com/europe/Haplogroup_J2_Y-DNA.shtml)


Nobody denies the spread of the Vlachs, who in their march North reached South Poland an Western Ukraine. The dispute is where they first came from. Most contemporary historians think it is very unlikely the centre of migration was somewhere North of the Danube, but rather the area which is now Eastern Serbia, Macedonia and North Western Bulgaria.
It is also irrelevant when exactly they came to Moldova - 14th or 18th c. as they are incredibly late migrants there as far as the age and the origin of haplogroup I goes.

I personally believe in a large spread of the Vlachs, or call them proto-Balkan Romance speakers if you want, all over the Balkans prior to the en masse migration of the Slavs and other groups of people. Gradually, they would be assimilated in the numerically dominant cultures and only some pockets of Romance would survive in the Balkans and North of it.

http://i.imgur.com/AOWUoD1.jpg

gravetti
01-27-2016, 10:30 AM
Dorkymon:Map of the North Balkans and present day Romania according to the narration in the 9th century Gesta Hungarorum...

According to Neagu Djuvara, the factual accuracy of Anonymus's work is likely high, because it is the earliest preserved Hungarian chronicle and is based on even older Hungarian chronicles.[113] On the other hand, Carlile Aylmer Macartney described Anonymus's work as "the most famous, the most obscure, the most exasperating and most misleading of all the early Hungarian texts" in his book of medieval Hungarian historians.[22] Paul Robert Magocsi also regarded the Gesta as an unreliable work.[114]

Dorkymon
01-27-2016, 11:10 AM
Find me a chronicle that does not suffer from controversy. And in this case the controversy comes from historians of Hungarian descent, so it should be quite obvious what I am hinting at. The document is there for those who doubt Romanian sources about the continuity North of Danube.

Gravetto-Danubian
01-27-2016, 11:38 AM
Interesting.The origins of the Thracians remain obscure, in the absence of written historical records.

Gravetti,
I have replied that question here (http://www.anthrogenica.com/showthread.php?6322-aDNA-of-South-Eastern-Europe-Catalogue-of-results-amp-impact-in-modern-groups)

Gravetto-Danubian
01-27-2016, 11:40 AM
Suggestion to MODERATOR. Maybe discussions about genetics of Balkan groups in particular be moved here (http://www.anthrogenica.com/showthread.php?6322-aDNA-of-South-Eastern-Europe-Catalogue-of-results-amp-impact-in-modern-groups)

epp
01-27-2016, 09:31 PM
How could they *possibly* know when I1 diverged from unknown I* that no one has sequences from? Sorcery?
So why act as if they are estimating something that they cannot possibly know? Better to say nothing, in my view. To state that I1 was "formed" long before there is any evidence to show that it existed is misleading, regardless of whether they explain in the small print that they mean something entirely different.

I am a 20th cousin of Queen Elizabeth II. Does this mean both I and Her Royal Majesty were formed in 1360, when we shared our most recent common ancestor?


Again this is ridiculous. If people look at information someone has kindly provided, don't understand what it means, don't bother to look up what it means, don't read the explanation that is linked for them *right frigging there*, and draw the wrong conclusions, that is certainly not the fault of the people who are presenting the information, for heaven's sake.
OK. Let's not get too up-tight about it ... it's only a tiny little bit of DNA we're talking about!


All of the public samples, and many of the private ones, have their origin labelled right next to them, as surely you noticed.
No, I didn't notice. But now that you have pointed that out, what I do notice is that there is only haplogroup I sample (out of over a thousand on yfull) that falls within a subclade (M423*) whose yfull TMRCA comes anywhere near the Ice Age (11,800 BC) - and this sample is from Luxembourg (the area of North West Europe from where my STR analysis estimates that haplogroup I developed after the Ice Age).
Similarly, samples falling within the I1 subclades that they estimate to be oldest are shown to be Finnish and British - nowhere near any Southern refugia.

