World Families Forums - so what does the new study mean for L21?

Welcome, Guest. Please login or register.
September 20, 2014, 09:14:42 PM
Home Help Search Login Register

+  World Families Forums
|-+  General Forums - Note: You must Be Logged In to post. Anyone can browse.
| |-+  R1b General (Moderator: rms2)
| | |-+  so what does the new study mean for L21?
« previous next »
Pages: [1] 2 3 ... 6 Go Down Print
Author Topic: so what does the new study mean for L21?  (Read 13072 times)
alan trowel hands.
Guru
*****
Offline Offline

Posts: 2012


« on: August 27, 2010, 07:02:58 PM »

L21 as a percentage of total male population in new study:

54.2% Ireland South
50.0% Ireland West
40.9% Ireland Southwest
33.3% Ireland North
31.3% Ireland East
25.0% England North
24.0% England Southeast
20.8% England Southwest
Northern France-missing but I think would fit in around here at 20odd%
19.4% France Alpes de Haute Provence
14.3% France West
09.5% Denmark North
09.1% Denmark Southeast
08.2% France Vaucluse (upstream Rhone)
08.0% France East
08.0% England Central
07.9% France South
07.4% Switzerland Northwest
06.3% France Bouches du Rhone (at mouth)
05.8% Sweden South (Malmö)
05.7% Netherlands
05.6% Austria
05.3% Spain Cantabria, Santander
05.3% Denmark West
03.1% Germany North
03.1% Switzerland Northeast
03.0% Portugal, Lisbon
02.9% France, Var (coastal, East of Rhone)
02.4% Spain, Castille and Leon, Leon
02.4% Poland
02.2% Germany South
02.1% Germany East
02.0% Switzerland (Lower Rhone Valley)
01.9% Croatia mainland
01.1% Czech Republic
01.1% Poland Southwest(Wroclaw)
01.0% Germany West
00.9% Spain, Valencia
00.9% Hungary
00.8% Andalusia, Sevilla
00.8% Russia Central
00.8% Italy North
00.6% Romania
00.4% Russia South
00.4% Slovakia



My first observation is that although disappointing in some ways it does not hugely contradict project findings.  it confirms that L21 is essentially an isles and French clade in terms of a substantial input.  However, it is interesting that although clearly its peak, L21 is nowhere near as massive in the isles, including Ireland (where I think people tend to think its anout 90% of the population) as has become commonplace to assume.

The only major contradiction with the project map is the low count in SW Germany.  There is without doubt a hotspot in the Rhineland that the study sampling has missed.  The projects indicate a lot of L21, U152 and U106 as well as some others in that area where if I recall correctly R1b1b2 runs about 50%.  So I am would be really baffled if at least in the Rhineland province and adjacent L21 is not over 10% of the population and more in line with parts of France.

Otherwise, the results confirm the project findings of very low numbers in Italy, Iberia (prior to recent chasing by matching), central and eastern Europe. The results confirm that the relatively small amount of L21 on the project map east of the Rhineland is real and not down to migration patterns. 

If we just look at France

20-25%?? Northern France
19.4% France Alpes de Haute Provence
14.3% France West
08.2% France Vaucluse (upstream Rhone)
08.0% France East
07.9% France South
06.3% France Bouches du Rhone (at mouth)
02.9% France, Var (coastal, East of Rhone)

Firstly despite somewhat dissapointing results I would still think looking at the main west, east, south and my guess at the missing north that over 10% of the entire male population of France is L21.  Although substantially less than some guesses (at one stage I thought it might be nearer 20%) it is still very significant and surely only the mad would try to explain it as a movement out of the isles (which variance does not support).

What I think this probably shows is something I think was emerging on the project map this year- that in France as well as an increase to the north, there was also a major east-west component with L21 stronger in the west.  There is a very significant drop off in the south and east in general.  I suspect 'France north and west' combined might come in about 18% while 'France south and east' combined might come in about 8%, nearly half.  The actual variation of L21 as a percentage of R1b1b2 may however be less as R1b1b2 also drops off from west to east I think but I think the trend is still there.   

There are any number of reasons why that might be. Dilution is one possibility.  The north to the west of Paris and the west in general did not get the same sort of substantial post-Roman Germanic influx as the east (Burgundy, Alsace, Lorraine, the extreme NE etc which actually brought about partial or temporary linguistic shifts/the renaming of the areas after Germanic peoples.  In the British Isles L21 is also stronger in the areas which missed out on the Germanic influx.  I wonder if the higher L21 in the north and west of France is echoing the same phenomenon as in the British Isles. 

One other thing about L21 is that the high Alpes de Haute Provence result from a very mountanous part of the extreme SE of France bucks this trend for L21 to be low in the south and east.  This study raises the possibility that French L21 tends to survive in retreat areas in pockets in the west and on difficult land elsewhere.  The presence of a pocket in a very mountainous area in the SE with levels of L21 as high as in the north and west of France does also hint that perhaps L21 was once higher in the east than it is now. There are other possibilities.   

With the surprisingly low rate in central Europe also comes a better than expected showing in northern Europe from Denmark.  This had not been uncovered by the project although the project had started to show a good showing in Norway for L21 (not covered in this study).  I suspect now that a prevalence map of L21 on the continent would have an area of raised strength running from western France, around NW France and (with a break in the U106-rich low countries) along the North Sea coasts of Denmark and Norway.  There would then be a band of lesser but not negligible presence in the rest of France, NE Spain, parts of Switzerland and as far as Austria (surely in SW Germany too as the stat seems to be a fluke).  Now that picture is broadly compatible and comparable with the project maps even if the overall scale of L21 is lower than expected. 

