World Families Forums - Estimating L21 using only recently tested samples in ysearch

Welcome, Guest. Please login or register.
July 25, 2014, 04:08:31 AM
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)
| | |-+  Estimating L21 using only recently tested samples in ysearch
« previous next »
Pages: [1] Go Down Print
Author Topic: Estimating L21 using only recently tested samples in ysearch  (Read 2290 times)
argiedude
Old Hand
****
Offline Offline

Posts: 146


« on: August 26, 2009, 06:31:03 PM »

I have a list of ysearch samples that existed in January, 2009, which listed their haplogroup as unknown. I searched for them again this week, and found that 10% of them now listed a haplogroup. I presume they recently took a 2nd test with FTDNA and updated their former unknown status. By using only these samples that presumably have been all retested after January, 2009, we can assume they were tested for the full spectrum of downstream R1b SNPs, such as SRY2627, U106, U152, L21, M222, etc. This should give us an accurate picture of their distribution. Unfortunately, the number of samples involved was very limited.



Some observations.

1) The results, when verifiable with studies, were accurate. Ireland has 17% M222, and the result here was also 17%. Scotland has 7% M222, here it had 13%, somewhat off. The study that looked at S21/S29 in Europe found U106 at 7% in Ireland, here it was also 7%. That study found U106 at 23% in England, here it was 20%. I pooled the German/French/Netherlands results, and got 16% U106; the study had very different results for these countries, and it would seem their average was 22%, somewhat off. The Adams study of Iberia found SRY2627 at 7%, here, admittedly with a ridiculously low amount of samples, the result was 9%. Only 1 or 2 samples, out of 196, were ht35 (one of them had 393 = 13 so it might have been mislabeled). This very low rate of ht35 in West Europe sans Italy is very close to what would be expected from the extremely low amount of R1b1b2 in West Europe with 393 = 12 (most of which must be ht15 which mutated from 393=13 to 393=12).

1.5) The rate at which customers misreported their samples seems to be relatively low. An Irish samples listed as ht35 was obviously M222, the only M222 that was mislabeled. Considering there are about 22 M222 samples overall, that's an acceptable error rate. Yet another ht35 had DYS393 = 13, plus he was from Scotland, which kind of suggests it's likely an error. U106 has a very high rate of DYS390 = 23, and in every single country the samples labeled as U106 had the appropriately high amount of DYS390 = 23, while the non-U106 had likewise an appropriately very low amount of DYS390 = 23. All in all, the results are trustworthy. I think these customers might have even achieved a greater accuracy then your typical peer-reviewed study.

2) I didn't include Italy because it was just 4 samples (3 U152 and 1 U106). I didn't bother with East Europe because it was certain to be even worse.

3) The pooled German/Low Countries/French results are actually 70% from Germany, 10% Belgium/Netherlands, and 20% French.

4) I used the retested samples only to subdivide the R1b section, but the total frequency of R1b in each country is from what I already know very well from other studies.

5) I'm the co-admin of the North Italy Project. I estimate there are only 150 north Italian samples in ysearch. In the last 8 months we have probably received only a dozen new members. 2 of them were L21. Maybe it was just a huge coincidence, or maybe it's indicating that a very large fraction of the uncategorized R1b1b2 results in the Project, most of them tested before FTDNA introduced L21, are in fact L21. Here's a link to the North Italy Project (by the way, I'm one of those 2 recently tested L21's):

http://www.familytreedna.com/public/northitaly/default.aspx?section=yresults

6) After I finished doing this, I realized I could have done the same with the January, 2009, samples that are also labeled as R1b, R1b1, or R1b1b2, to see if they've recently upgraded their  R1b status. I think I could double or triple the results, this way. We'll see tomorrow...