I can appreciate that analyses based on STR and SNP mutation are both valid, and in general give proportionately consistent answers. The only difference seems to be in the magnitude of ageing. So I think I will recalculate based on an average of the two to see how it comes out.

epp
01-28-2016, 11:14 AM
Analysis of the limited sample locations identified in yfull indicates the following migratory patterns as most likely in the haplogroup I population:
1. Most recent common ancestor in Germany/Britain/North Sea/Western Baltic region - 25,300 BC
2. First surviving migration (I2a1a1 L158) South West to Spain - 6,500 BC (three quarters of the way into haplogroup I's estimated lifespan)
3. Migration (I2a1b2 L621) East to Poland - 4,500 BC
4. Migrations (I2a2a1a Y3721 and a branch of I2c Y16419) South East to Caucasus/Balkans - 3,400-2,200 BC
5. Migrations (CTS6868 & Z2040 branches of I1a1b & I11a2a) East to Finland - 2,000 BC
No other expansions are observed before 1,000 BC.
Apart from the dating, this is very similar to analysis previously obtained from FTDNA's STR data.

It suggests haplogroup I's foundation and initial development in North Western Europe at around the time of the last glacial maximum, and a much later migration to Southern Europe just as the continent was significantly warming up. It doesn't indicate escape from extreme cold to a Southern refuge (in fact, exactly the opposite), and its estimate of Northern growth & development near the point of the last glacial maximum is highly implausible.

Later, I will analyse dates & locations based on an average of yfull's SNP data and FTDNA's STR data. Hopefully, this will yield more plausible results.

Megalophias
01-29-2016, 12:41 AM
It suggests haplogroup I's foundation and initial development in North Western Europe at around the time of the last glacial maximum, and a much later migration to Southern Europe just as the continent was significantly warming up. It doesn't indicate escape from extreme cold to a Southern refuge (in fact, exactly the opposite), and its estimate of Northern growth & development near the point of the last glacial maximum is highly implausible.
Why do you figure there was growth at the time of the LGM? There is no star-like expansion or even rapid bifurcation on the tree until the Neolithic or later, what's there to tie the coalescence dates to a population expansion?


Later, I will analyse dates & locations based on an average of yfull's SNP data and FTDNA's STR data. Hopefully, this will yield more plausible results.
That sounds like an interesting project. I hope FTDNA has a decent number of French and Spanish samples. Y-Full's I2 tree has 2 of each, which is a little problematic if one is trying to find traces of the Franco-Cantabrian refuge population. :biggrin1:

epp
01-29-2016, 08:44 PM
Why do you figure there was growth at the time of the LGM? There is no star-like expansion or even rapid bifurcation on the tree until the Neolithic or later, what's there to tie the coalescence dates to a population expansion?
It's likely there was growth relatively shortly after the LGM, because we see the haplogroup split into five surviving lineages (pre-I1, I2, I2a2, pre-I2a1b and I2a1a) during a period of a few thousand years (according to yfull). This was unlikely to have been rapid growth, but shows more signs of successful development than in the preceding 20,000 years.


That sounds like an interesting project. I hope FTDNA has a decent number of French and Spanish samples. Y-Full's I2 tree has 2 of each, which is a little problematic if one is trying to find traces of the Franco-Cantabrian refuge population. :biggrin1:
I have 95 samples of I2 from France, Iberia and Hispanic America.

Gravetto-Danubian
01-29-2016, 09:20 PM
It's likely there was growth relatively shortly after the LGM, because we see the haplogroup split into five surviving lineages (pre-I1, I2, I2a2, pre-I2a1b and I2a1a) during a period of a few thousand years (according to yfull). This was unlikely to have been rapid growth, but shows more signs of successful development than in the preceding 20,000 years.

Makes sense on balance, but I suspect I* was already split before the LGM .

epp
01-29-2016, 09:38 PM
Later, I will analyse dates & locations based on an average of yfull's SNP data and FTDNA's STR data. Hopefully, this will yield more plausible results.
So here is my approximate average of yfull SNP and FTDNA STR calculations. Both sets of data give a very similar estimates of geographical flows. The dating is different in magnitude between the two (the SNP analysis gives older results), but they are proportionately similar.

Initial development - formation of pre-I1, I2, I2a2, pre-I2a1b and I2a1a around the time of the first significant glacial melting (between 15,500 and 10,000 BC) - all North Sea and the Rhine, as far South as Switzerland.

First significant migrations away from this region - I2a1a1 from Northern & Eastern France towards Iberia 4,000 BC ; I2a1b2 from Germany towards Poland 2,700 BC; I1a1 from Sweden towards Finland 1,230 BC.

Does this look plausible?

epp
01-29-2016, 09:41 PM
Makes sense on balance, but I suspect I* was already split before the LGM .
Yes, I would at least imagine that there were a number of haplogroup I lineages that didn't survive (or haven't yet been found).