All in all I think there is no doubt that in continental terms L21 is significant in much of France and adjacent areas and also along the North Sea coast. However it does look rare in cenral Europe (south Germany eastwards).  I think it is clear from the relative strengths that L21 in Europe in France as in the isles in pre-Germanic.  Its hard to say what its relationship is to U152.  In later times L21 must have been reasonably common (though not as much as I thought) among the Gauls of France although perhaps the lack of it in Italy is evidence that even in the last centuries BC it was (as now) not common in the east and SE of Gaul (the more central European areas).  There is no reason to believe S116 clades were similar throughout Gaul and indeed the proportion of clades of the Celts as a whole is very different in Iberia, France, Britain, Ireland etc so any talk of a single Gaulish clade is probably as silly as the idea of a single wider Celtic clade.  The only constant seems to be a high level of S116 clades. Even the question of the origin of the Celts is no longer safely tied to central Europe and some now favour the Atlantic fringes.  Other than belonging to S116 in general there is no single common denomenator clade among the Celts.   
Logged
rms2
Board Moderator
Guru
*****
Offline Offline

Posts: 5023


« Reply #1 on: August 27, 2010, 07:45:34 PM »

I would like to know where they got their samples.

Do we know the sample sizes from each region from Table S4?

I have a hard time believing the low frequencies for Germany West. How did FTDNA just stumble on so many L21+ western Germans, if L21 is only 1% in western Germany?

Incredible.

For me this new study is tremendously disappointing. The only consolations are that L21 variance is higher on the Continent than in the Isles and that Myres et al did not get a representative sample, so higher L21 figures might still be lurking in untested places (and there were plenty of them).

Unfortunately, a shot at a better assessment might not come for years - if ever - and that makes this study even more heartbreaking.

I just wish the University of Santiago de Compostella would test its French samples for P312 and L21 and publish the results. As I recall, their sampling was much more representative and evenly distributed.

This Myres et al report has really cut into my enthusiasm, to be quite honest.

« Last Edit: August 27, 2010, 07:46:41 PM by rms2 » Logged

rms2
Board Moderator
Guru
*****
Offline Offline

Posts: 5023


« Reply #2 on: August 27, 2010, 08:11:04 PM »

A subtitle for the Myres et al report could be, "Yes, U152 and U106 are even bigger deals than we thought they were - hip, hip, hooray!".

They didn't even bother with SRY2627 and M153.
« Last Edit: August 27, 2010, 08:16:43 PM by rms2 » Logged

jerome72
Senior Member
***
Offline Offline

Posts: 64


« Reply #3 on: August 27, 2010, 11:21:47 PM »

I did not have this study under the eyes, I read with interest all your comments.

This is a confirmation (for me), that in France, it is in the south-eastern than R1b1b2 is the most important.
Maybe, Many think it is in the northwest, in fact, a high proportion of I (I1 and I2) of about 25%, resulting probably R1b1b2 is only about 50%.

I think the proportion of L21 compared to the rest of R1b1b2 must be very important in the north-west and much lower in the south-east
The Alps of Haute-Provence, are a little away of path of communication, as rhodannien corridor and the Mediterranean Sea.
This may reinforce the idea that L21, was one of the first groups R1b1b2 has arrived (unless it is original) in Western Europe (Essentially France and British Islands).

The following successive migrations from the east, would be, dilute L21 or L21 has been pushing to the Atlantic coast to the west-central France to the islands and some pockets of L21 in the east, isolated of lines of communication, have survived until today.

I find, on the contrary, the fact that L21 is not as great as expected in Europe can help us better understand who they are.
Unfortunately, this study does not seem enough to study the geographical areas of Europe where L21 is probably the most present.
« Last Edit: August 27, 2010, 11:29:40 PM by jerome72 » Logged
alan trowel hands.
Guru
*****
Offline Offline

Posts: 2012


« Reply #4 on: August 28, 2010, 06:33:13 AM »

I think too what we have to remember is if R1b1b2 is only around half of the French population and France has a lot of different R1b1b2 clades including high amounts of U152, L21, S116*, the Iberian clades, some U106  etc then that 50% is very divided.  So 10% of the country and areas between with 10 and 20% L21 is actually highly significant chunk of the R1b1b2 population. 

I also have to agree with Rich about SW Germany, especially the Rhineland.  The sampling is maybe too broadbrush to pick up localised concentrations.  It is noticeable that the much more intensive testing of SE France showed a lot of variation including an L21 hotspot.  However, this was the exception and there is no doubt that their testing somehow missed other hotspots like the Rhineland.  I have always thought of the Rhineland as the best self-tested (i.e. hobby tested) area on the continent and perhaps the one area where there hobby testing was so strong that we should have had a very good idea of the y-DNA there.  There is absolutely no reason why the hobby testing would favour L21.  Migration patterns to America favour certain areas (the Rhineland being one) but they should not favour certain clades. 

One other thing, although the numbers are low, even in Iberia I think the projects identification of small hotspots in the NE and Portugal is backed up by the Iberian results. 

I know Rich is feeling a bit down about it but I think if we just get over the fact that the scale is smaller I think by and large the patterns identified by the project maps have been vindicated.  We had wondered if there was distortion down to migration patterns to America but the patterns have turned out to be real.  The project showed most L21 in north and to some degree west France, a modest amount in the SE of France and NW Switzerland, minor amounts in the Low Countries, almost none in Italy and central Europe beyond the Rhine/Main, very little non-Jewish in eastern Europe.  I also think the project for a long time showed very little in Iberia.  However, there was from an early stage a hint of a small group in the NE and the west which Rich recently expanded through matching and identifying clusters.  Although the Iberian L21 is small, the results of this paper are  in line with the projects identification that most L21 in Iberia is in the extreme NE and around Portugal (the paper has the highest levels where it tested in Cantabria and Portugal).  Other projects too indicated things that have since been confirmed by this study such as the superconcentration of U152 in Switzerland, Italy, south Germany etc, the concentration of U106 in Holland etc.  So, I think the projects have been a success and the worries about distortion have not been vindicated.   

I think it is remarkable that not only has the project identified the European pattern well but it has even been able to accurately pick up the modest peas in a low L21 area like Iberia.  I think this all vindicates that project work/hobbiest testing is very useful at reconstructing.  So I think there is a lot to be proud of in the hobbiest and project efforts.

So, the patterns identified by the project are surprisingly accurate even down to small hotspots.   What we have to get used to is

1. The scale is smaller everywhere, probably half of what we thought (and that includes much of Ireland).  I think we just need to get used to the fact it has turned out we are a more minor clade than we thought and that our distribution really is pretty well that indicated on the project map.   