7) A similar method could be done using current ysearch results. One would have to grab all the currently listed R1b's, then wait a couple months, download the list again, identify which ones were added and then use only those to calculate the frequencies of L21, U152, etc.
Logged

y-dna: R1b L21
mtdna: U5
NealtheRed
Old Hand
****
Offline Offline

Posts: 930


« Reply #1 on: August 26, 2009, 07:51:51 PM »

Are you saying that many of those in the North Italy project who haven't tested for L-21 are most likely L-21?


Neal
Logged

Y-DNA: R-Z255 (L159.2+) - Downing (Irish Sea)


MTDNA: HV4a1 - Centrella (Avellino, Italy)


Ysearch: 4PSCK



argiedude
Old Hand
****
Offline Offline

Posts: 146


« Reply #2 on: August 26, 2009, 09:09:00 PM »

Well, I received my result earlier this year and I thought it was crazy luck. Now 8 months later, only a dozen new members have joined the North Italy Project. These are the only people who have had the chance of being tested for L21, all previous R1b1b2 results, just consider that they don't exist. And what do we have? ANOTHER L21! So with just a dozen samples we already have 2 L21's?? My single result could have been wild luck, but now 2 samples with only a dozen people tested...

Also, keep in mind the following. There are probably only 150 samples from north Italy in ysearch. But there are 10,000+ from Scotland and Ireland!
Logged

y-dna: R1b L21
mtdna: U5
NealtheRed
Old Hand
****
Offline Offline

Posts: 930


« Reply #3 on: August 26, 2009, 09:33:08 PM »

Ok, so there are two samples from North Italy so far: you and someone else who just tested positive? And this is out of 12 people so far, but there are more to be tested obviously?

I just want to understand what you're saying! If the case means that more L21 will be found in North Italy, this will certainly be a joyous occasion.


Neal
Logged

Y-DNA: R-Z255 (L159.2+) - Downing (Irish Sea)


MTDNA: HV4a1 - Centrella (Avellino, Italy)


Ysearch: 4PSCK



argiedude
Old Hand
****
Offline Offline

Posts: 146


« Reply #4 on: August 26, 2009, 10:08:06 PM »

FTDNA has only been testing for L21 since about late 2008. In that time only a dozen or so members have joined the North Italy Project, and 2 of them are L21+. There aren't more to be tested. I think what you're referring to is what I said about using R1b, R1b1, and R1b1b2 samples to make a similar comparison like I did with the unknown samples.

Can someone explain when FTDNA started selling L21 as part of a deep clade test and when they included it as part of the standard set of SNPs tested?
« Last Edit: August 26, 2009, 10:12:06 PM by argiedude » Logged

y-dna: R1b L21
mtdna: U5
Maliclavelli
Guru
*****
Offline Offline

Posts: 2146


« Reply #5 on: August 27, 2009, 02:22:40 AM »

I was waiting for many time a R-L21+ from Italy when the SNP L21 was discovered and I supported the origin also of this haplotype/haplogroup in Italy, and then, that I was convinced that it arose in the Rhine Valley, you are finding all these R-L21+ in North Italy. But the second (I don’t know your surname you haven’t ever communicated), Bonnet, I have demonstrated he is of French origin, not only for his surname, but also because he comes from a town of Waldensians from France. Then, please, if you retain to be of Italian descent, consider you so far the unique Italian R-L21+!
Logged

Maliclavelli


YDNA: R-S12460


MtDNA: K1a1b1e

Maliclavelli
Guru
*****
Offline Offline

Posts: 2146


« Reply #6 on: August 27, 2009, 02:46:59 AM »

Certainly your data are so strange that, if your family isn’t migrated from elsewhere, I begin to think again that also R-L21+ arose in the Italian refugium in the Alpine region. Don’t look at the modal (it changes frequently and have many mutations during the time around itself), look at your DYS19=10; DYS385=11,11; DYS426=13; also DYS458=16, that it is modal for us R-L23+ and then very ancient.
Logged