Gravetto-Danubian
01-30-2016, 12:09 AM
Analysis of the limited sample locations identified in yfull indicates the following migratory patterns as most likely in the haplogroup I population:
1. Most recent common ancestor in Germany/Britain/North Sea/Western Baltic region - 25,300 BC
2. First surviving migration (I2a1a1 L158) South West to Spain - 6,500 BC (three quarters of the way into haplogroup I's estimated lifespan)
3. Migration (I2a1b2 L621) East to Poland - 4,500 BC
4. Migrations (I2a2a1a Y3721 and a branch of I2c Y16419) South East to Caucasus/Balkans - 3,400-2,200 BC
5. Migrations (CTS6868 & Z2040 branches of I1a1b & I11a2a) East to Finland - 2,000 BC
No other expansions are observed before 1,000 BC.
Apart from the dating, this is very similar to analysis previously obtained from FTDNA's STR data.

It suggests haplogroup I's foundation and initial development in North Western Europe at around the time of the last glacial maximum, and a muc later migration to Southern Europe just as the continent was significantly warming up. It doesn't indicate escape from extreme cold to a Southern refuge (in fact, exactly the opposite), and its estimate of Northern growth & development near the point of the last glacial maximum is highly implausible.

Later, I will analyse dates & locations based on an average of yfull's SNP data and FTDNA's STR data. Hopefully, this will yield more plausible results.

the data is telling me a different story - esp 1-3. Perhaps what's muddling you is your only working with modern TMRCAs from solely modern samples. IMO- it fails a basic archaeological - historical "litmus test"- like how I* expanded from NW europe 25 KYA, and when it was under ice

___________________
Europe during LGM.

http://www.anthrogenica.com/attachment.php?attachmentid=7534&d=1454117110&

epp
01-30-2016, 01:29 AM
the data is telling me a different story - esp 1-3. Perhaps what's muddling you is your only working with modern TMRCAs from solely modern samples. IMO- it fails a basic archaeological - historical "litmus test"- like how I* expanded from NW europe 25 KYA, and when it was under ice

7534
What is the data, archaeological or otherwise, that identifies the earliest surviving haplogroup I lineages as having lived other than in the general vicinity of the areas that I have identified (Switzerland to the North Sea)? And what evidence is there to counter the premises that L158 migrated to Spain from the North and that L621 migrated to Poland from the West? (By the way, the sample from yfull's estimated oldest surviving subclade is an archaeological one, discovered in Luxembourg - right in the middle of the area in which both yfull and FTDNA data suggest I's earliest development most likely occurred.)

bix
01-30-2016, 02:22 AM
The Swiss Plateau was recolonized by Magdalenians from the southwest who brought their tools that were made out of material that was sourced in southern France, initially. As pioneers they didn't know where to find all the resources they needed to make those tools. They were all about portability and adaptability in the face of an increasingly erratic climate. Once they got familiar with an area, they began to exploit more local resources, but in a style that they'd developed elsewhere... somewhere down to the south.

Look at: Antiquity / Volume 87 / Issue 336 / June 2013, pp 384 - 404

Nobody knows what the maternal or paternal lines are of these folks were, yet, though Bichon could be the smoking gun. The Magdalanians had their act together enough to survive and evidently were an expanding population post LGM, likely more numerous than anyone, if there was anyone to the north and or east. Why is it such a stretch to think that I2 was among them, expanding out of the Franco-Cantabrian refuge?

Gravetto-Danubian
01-30-2016, 02:42 AM
The Swiss Plateau was recolonized by Magdalenians from the southwest who brought their tools that were made out of material that was sourced in southern France, initially. As pioneers they didn't know where to find all the resources they needed to make those tools. They were all about portability and adaptability in the face of an increasingly erratic climate. Once they got familiar with an area, they began to exploit more local resources, but in a style that they'd developed elsewhere... somewhere down to the south.

Look at: Antiquity / Volume 87 / Issue 336 / June 2013, pp 384 - 404

Nobody knows what the maternal or paternal lines are of these folks were, yet, though Bichon could be the smoking gun. The Magdalanians had their act together enough to survive and evidently were an expanding population post LGM, likely more numerous than anyone, if there was anyone to the north and or east. Why is it such a stretch to think that I2 was among them, expanding out of the Franco-Cantabrian refuge?