2. The blank or poorly represented areas on the project maps are real - there really is little L21 in those areas.

One caveat to the last statement is the possibility that L21 is patchy with elevated pockets some of which may be missed .  Some small pockets of modest L21 rise (such as the two in Iberia) were picked up by both the project and the study while the very strong probably very localised pocket in the SE Alps of France was detected in this new study but not the project.  Most weird was  the Rhineland concentration that the project emphatically found but this study has somehow missed.  That is probably the single most worrying thing about this studies sampling.  There is simply no doubt that that is a fluke of sampling.   I think the fact that Austria has a modest but significantlly higher result than SW Germany suggests this is wrong and that there is at least a modest trail of L21 along the Upper Danube into the central European part of the old Celtic world. 

Logged
alan trowel hands.
Guru
*****
Offline Offline

Posts: 2012


« Reply #5 on: August 28, 2010, 08:04:07 AM »

A few other thoughts regarding the origins of L21 or any clade.  Variance indicates the origin of L21 is in France and I believe from what I hear that the odd individuality of many French L21 rather than belonging to late clusters means that more variance could yet be discovered.  However, variance essentially is rooted in the point of initial expansion.  Where an SNP first occurs as a unique individual and soon after  small minority among a group who settle down in an area, it will not have the same opportunities for rampant expansion as its descendants who head on into virgin land.  So, there is every chance that an SNP that had just occurred could easily daughter out (or die out) in its first location of occurance.  Even in areas where a small 2nd  generation tree of immediate male descendant lines  have been produced many of the lines could die out, meaning the MRCA is later than the real founder date of L21 in the locality (and therefore leave less variance).  Basically, even in an LBK style scenario of huge expansion opportunities and rapid geographical expansion, it is likely that the origin point/the 1st generation will be missing or at least not be obvious from variance. 

The model that the new paper uses is one I totally agree with in a Neolithic farmer scenario - each generation surfing the head of the wave of migration.  That model is based on a family moving forward into an area, settling down and then some of their offsping moving on back on trail trail west while others stay put.  That clearly will tend to mean SNPs expand and have their variance dictated by the generations after the SNP occurred who have the advantage of moving into virgin territory and unchecked expansion that the first generation (and perhaps the 2nd) would not.  In a period like LBK of fairly rapid expansion west.  There is every chance in the location where the first guy with the SNP was a singleton among a load of S116* and other guys that his line may have died out or at least many of his descendant lines would have.  Basically you need constant expansion for a while in the early stages for variance to work.  I believe that in an LBK scenario of huge expansion opportunities on the trail west that variance works but I do think it will tend to lose its earliest end of its trail or at least the earliest part of the SNPs trail where opportunities were limited may have less variance than expected. 

For that reason although L21 has its highest variance in France I would still not rule out the possibility that L21 first occurred somewhere else somewhat (although perhaps not spectacularly) to the east of its highest variance area today.   L21 obviously must have occurred among S116* people of slightly (but not much more) variance.  Clearly L21 has to have happened close to and soon after S116 itself occurred judging from most calculations.  The study suggests SE France but I personally think in an LBK scenario it would more likely have been a little further east in the Upper Danube area. 

The study also suggests the alternative possibility that during the LBK period perhaps it arrived at and crossed the Rhine as P310* and that S116 and downstream only occurred during secondary middle Neolithic expansions in western Europe in subsequent generations, possibly after a hiatus in expansion and the end of the unified LBK culture.  There could be an element of both IMO.  I quite like the idea that a lot of the expansion of the downstream S116 groups like U152 and L21 could have occurred in post-LBK times when after a hiatus these groups expanded in the form of successor cultures beyond the core LBK area.  In fact this is a necessary part of this model because LBK is not found in many areas where R1b1b2 or S116 is found today in high numbers.   This suggested middle Neolithic expansion of former LBK lineages was apparently mainly into areas where there pre-farming populations had held onto the land (nearly all of north facing coastal Europe) or where there were late local hunters with weak farming traits from Cardial influences (substantial parts of central and upland France, the Alpine areas, parts of north Italy, most of interior and north of Iberia).  These groups may have had the advantage of a far more developed farming culture than the partly Neolithiiced hunting groups they supplanted.   

In terms of L21 it may have especially done well in the post-LBK expansion to the French Atlantic coasts in the 5th millenium BC which in turn meant they were almost certain to have a big impact on the isles.   Southern England now has similar L21 similar to northern France and it is likely that it was a founder effect in the crossing from Britain (and perhaps direct from NW France) that meant that L21 became raised in Ireland and Scotland (which were closely linked).     

Note that this is just working through the LBK and post-LBK middle Neolithic scenario that the paper presents.  I am not saying that they are correct to chose this scenario although I do see advantages in it.  I am not ruling out a copper age/beaker model but the evidence for it is much much more subtle. 
Logged
rms2
Board Moderator
Guru
*****
Offline Offline

Posts: 5023


« Reply #6 on: August 28, 2010, 09:17:51 AM »

So, basically, it looks like L21 has been in the British Isles since the Neolithic Period. It's unlikely, then, that it is responsible for bringing Celtic languages to the Isles. (More great news courtesy of Myres et al.) Who are we to thank for that? U152?

Perhaps the greatest benefit of being L21+ for me, as an American with a British surname but no y-dna line paper trail out of North America, is confirmation that I am probably descended from a British, Irish or Scots-Irish immigrant.

So, in the end, the "Deep Ancestry" trail has proven very disappointing, involving a lot of apparently wasted time and effort and emotional investment. Probably the best thing to do now is to return to the old paper trail, making use of the genetic information simply to narrow the search.

« Last Edit: August 28, 2010, 09:19:00 AM by rms2 » Logged

Mike Walsh
Guru
*****
Offline Offline

Posts: 2964


WWW
« Reply #7 on: August 28, 2010, 10:18:02 AM »

.... I also have to agree with Rich about SW Germany, especially the Rhineland.  The sampling is maybe too broadbrush to pick up localised concentrations.  It is noticeable that the much more intensive testing of SE France showed a lot of variation including an L21 hotspot.  However, this was the exception and there is no doubt that their testing somehow missed other hotspots like the Rhineland.  I have always thought of the Rhineland as the best self-tested (i.e. hobby tested) area on the continent and perhaps the one area where there hobby testing was so strong that we should have had a very good idea of the y-DNA there.  There is absolutely no reason why the hobby testing would favour L21.  Migration patterns to America favour certain areas (the Rhineland being one) but they should not favour certain clades.  
...
In Table S4 here are R-L21*'s frequency percentages.
2.1% in Gemany East
3.1% in Germany North
2.2% in Germany South
1.0% in Germany West

In Table S2 there are is Count (N) = 6 of R-L21*'s in Germany.   I didn't see the total sample size for Germany so I was going to reverse calculate the total sample size but Table S2 and S4 don't totally agree. For example, M222 does not have any Germany listed in S2 but has 5.3% listed for it in S4.

I think there is a possibility of an error in S4 or at least an anomaly for M222 to be greater than L21* in Germany.  Am I missing something?

« Last Edit: August 28, 2010, 10:20:08 AM by Mikewww » Logged

R1b-L21>L513(DF1)>S6365>L705.2(&CTS11744,CTS6621)
Mike Walsh
Guru
*****
Offline Offline

Posts: 2964


WWW
« Reply #8 on: August 28, 2010, 11:09:56 AM »

We've got 924 R-L21* haplotypes of 25 or greater STRs that have Old World MDKA's.  

Below are the diversity numbers at a granular regional level.  I was more granular where we have high numbers of MDKA's.

R-L21* Relative Diversity & Count for 25 marker Hts:

106.8  19  France - North & Central
105.4  16  France - North Atlantic
103.2  04  France - Northeast
095.2  16  Germany - Middle
094.2  06  Germany - South
093.5  06  Alpine & Cisalpine (inc. N. Italy)
087.3  07  Aquitaine & Pyrenees
084.7  25  Iberian Peninsula (exc. Pyrenees)
067.0  06  Low Countries

082.1  16  Scandinavia - North Sea side & Straits
074.0  10  Scandinavia - Baltic Sea side

109.8  13  England - West Midlands
108.1  09  England - Yorkshire and Humbria
097.0  09  England - East
097.0  14  England - South East
096.6  24  England - South West
095.7  09  England - North West
093.8  12  England - East Midlands
092.6  19  England - London
085.9  52  Wales
064.3  04  England - North East

096.2  26  Scotland - East
084.4  38  Scotland - North
084.4  43  Scotland - Strathclyde
071.2  18  Scotland - South

099.0  72  Ireland - Munster
092.9  91  Ireland - Ulster
092.3  33  Ireland - Connacht
091.7  52  Ireland - Leinster

Surprised me a bit that Scotland has lower diversity than England.  Also that Wales was so low. West Midlands is very diverse.  Is it a mixing place or an origin place?  I think this is an important consideration.  Perhaps Wales and Scotland represent Old Britons whereas much of England represents a "melting pot" of a immigrant Gaulish, Belgae and some Scandinavian R-L21* folks.

Given our comparatively low and limited sampling of France, that still clearly looks like more of expansion point.  There is almost a "Paris-London" axis.

I don't think these numbers should be considered stand-alone.  I think, similar to what Vince V. does with the Ht35 project, the placement of brother and ancestor clades is important, probably more important.   In our hobbyist DNA projects files I get that the Alpine area has very high diversity for P312 overall as well as points east, but points east are so scattered it's hard to figure out.

Note: We have more R-L21* people in our various "hobbyist" DNA projects and we also have many more STR's cataloged in contrast to what the Myres study evaluated. On the other side, our DNA projects are American immigrant biased and dependent on individuals' MDKA info.
« Last Edit: August 28, 2010, 12:14:50 PM by Mikewww » Logged

R1b-L21>L513(DF1)>S6365>L705.2(&CTS11744,CTS6621)
alan trowel hands.
Guru
*****
Offline Offline

Posts: 2012


« Reply #9 on: August 28, 2010, 12:24:43 PM »

I really do not want to see L21 in terms of a head count, especially an L21 v U152 head count.  However, it also worth saying that while U152 is the bigger of the two in France it is not by an enormous amount:

.................. L21..........U152..........................S116*
W France......14%.........21.4%........................21.4%
S France.......7.9%........10.5%........................28.9%
E France........8%..........16%..........................24%
N France-not covered but I think L21 would eclipse U152 there

Now, while U152 is larger than L21, its hardly a landslide.  Its around the 3:2, 5:4 and 2:1 mark in U152's favour but L21's peak in the north has gone unlisted.  In fact that group of results is compatible with the Santiago de Compostella study which indicated a really major advantage for U152 in Alsace but not elsewhere.  I think U152 people are probably also feeling surprised at their unspectacular performance in France.  It actually looks like in France L21 might be about 12%ish of the population and U152 maybe 15%ish.  Not a huge difference.   So I do not think the results really give the people who want to claim U152 as THE Gaulish clade really have a very strong case.  In fact the real shock is that S116* (which include the 'Iberian' clades is dominant, probably about 25% of the French and similar to L21 and U152 combined.  To some degree single testing for L21, U152 etc when it was negative and not taken to resolution has hidden a lot of them because they cannot be defined as S116* and therefore do not end up on the S116* map.  Another reason may be that there are no  ways to an S116* definition by a stand alone test for an SNP.

There were apparently Celts who were predominantly L21 (Irish), Celts who were predominantly S116* (using the paper's terminology) (Iberia) and there were probably Celts who were predominantly U152 (Switzerland and adjacent) but the Celts of France seems to have been a varying mix of all the S116 clades.  I know DF tends to claim glory for U152 but please note that despite the fact the name La Tene comes from a place in Switzerland this is not where La Tene culture originated.  La Tene originated in and adjacent to NE France in the Mosselle area.  Switzerland is peripheral to La Tene development.  That may be an irony but its a fact (similarly the site that gave the name Hallstatt to the culture is located in Austria which is peripheral to the formation of that culture).  So, the U152=Switzerland=La Tene stuff is flawed.  The reality seems to be from the variances that the structure of R1b1b2 in Europe was set out long before La Tene or Hallstatt etc.