Your story keeps changing ! ;)
You do realize that Franco-Cantabria isn't northwest Europe , right ? Is part of Southern Europe

bix
01-30-2016, 03:16 AM
Of course, around the Pyrenees, from the Gulf of Lion to the Bay of Biscay.

Gravetto-Danubian
01-30-2016, 03:33 AM
Of course, around the Pyrenees, from the Gulf of Lion to the Bay of Biscay.

That's my theory, and I'm sticking to it.

The other guy thinks Doggerland or Russia or something.

I know, it was under a mile of ice--rather a tough slog, not conducive to survival methinks.

Oh so sorry. I confused you with Epp. ! (must be the 3 letter name)
I agree

:amen:

But I suspect other refugia also played a role, albeit lesser.

bix
01-30-2016, 03:46 AM
No doubt they did.

I remember reading somewhere that eastern Magdalenian probably involved Balkan refuge, which would make sense.

In here...

http://link.springer.com/chapter/10.1007/978-94-017-7206-8_8

Time will tell.

Cheers.

Megalophias
01-30-2016, 06:31 AM
So here is my approximate average of yfull SNP and FTDNA STR calculations. Both sets of data give a very similar estimates of geographical flows. The dating is different in magnitude between the two (the SNP analysis gives older results), but they are proportionately similar.

Initial development - formation of pre-I1, I2, I2a2, pre-I2a1b and I2a1a around the time of the first significant glacial melting (between 15,500 and 10,000 BC) - all North Sea and the Rhine, as far South as Switzerland.

First significant migrations away from this region - I2a1a1 from Northern & Eastern France towards Iberia 4,000 BC ; I2a1b2 from Germany towards Poland 2,700 BC; I1a1 from Sweden towards Finland 1,230 BC.

If all modern I is from Northwestern Europe over the last 6000 years or so, what the heck Y haplogroups did people in the rest of Europe have? And what happened to them?

Some ancient I DNA to help out with the timeline (not exhaustive):

Bichon - northern Switzerland - 11 500 BC - Azilian (post-Magdalenian) - I2a. Per genetiker pre-I2a1a2a1-L1287 (IMO very high confidence).
Menteşe - northwestern Anatolia - 6300 BC - Early Neolithic - I2c-L596.
Loschbour - Luxembourg - 6100 BC - Rhine-Meuse-Schelde? - I2a1b*.
Motala - interior southern Sweden - 5700 BC - Maglemosian?? - 1 x I2c, 2 x (pre?-)I2a1b, 1 x I2a1-P37 (pre-L1287 per genetiker, fairly high confidence), 1 x I2a1a1a-L672.
KO1 - Hungary - 5700 BC - Körös - I2a. Again per genetiker pre-I2a1a2a1-L1287 (again IMO very high confidence).
BAB5 - Hungary - ~5-6000 BC - Transdanubian LBK - I1-M253 (but presumably pre-I1, don't know how divergent).
Stora Förvar 11 - Gotland - 5500 BC - Kongemose?? - pre-I1 per genetiker (high confidence, very old split).
Els Trocs - Spanish Pyrenees - 5300 BC - Epicardial - I2a1b1-L161.
NE7 - Hungary - 4400 BC - Lengyel - I2a - same as KO1 per genetiker.
La Mina - northern Spain - 3700 BC - Megalithic - I2a2a1b2-Z161.
Remedello - northern Italy - 3300 BC - Copper Age - 1 x I2a1a1-PF3983 per Tagankin; 1 x I2a1a1, 2 x I2a1a1a-L672 per genetiker.
Salzmünde/Bernburg - central Germany - 3200 BC - Funnelbeaker - I2a1b1a-L1498.
Atapuerca/Burgos - northern Spain - 2700-2900 BC - Copper Age - I2a1a1 x 1, I2a2 x 1, I2a2a x 1, I2a2a1 x 1, I2a2a2 x 1.
Lanycsok - Hungary - 2700 BC - Copper Age - I2a2a-M223.
RISE552 - European Russia - 2900-2100 BC - Catacomb - I2a2a1b1b-L699.
Ajvide - Gotland - 2800-2000 BC - Pitted Ware - pre-L1287 x 2, I2a1b1a1-S2703 x 1.
Unetice - Germany - 2200-1600 BC - Early Bronze Age - I2, I2a2, I2c2.
Vatya - Hungary - 2000 BC - Early Bronze Age - I2a2a1a2a-L1229, 1 x I2a1 per Tagankin.
Angmollan - Sweden - 1400 BC - Nordic Bronze Age - I1-M253 (first known modern I1).
Lichtenstein Cave - Germany - 1000 BC - Urnfield - 5 x I2a2b (probably related).