I really do not think numbers mean much in terms of deep ancestry research.  We have found that again and again.  That idea made people think R1b originated in Iberia and variance studies show that numbers are not the thing that matters.  Its variance, phyogeny etc.  However, to work that out a sample was needed and without the sample the project built the hobbiest would have had nothing to work with.  Its interesting that the paper came to very similar conclusions about L21 variance as the most recent hobbiest ones.  Not only is L21's variance slightly higher in France but S116* is much higher in France than anywhere, about 50% higher than England and over twice as high as in Ireland.  So, the obvious way to interpret that is that there is no way L21 could have arisen among S116* in Ireland or England. It looks from that data that the first major expansion of both S116 and then L21 must has their first serious expansions within France. The difference, judging from the  relative similarity of variance in L21 seems to be that L21 happened not too long before its first major expansion.  Hobbiests have worked out a lot of this stuff from project data already but it always lacked the backup of an academic paper.

Regarding low numbers of L21 in some areas, if as I have long thought, L21 or its S116* ancestors basically appeared as migrants passed into northern France from the east in what was essentially a south-east to north-west population movement then large numbers are not to be expected to the east or south of the area where L21 expanded.  If the LBK idea is correct (and who knows) the largest onwards expansion from northern France was into the isles.  I am not aware of any major movements out of northern France in prehistory that headed back into central Europe so I do not see how it could be expected to have a really substantial L21 presence in central or southern Europe.  There is little trace of the U152 Galations in Turkey we often heard suggested.    

I still think there are some weird anomolies that could indicate the sampling has distorted things.  One is the weird fact that Austria's R1b1b2 is about 25% L21 in the paper.  However at the same time it misses out the apparently strong showing of L21 in SW Germany.  That really is odd.  Some of Denmark is 20-25% of R1b1b2 is L21 in other parts of the same country its zero.  Then there is the strong pocket in the SE French Alps.  Are the proportions of R1b1b2 clades in Europe subject to all sorts of micro-patterns that we cannot see in most studies?  That leaves the possibility that a 'dotty' survey could provide some weird results if L21 is also 'dotty'.  
« Last Edit: August 28, 2010, 01:52:53 PM by alan trowel hands. » Logged
Mike Walsh
Guru
*****
Offline Offline

Posts: 2964


WWW
« Reply #10 on: August 28, 2010, 12:56:44 PM »

.... La Tene originated in and adjacent to NE France in the Mosselle area.  ......

I still think Hans is on to something.  I-L38 may be a cultural brother to R-L21.
http://tiny.cc/fb3zx
Quote from: Hans De Beule
Since the trendlines only are very rough indicators, the comparison of the trendlines of several clades only leads to hypotheses: - Three trendlines, I-L38 (orange), R-U152 (dark blue) and R-L21 (light blue), cross near contemporary Frankfurt, suggesting a common origin that could be linked to the La Tène culture that started to spread from the middle Rhine region.
Doesn't the Lichtenstein Cave have I-L38, R-U106 and R1a guys?  Could the I-L38 guy have joined an IE speaking culture?  The cave is just east of the major Celtic expansion origin points.  In the Bell Beaker discussion, didn't some eastern/northern Corded Ware (or some Central Europe people) and southern/western Beaker types exchange cultural practices/products?

I'm not a necessarily big proponent that L21 was spread by La Tene verus Halstatt or something else. However, L21 seems to have such a strong correlation with Celtic or potential pre-Celtic (possibly Beaker) speaking areas that it can hardly be explained any other way than a Celtic connection.

I  guess I'm saying I don't think R-L21 in Ireland learned to speak Celtic there.  I think they brought it with them from the continent.

That leaves the possibility that a 'dotty' survey could provide some weird results if L21 is also 'dotty'.
I could not find in my reading of the Myres study that they did a proper cross-sectional sampling by country, so that they could report a valid projected  overall frequency by country.  Maybe they did, but I just don't see it.
« Last Edit: August 28, 2010, 01:03:59 PM by Mikewww » Logged

R1b-L21>L513(DF1)>S6365>L705.2(&CTS11744,CTS6621)
rms2
Board Moderator
Guru
*****
Offline Offline

Posts: 5023


« Reply #11 on: August 28, 2010, 03:31:46 PM »

. . .
I could not find in my reading of the Myres study that they did a proper cross-sectional sampling by country, so that they could report a valid projected  overall frequency by country.  Maybe they did, but I just don't see it.

Yet they formulated maps showing the supposed distribution and frequency of the clades included in that report.
« Last Edit: August 28, 2010, 03:32:03 PM by rms2 » Logged

argiedude
Old Hand
****
Offline Offline

Posts: 146


« Reply #12 on: August 28, 2010, 03:35:39 PM »

West Germany was just 100 samples. It happens. The Cruciani study looked at 250 North Africans and found only 1 R1b1b2, a huge disappointment. But judging from a database of 1800 North African samples, North Africa's R1b1b2 is about 4%, so Cruciani's frequency is off by an order of magnitude from reality (0,4% versus 4%). It happens. It's statistics. Most of the results will reflect reality very well, some will be off significantly, and a few will be complete baloney. West Germany must have drawn the short straw in this particular study. The fact alone that south Germany had 2% L21 should tell you that the gradient between France and south Germany will result in west Germany having 5% or 10% L21.

By the way, we know of several haplotype clusters in L21 by now. How likely is it for an L21 to belong to one of these clusters?
Logged

y-dna: R1b L21
mtdna: U5
rms2
Board Moderator
Guru
*****
Offline Offline

Posts: 5023


« Reply #13 on: August 28, 2010, 03:45:45 PM »

I really do not want to see L21 in terms of a head count, especially an L21 v U152 head count.  However, it also worth saying that while U152 is the bigger of the two in France it is not by an enormous amount:

.................. L21..........U152..........................S116*
W France......14%.........21.4%........................21.4%
S France.......7.9%........10.5%........................28.9%
E France........8%..........16%..........................24%
N France-not covered but I think L21 would eclipse U152 there

. . .  In fact the real shock is that S116* (which include the 'Iberian' clades is dominant, probably about 25% of the French and similar to L21 and U152 combined.  To some degree single testing for L21, U152 etc when it was negative and not taken to resolution has hidden a lot of them because they cannot be defined as S116* and therefore do not end up on the S116* map.  Another reason may be that there are no  ways to an S116* definition by a stand alone test for an SNP.

. . .


I know you mentioned the "Iberian clades", but I think it is important to state that Myres et al did not test for SRY2627, and SRY2627 is pretty common in France. In Myres' testing regime, SRY2627 would be hidden in that S116* figure. So, the actual S116* frequencies are probably much lower.

Now the interesting thing is that University of Santiago de Compostella study of France did test for SRY2627 and found that in every region tested, except Alsace, R1b1b2* was the most frequent group. Since they didn't test for P312 or L21, that R1b1b2* would include both L21 and P312* (S116*).

It really is a shame those Spanish scientists didn't test for L21 and P312. That would have provided some balance to the Myres et al results, at least for France.
Logged

alan trowel hands.
Guru
*****
Offline Offline

Posts: 2012


« Reply #14 on: August 28, 2010, 04:52:19 PM »

West Germany was just 100 samples. It happens. The Cruciani study looked at 250 North Africans and found only 1 R1b1b2, a huge disappointment. But judging from a database of 1800 North African samples, North Africa's R1b1b2 is about 4%, so Cruciani's frequency is off by an order of magnitude from reality (0,4% versus 4%). It happens. It's statistics. Most of the results will reflect reality very well, some will be off significantly, and a few will be complete baloney. West Germany must have drawn the short straw in this particular study. The fact alone that south Germany had 2% L21 should tell you that the gradient between France and south Germany will result in west Germany having 5% or 10% L21.

By the way, we know of several haplotype clusters in L21 by now. How likely is it for an L21 to belong to one of these clusters?

Yes one the project maps you could see L21 had a lot of competition from U152 and U106 with other clades probably underrepresented.  When I did a tot up there was something of a three way split and the possibility of undertested clades too.  However, L21 was holding its own and I think when I counted L21 and U106 were neck in neck and U152 was actually less.  From that I thought L21 would be about a quarter of R1b1b2.  R1b1b2 stands at about 50% in south and west Germany so i would have thought at least 10% of y-DNA was L21.  However,  that wasa very rough guesstimate with loads of unknowns.  Now we know NW Switzerland and E France are around the 8% mark for L21 in the male population, maybe that guess was a little high but Austria has 5.6% (which is actually about about 20% of R1b1b2 there).  I must say I would have found a figure more around the 7% much more believable than the 1 or 2 % noted for L21 among males in south and west Germany.  I would be simply astonished if that is not some sort of quirk of choice of sampling spots. 

Its a shame because a realistic SW German result would have given it a sort of a smooth distribution gradient for continental L21 over the old Gaulsih area varying from a peak of 14% in western France (with a likely higher peak in northern France) with a smooth decline in all directions to about 5-6% in Austria and north-east Spain at the extremities of this 'block' and therefore a very good match for the project map. 

It is interesting too that the likely real frequency peak of L21 in northern France is not likely to be too far from the continental variance peak.  As we are aware the frequency and variance peaks do not always coincide.  So that is interesting.  I think someone posted that in detail the variance peak (if you  the out of Birmingham school of though :0)) was in northern (not NW) France whatever that implies (presumably the NE).  I suppose we do not really know where the frequency peak is although its normally assumed its in the NW.  So, it may not be a close correlation between the two peaks.  If you look at the geography of northern France it would make sense (following the paper's LBK model) if L21 happened in the north-east or north-centre and a westward movement from there ended up in a pile up due to the sea on all sides in NW France leaving no expansion options that did not involve getting into a boat (the area south of the Loire was already settled by Cardial elements)
Logged
Mike Walsh
Guru
*****
Offline Offline

Posts: 2964


WWW
« Reply #15 on: August 28, 2010, 09:39:27 PM »

. . .I could not find in my reading of the Myres study that they did a proper cross-sectional sampling by country, so that they could report a valid projected  overall frequency by country.  Maybe they did, but I just don't see it.
Yet they formulated maps showing the supposed distribution and frequency of the clades included in that report.
My wife use to work for Gallup Polls and I've seen also had some Marketing Research experience.  I just don't see how what they did is representative, but maybe I'm stupid.  Once I criticized a formal academic study and and just about had my head bit off because I needed to make such criticisms in the form of another paper or in the right kind of form.  So I won't say Myes was wrong, but I will say I sure don't understand the validity of the country level frequency numbers.
Logged

R1b-L21>L513(DF1)>S6365>L705.2(&CTS11744,CTS6621)
Jdean
Old Hand
****
Offline Offline

Posts: 678


« Reply #16 on: August 29, 2010, 05:25:15 AM »

I can't help wonder about the M222 distribution in this study.

It's most common in Ireland, obviously, helping to bring the frequency of L21 up to 94% in West Ireland, but it forms 18% of Northern England’s M269 trumping the South of Ireland. It also forms 7% of M269 in Malmö, though how useful a DNA study in a city could be for deep ancestry I'm not entirely sure.
Logged

Y-DNA R-DF49*
MtDNA J1c2e
Kit No. 117897
Ysearch 3BMC9

alan trowel hands.
Guru
*****
Offline Offline

Posts: 2012


« Reply #17 on: August 30, 2010, 09:43:30 AM »

Hmm.. according to the Santiago University study U152 is meant to outnumber S116* (negative for SRY 2627) and L21* combined in eastern France.  However in this new paper U152 is 16% in eastern France while L21 (8%) and S116* (16%) (including SYR 2627) comes to 24%.  In fact the maximum that true S116* can be if the Santiago study is correct is 8%, otherwise U152 would not outnumber L21 and S116.  That if taken literally could mean AT LEAST half of the 16% S116* in eastern France in the new paper is SYR 2627.  Otherwise U152 wouldnt be the biggest clade there as the Santiago study suggests.  I know this is not on any sort of firm ground (and I realise its very unlikely SYR 2627 is that evenly spread) but if the estimate that at least half of S116* is SRY 2627 is projected onto this new paper's stats this would split French S116 roughly as follows

.................. L21..........U152..........................S116*...........SYR 2627
W France......14%.........21.4%........................10.7%..............10.7%
S France.......7.9%........10.5%........................14.45%...........14.45%
E France........8%..........16%..........................12%...............12%
N France.......????????????????????????????????????????????????????



Something like that would be compatible with the Santiago paper's statements that U152 is bigger than the S116*/L21* clade in eastern France but the latter is bigger elsewhere.