Unfortunately there is no ancient Y DNA from southern Europe proper and only one Paleolithic sample. There are not enough samples to provide any negative evidence, but we have positive evidence of presence at least.

Pre-I1 is found in both Hungary and Sweden at about the same time. I2c is attested in Turkey even before it is attested in Northern Europe. I2a2a does not show up until the Middle Neolithic in Spain, then we find it across Copper Age Europe; I2a2b not until well into the Bronze Age. I2a1a1-M26, associated with Spain and Sardinia, turns up first in Sweden, while I2a1b1-L161, associated with Britain and Ireland, is first found in Spain. L161 also has some connection with Funnelbeaker (<Ertebolle?).

Unless genetiker's calls are totally wrong (which they have never been before) pre-I2a1a2a1-L1287 was all over the place from an early period.

epp
01-30-2016, 11:29 AM
The Swiss Plateau was recolonized by Magdalenians from the southwest who brought their tools that were made out of material that was sourced in southern France, initially. As pioneers they didn't know where to find all the resources they needed to make those tools. They were all about portability and adaptability in the face of an increasingly erratic climate. Once they got familiar with an area, they began to exploit more local resources, but in a style that they'd developed elsewhere... somewhere down to the south.

Look at: Antiquity / Volume 87 / Issue 336 / June 2013, pp 384 - 404

Nobody knows what the maternal or paternal lines are of these folks were, yet, though Bichon could be the smoking gun. The Magdalanians had their act together enough to survive and evidently were an expanding population post LGM, likely more numerous than anyone, if there was anyone to the north and or east. Why is it such a stretch to think that I2 was among them, expanding out of the Franco-Cantabrian refuge?
I guess I2 was among them, although not necessarily in Franco-Cantabria. The evidence that I've seen to date (yfull, FTDNA & Bichon) points to a more Northerly presence for them (Switzerland & the Rhine).

epp
01-30-2016, 12:09 PM
If all modern I is from Northwestern Europe over the last 6000 years or so, what the heck Y haplogroups did people in the rest of Europe have? And what happened to them?
DNA seems to evidence that most lineages completely died out. Current lineages largely descend from a few individuals who managed to survive bottlenecks. My guess is that the y-DNAs of other European populations were now-extinct forms of I or IJ or K or DE, or possibly LT or A. There's no point in speculating unless there is some indicative evidence that is consistent with other reliable evidence.

Your ancient I DNA is a good summary. The Bichon sample is clearly the most useful, and is consistent with estimates from modern DNA based on yfull and FTDNA data - indicating early presence North West of the Alps. Genetiker's call of pre-I2a1a2a1 (i.e. I2a1a2a*) seems consistent with modern DNA, as this is one of the earliest predicted I2 subclades, although I don't know the basis for this conclusion.


Unfortunately there is no ancient Y DNA from southern Europe proper and only one Paleolithic sample.
Unless this evidence or any indicative modern DNA evidence emerges, I would exclude it from consideration.

Megalophias
01-30-2016, 06:35 PM
DNA seems to evidence that most lineages completely died out. Current lineages largely descend from a few individuals who managed to survive bottlenecks. My guess is that the y-DNAs of other European populations were now-extinct forms of I or IJ or K or DE, or possibly LT or A. There's no point in speculating unless there is some indicative evidence that is consistent with other reliable evidence.

Suppose that a few northern European lineages survived by founder effects while the southern ones went completely extinct, as you suggest. Scenario 1: I-M170 originated from pre-I in the north, and later some lineages expanded southward; all southern lineages (which might include pre-I/IJ* cousins of I) went extinct. Scenario 2: I-M170 originated from pre-I in the south, descendants populated the north, later some lineages of northern I re-expanded southward; all southern lineages (which would include various pre-I1 and I2*) went extinct.

The only direct evidence that would distinguish between these scenarios is what the southern lineages were. But if there are neither surviving southern lineages nor ancient DNA from there, then there is no way to directly distinguish them. You cannot argue from present-day diversity that the origin should be in the north, because *all* original southern diversity has been erased. Nor can you appeal to parsimony, because obviously there *were* many men in the southern parts of Europe and they *did* expand northward after the LGM.