Certainly I think the picture emerging is that France is likely  split into all the major S116 clades with none dominant and that is the main cause for the fairly low Percentages.  In fact France seems to be the only place where a fairly even split among all the main S116 clades with none dominant is the case.  Again along with variance consideration France keeps standing out as the likely source of the earliest expansion of all the main S116  clades.  Surrounding France in all directions it seems that founder effects skewed this even mix towards the dominance of one clade (Iberia, North Italy, Ireland etc).  
« Last Edit: August 30, 2010, 01:01:33 PM by alan trowel hands. » Logged
Mike Walsh
Guru
*****
Offline Offline

Posts: 2964


WWW
« Reply #18 on: August 30, 2010, 04:37:41 PM »

... Certainly I think the picture emerging is that France is likely  split into all the major S116 clades with none dominant and that is the main cause for the fairly low Percentages.  In fact France seems to be the only place where a fairly even split among all the main S116 clades with none dominant is the case.  Again along with variance consideration France keeps standing out as the likely source of the earliest expansion of all the main S116  clades.  Surrounding France in all directions it seems that founder effects skewed this even mix towards the dominance of one clade (Iberia, North Italy, Ireland etc).  
As Alan noted, I don't think we can infer too much about the origins from frequency %'s.  I just went through diversity and variance calculations with Ken N, I think it is wise not to infer too much from them either.  I'm not saying the stats are useless, particularly where they seem to all align, but they are just additional data points.

I think the most important thing is just the finding of any significant (more than a Wild Goose find) presence of a subclade, and then lining the subclades to see who is where.  In that regards, France is certainly a point of origin or a crossroads.  It appears SE France and possibly the Alpine region are critical for P312 subclades.  I think it should be notes that U106 has a presence in the Alpine area as well.  I do think U106 is slightly older than P312 so even if they came from the same place, U106's trail might be thicker back towards the origin.  Same goes for U152 and L21.  U152 might be older.
« Last Edit: September 01, 2010, 10:30:34 AM by Mikewww » Logged

R1b-L21>L513(DF1)>S6365>L705.2(&CTS11744,CTS6621)
alan trowel hands.
Guru
*****
Offline Offline

Posts: 2012


« Reply #19 on: August 30, 2010, 05:03:02 PM »

... Certainly I think the picture emerging is that France is likely  split into all the major S116 clades with none dominant and that is the main cause for the fairly low Percentages.  In fact France seems to be the only place where a fairly even split among all the main S116 clades with none dominant is the case.  Again along with variance consideration France keeps standing out as the likely source of the earliest expansion of all the main S116  clades.  Surrounding France in all directions it seems that founder effects skewed this even mix towards the dominance of one clade (Iberia, North Italy, Ireland etc).  
As Alan noted, I don't think we can infer too much about the origins from frequency %'s.  I just went through diversity and variance calculations with Ken N, I think it is wise not to infer too much from them either.  I'm not saying the stats are useless, particularly where they seem to all align, but they are just additional data points.

I think the most important thing is just the finding of any significant (more than a Wild Goose find) presence of a subclade, and then lining the subclades to see who is where.  In that regards, France is certainly a point of origin or a crossroads.  It appears SE France and possibly the Alpine region are critical for P312 subclades.  I think it should be notes that U106 has a presence in the Alpin area as well.  I do think U106 is slighltly older than P312 so even if they came from the same place, U106's trail might be thicker back towards the origin.  Same goes for U152 and L21.  U152 might be older.

Is it still fair to say that S116, U152, L21 and U106 have variances is a similar enough ballpark to suggest they first appeared in the same period broadly speaking?  By that I mean that it doesnt look like that one of those clades is radically different in age. I got the impression that the variance total for each of these clades in the supplimentary data is fairly similar.    
« Last Edit: August 30, 2010, 05:04:42 PM by alan trowel hands. » Logged
Mike Walsh
Guru
*****
Offline Offline

Posts: 2964


WWW
« Reply #20 on: August 30, 2010, 05:24:49 PM »

... Certainly I think the picture emerging is that France is likely  split into all the major S116 clades with none dominant and that is the main cause for the fairly low Percentages.  In fact France seems to be the only place where a fairly even split among all the main S116 clades with none dominant is the case.  Again along with variance consideration France keeps standing out as the likely source of the earliest expansion of all the main S116  clades.  Surrounding France in all directions it seems that founder effects skewed this even mix towards the dominance of one clade (Iberia, North Italy, Ireland etc).  
As Alan noted, I don't think we can infer too much about the origins from frequency %'s.  I just went through diversity and variance calculations with Ken N, I think it is wise not to infer too much from them either.  I'm not saying the stats are useless, particularly where they seem to all align, but they are just additional data points.

I think the most important thing is just the finding of any significant (more than a Wild Goose find) presence of a subclade, and then lining the subclades to see who is where.  In that regards, France is certainly a point of origin or a crossroads.  It appears SE France and possibly the Alpine region are critical for P312 subclades.  I think it should be notes that U106 has a presence in the Alpin area as well.  I do think U106 is slighltly older than P312 so even if they came from the same place, U106's trail might be thicker back towards the origin.  Same goes for U152 and L21.  U152 might be older.

Is it still fair to say that S116, U152, L21 and U106 have variances is a similar enough ballpark to suggest they first appeared in the same period broadly speaking?  By that I mean that it doesnt look like that one of those clades is radically different in age. I got the impression that the variance total for each of these clades was in the supplimentary data is fairly similar.    

Yes, U106, S116(P312), U152 and L21 all happened in fairly rapid order.  I think U106 is considered older than S116(P312).  Perhaps Vince V. will chime in his perspective.

I think L176 and L176/SRY2627 are younger than any of the above but in the ballgame with the above.

I'm pretty sure that M222 and the other L21 subclades, like L193, P314, L226 are all quite a bit younger.

U152's big subclade, L2, seems to be pretty close to the age of U152.

M153, the so-called Basque marker, seems to be younger like L193, etc., but there is limited data on it.

P312* isn't really a group but we could probably look at the North/South Cluster. It seem to be a about LK176/SRY2627's age. Given their spreads,  I almost wonder if L176 and North-South weren't along for the same ride.

« Last Edit: August 30, 2010, 05:49:43 PM by Mikewww » Logged

R1b-L21>L513(DF1)>S6365>L705.2(&CTS11744,CTS6621)
Mike Walsh
Guru
*****
Offline Offline

Posts: 2964


WWW
« Reply #21 on: August 30, 2010, 05:59:57 PM »

... Certainly I think the picture emerging is that France is likely  split into all the major S116 clades with none dominant and that is the main cause for the fairly low Percentages.  In fact France seems to be the only place where a fairly even split among all the main S116 clades with none dominant is the case.  Again along with variance consideration France keeps standing out as the likely source of the earliest expansion of all the main S116  clades.  Surrounding France in all directions it seems that founder effects skewed this even mix towards the dominance of one clade (Iberia, North Italy, Ireland etc).  
As Alan noted, I don't think we can infer too much about the origins from frequency %'s.  I just went through diversity and variance calculations with Ken N, I think it is wise not to infer too much from them either.  I'm not saying the stats are useless, particularly where they seem to all align, but they are just additional data points.

I think the most important thing is just the finding of any significant (more than a Wild Goose find) presence of a subclade, and then lining the subclades to see who is where.  In that regards, France is certainly a point of origin or a crossroads.  It appears SE France and possibly the Alpine region are critical for P312 subclades.  I think it should be notes that U106 has a presence in the Alpin area as well.  I do think U106 is slighltly older than P312 so even if they came from the same place, U106's trail might be thicker back towards the origin.  Same goes for U152 and L21.  U152 might be older.
Perhaps this is the big new learning for L21+, the fact M222+ may exist enough in Germany beyond Wild Geese type stories.

If that is the case, then L21*  must have been there too and probably either further directly east or further south.
Logged

R1b-L21>L513(DF1)>S6365>L705.2(&CTS11744,CTS6621)
alan trowel hands.
Guru
*****
Offline Offline

Posts: 2012


« Reply #22 on: August 30, 2010, 06:17:40 PM »

... Certainly I think the picture emerging is that France is likely  split into all the major S116 clades with none dominant and that is the main cause for the fairly low Percentages.  In fact France seems to be the only place where a fairly even split among all the main S116 clades with none dominant is the case.  Again along with variance consideration France keeps standing out as the likely source of the earliest expansion of all the main S116  clades.  Surrounding France in all directions it seems that founder effects skewed this even mix towards the dominance of one clade (Iberia, North Italy, Ireland etc).  
As Alan noted, I don't think we can infer too much about the origins from frequency %'s.  I just went through diversity and variance calculations with Ken N, I think it is wise not to infer too much from them either.  I'm not saying the stats are useless, particularly where they seem to all align, but they are just additional data points.

I think the most important thing is just the finding of any significant (more than a Wild Goose find) presence of a subclade, and then lining the subclades to see who is where.  In that regards, France is certainly a point of origin or a crossroads.  It appears SE France and possibly the Alpine region are critical for P312 subclades.  I think it should be notes that U106 has a presence in the Alpin area as well.  I do think U106 is slighltly older than P312 so even if they came from the same place, U106's trail might be thicker back towards the origin.  Same goes for U152 and L21.  U152 might be older.
Perhaps this is the big new learning for L21+, the fact M222+ may exist enough in Germany beyond Wild Geese type stories.

If that is the case, then L21*  must have been there too and probably either further directly east or further south.

Mike-sorry I do not follow you.  What is that about M222 in Germany etc.  I hadnt really paid much attention to M222 so I may have missed something. 
Logged
alan trowel hands.
Guru
*****
Offline Offline

Posts: 2012


« Reply #23 on: August 30, 2010, 07:06:50 PM »

... Certainly I think the picture emerging is that France is likely  split into all the major S116 clades with none dominant and that is the main cause for the fairly low Percentages.  In fact France seems to be the only place where a fairly even split among all the main S116 clades with none dominant is the case.  Again along with variance consideration France keeps standing out as the likely source of the earliest expansion of all the main S116  clades.  Surrounding France in all directions it seems that founder effects skewed this even mix towards the dominance of one clade (Iberia, North Italy, Ireland etc).  
As Alan noted, I don't think we can infer too much about the origins from frequency %'s.  I just went through diversity and variance calculations with Ken N, I think it is wise not to infer too much from them either.  I'm not saying the stats are useless, particularly where they seem to all align, but they are just additional data points.

I think the most important thing is just the finding of any significant (more than a Wild Goose find) presence of a subclade, and then lining the subclades to see who is where.  In that regards, France is certainly a point of origin or a crossroads.  It appears SE France and possibly the Alpine region are critical for P312 subclades.  I think it should be notes that U106 has a presence in the Alpin area as well.  I do think U106 is slighltly older than P312 so even if they came from the same place, U106's trail might be thicker back towards the origin.  Same goes for U152 and L21.  U152 might be older.
Perhaps this is the big new learning for L21+, the fact M222+ may exist enough in Germany beyond Wild Geese type stories.

If that is the case, then L21*  must have been there too and probably either further directly east or further south.

I see now what you are saying.  The France and Germany totals have some M222.  One thing I do not understand is the totals.  How can 'France' or 'Germany' have M222 but the totals that break it down into regions list none?

Personally as M222 appears to be only about a third of the age of L21 etc, I do not really see how the M222 in Germany or France can be native to those areas.  Certainly germline dates for L21 of say 4000 years ago for L21 would only make M222 around the the usual 400AD date , far too late to have come to north Britain and Ireland from France or Germany by any means I can think of.  

I do also wonder again about the figures.  Hobbiest testing has found a lot lot less continental M222 and  think the Santiago de Compostella University paper on French y-DNA only found instances in Paris (hardly surprising) and Tolouse.  
« Last Edit: August 30, 2010, 07:56:01 PM by alan trowel hands. » Logged
argiedude
Old Hand
****
Offline Offline

Posts: 146


« Reply #24 on: August 30, 2010, 08:22:48 PM »

Does anyone have that y-dna study of France (Santiago de Compostela)?
Logged

y-dna: R1b L21
mtdna: U5
Pages: [1] 2 3 ... 6 Go Up Print 
« previous next »
Jump to:  


SEO light theme by © Mustang forums. Powered by SMF 1.1.13 | SMF © 2006-2011, Simple Machines LLC

Page created in 0.181 seconds with 19 queries.