So, even granting that all surviving I comes from recent northern expansions (certainly a good deal does), I see no reason to lower the TMRCA of I to fit a post-glacial date.


Genetiker's call of pre-I2a1a2a1 (i.e. I2a1a2a*) seems consistent with modern DNA, as this is one of the earliest predicted I2 subclades, although I don't know the basis for this conclusion.
He checks the DNA files made public by the study authors for a large number of equivalent SNPs. There are a small number of false positives and negatives but for good quality genomes the results are very consistent. He publishes all the Y SNP calls on his blog, you can just Google for it. (Be warned he is kind of crazy but his SNP calls are fine.)

epp
01-31-2016, 12:21 AM
Suppose that a few northern European lineages survived by founder effects while the southern ones went completely extinct, as you suggest. Scenario 1: I-M170 originated from pre-I in the north, and later some lineages expanded southward; all southern lineages (which might include pre-I/IJ* cousins of I) went extinct. Scenario 2: I-M170 originated from pre-I in the south, descendants populated the north, later some lineages of northern I re-expanded southward; all southern lineages (which would include various pre-I1 and I2*) went extinct.

The only direct evidence that would distinguish between these scenarios is what the southern lineages were. But if there are neither surviving southern lineages nor ancient DNA from there, then there is no way to directly distinguish them. You cannot argue from present-day diversity that the origin should be in the north, because *all* original southern diversity has been erased. Nor can you appeal to parsimony, because obviously there *were* many men in the southern parts of Europe and they *did* expand northward after the LGM.

So, even granting that all surviving I comes from recent northern expansions (certainly a good deal does), I see no reason to lower the TMRCA of I to fit a post-glacial date.
I agree that the TMRCA of I could well be pre-LGM, although my interest here is not in I per se, but surviving I (which the modern DNA evidence appears to indicate is North West European post-glacial). Scenario 2 is the essentially the same as scenario 1, apart from the fact that it looks back to an earlier point. If I were to look back further than I do, the indications for the closest surviving relative of surviving I would be, in my view, the Middle East (from where J appeared to arise), rather than Cantabria. There appears no evidence to suggest that I itself either arose or didn't arise in Cantabria (apart from the fact that Spain is not located directly between the Middle East and North West Europe), so it does not come into my equation.


He checks the DNA files made public by the study authors for a large number of equivalent SNPs. There are a small number of false positives and negatives but for good quality genomes the results are very consistent. He publishes all the Y SNP calls on his blog, you can just Google for it. (Be warned he is kind of crazy but his SNP calls are fine.)
My read of these calls is that Bichon is probably I2a1*, rather than pre-I2a1a2a1*, which would set it back much earlier in I's development and put it into the earliest populated subclade yet discovered in haplogroup I. As the second earliest such sample is I2a1b* from Luxembourg, these represent further indication of the earliest surviving I lineages developing in and around the Rhine basin, presumably in the post-LGM era.
By the way, I note that Genetiker seem to be of the view that the Cantabrian refuge was populated by R1b, rather than I, and that North West European I's ancestor was Balkan.

Gravetto-Danubian
01-31-2016, 12:40 AM
Epp
Which ftdna group do you manage; and which dataset are your analyses based on ?
(which subclades, which countries )

epp
01-31-2016, 11:41 AM
Scenario 1: I-M170 originated from pre-I in the north, and later some lineages expanded southward; all southern lineages (which might include pre-I/IJ* cousins of I) went extinct. Scenario 2: I-M170 originated from pre-I in the south, descendants populated the north, later some lineages of northern I re-expanded southward; all southern lineages (which would include various pre-I1 and I2*) went extinct.

The only direct evidence that would distinguish between these scenarios is what the southern lineages were.
I suppose the only indication tipping it in favour of scenario 1 is that it is a less convoluted route - a two-stage general migration - a move of one section of IJ from the South East (where it shared most recent ancestry with J), followed by a move of one subclade of this section I2a1a1 to Iberia. Whereas scenario 2 would have required a four-stage migration - a move from the South East into the heart of Europe, followed by a move South West to Iberia, followed by a move back past the Alps, followed by a move of one subclade back to Iberia again (possible, but I would say less likely/plausible).

Megalophias
01-31-2016, 04:58 PM
My read of these calls is that Bichon is probably I2a1*, rather than pre-I2a1a2a1*, which would set it back much earlier in I's development and put it into the earliest populated subclade yet discovered in haplogroup I.
How do you come to that conclusion? He no negative calls for P37 or CTS595 equivalent SNPs and has about 35/70 positive calls for L1287 equivalent SNPs. There is no way that number of calls can be false positives, either it's pre-L1287 or the whole thing is completely spurious.


I suppose the only indication tipping it in favour of scenario 1 is that it is a less convoluted route - a two-stage general migration - a move of one section of IJ from the South East (where it shared most recent ancestry with J), followed by a move of one subclade of this section I2a1a1 to Iberia. Whereas scenario 2 would have required a four-stage migration - a move from the South East into the heart of Europe, followed by a move South West to Iberia, followed by a move back past the Alps, followed by a move of one subclade back to Iberia again (possible, but I would say less likely/plausible).

But people did move into Europe from the SE, at least once and likely several times. They did spread into SW Europe, at least once as the Aurignacian; after that the Gravettian technocomplex moved into the SW, which at least required human contact if not major migration, and there is evidence of people moving into the SW during the height of the LGM as well. Then after the coldest period people did spread out from the SW again (and from the SE too for that matter). This is quite minimalist, people moved long distances hunting over large areas and there were all sorts of changes in material culture which might be associated with the movement of people.

We don't need to invent migrations to fit the theory, we already have plenty of migrations. And in your version the same migrations still occurred, the only difference is when and where the ancestor of I lived. So there is nothing to be gained by invoking the principle of least moves. We can only speculate, really.

Gravetto-Danubian
01-31-2016, 08:36 PM
Yes it will be very good when we get some more ancient Sample's from the paleaolithic.
At the moment we can only make inferential hypotheses, some of which are a little half- baked.

epp
01-31-2016, 10:40 PM
How do you come to that conclusion? He no negative calls for P37 or CTS595 equivalent SNPs and has about 35/70 positive calls for L1287 equivalent SNPs. There is no way that number of calls can be false positives, either it's pre-L1287 or the whole thing is completely spurious.
I came to the conclusion he was probably I2a1, precisely because he does not have a negative call for I2a1 P37. He has no negative calls for CTS595, because he has no calls for CTS595 at all (either positive or negative). And he has a negative call for L1287, so does not appear to be I2a1a2a1, although does appear to be positive for its subclade I2a1a2a1a A417! Of course, as you say, the whole thing could be completely spurious.

Given this data, I am curious why the researcher himself would have only claimed I2a per se to be his "best guess" - and wasn't even convinced enough to call it a "conservative guess". Does anyone have a link to exactly where the researcher published all this data that Genetiker has used?


But people did move into Europe from the SE, at least once and likely several times. They did spread into SW Europe, at least once as the Aurignacian; after that the Gravettian technocomplex moved into the SW, which at least required human contact if not major migration, and there is evidence of people moving into the SW during the height of the LGM as well. Then after the coldest period people did spread out from the SW again (and from the SE too for that matter). This is quite minimalist, people moved long distances hunting over large areas and there were all sorts of changes in material culture which might be associated with the movement of people.

We don't need to invent migrations to fit the theory, we already have plenty of migrations. And in your version the same migrations still occurred, the only difference is when and where the ancestor of I lived. So there is nothing to be gained by invoking the principle of least moves. We can only speculate, really.
You're probably right, although they were not necessarily the same people and their descendants who made all these moves, and if they were, they were not necessarily haplogroup I. There is evidence, for instance, to indicate that haplogroup K moved massive distances, but that haplogroup I was relatively static.

Even according to yfull estimates, which FTDNA data suggests might be substantially overstating TMRCAs, neither I1 nor I2 had arisen by the time of the last glacial maximum. So if people did move into Southern refuges at the LGM, only one of these people at most would have been a bearer of a surviving haplogroup I lineage. The question is - if this person did move that early (and not later, as indicated by FTDNA data) did he move South East or South West? As we're talking about one person, it could not have been both. So what is the most likely answer? Based on an average of yfull and FTDNA data analysis, I would estimate the likelihood is that he moved later - and as this would then have been at a time of continental warming, in the opposite direction. From which direction? I would say most likely South East, as that is the area where the descendants of his closest relatives are found.