Maliclavelli


YDNA: R-S12460


MtDNA: K1a1b1e

argiedude
Old Hand
****
Offline Offline

Posts: 146


« Reply #7 on: August 27, 2009, 06:06:55 PM »

Ok, this is a much better estimate, using not only the Unknown samples like I did yesterday, but also the R1b, R1b1, and R1b1b2. To recap, I took the ysearch samples that in January were listed as Unknown, or  R1b, and searched for them again this week. Some of them changed their haplogroup to a detailed result, indicating that they've recently retested their sample. By doing this, I ensure that all haplogroups were submitted to the full battery of R1b SNP tests and the results are unbiased. What follows are some observations about the accuracy of the data. At the bottom of the post is the graph with the data.

ht35

This is the only haplogroup where a lot of people made bad calls. 8 out of 18 R1b1b2a (ht35) had DYS393 = 13, which is too high (maybe 10% of ht35 might have 393=13). I simply eliminated these samples. And to top it off, one of the good ones, with 392=12, actually had a perfect haplotype for M222 and it was from Ireland. Groan... I moved him to M222. So about half of the R1b1b2a samples were erroneous. Oh well. The rest of the haplogroups were accurate. After eliminating the bad apples, I found the ht35 results to be supremely agreeable. They're exactly what I would have expected from previous estimates using haplotypes as proxies. Only 1 in 250 West-European-sans-Italian R1b1b2 is ht35. The Italian results were also very expected, note that most of the Italian R1b samples came from the south and Sicily.

SRY2627

I think about half of SRY2627 have DYS490 = 10, while the rest of R1b1b2 have 490 = 12. Of 6 SRY2627 samples, 2 had 490 = 10, which I guess is reasonable. Of 345 non-SRY2627 samples, absolutely none had 490 = 10. This is all very good regarding the reliability of the results.


M153

Interesting there were no M153 results from Spain. It's present at about 5% of its R1b, so we would have expected 2 or 3 samples. But it's also very unevenly distributed in Iberia, so it all depends on where the Iberian ysearch samples come from exactly within Iberia. The same applies to SRY2627, which together with M153 are heavily concentrated in the northeast. As a secondary confirmation of the accuracy of ysearch, it should be noted that outside of Iberia, a dozen SRY2627 were reported but absolutely no M153. That argues against customers randomly and erroneously picking a haplogroup label, because otherwise they would have chosen R1b1b2a1b2 (M153) just as much as R1b1b2a1b3 (SRY2627). And previous studies have shown that SRY2627 is indeed more of a West European haplogroup which has its peak frequency in Iberia, while M153 seems to be mostly found in Iberia.

U152

Highest frequencies in Italy and Switzerland, as expected, once again confirming the accuracy of the ysearch samples.

U106

Netherlands/Belgium had 45% U106. The 2007 study that tested S21 (U106) found 60% of R1b in Netherlands was U106. Remember I lumped Netherlands together with Belgium. Also, that study found 40% of R1b in Germany was U106. So did I... 39%. The U106 results, pooled together, had 390=23 at a rate of 55%. Great. The non-U106 samples had 390=23 at a rate of 10%. Great again.

DYS492 = 13 in U106, and 12 in everybody else. There were only 3 out of 300 non-U106 samples that had 492 = 13, and there were only 4 out of 54 U106 samples that had 492 = 12. If they were all errors, it would result in an error rate of 4% or 5%. But most of them are almost certainly simply STRs that mutated outside of their modal value. For example, there are 4 U106 samples with 492=12, but there are also 5 with 492=14; same with non-U106 samples. So the real error rate is probably just 1% or less. Excellent.

M65

Nobody claimed to be R1b1b2a1b1, the supposedly private SNP from Iberia that is extremely rare and was found in just 2 Basques in the Adams study of 1000+ Iberians. Great, once again this shows that ysearch customers aren't morons who mindlessly enter any haplogroup for their ysearch sample.

M222 - trans-Brittania

None of the 250+ samples from outside of Britain claimed to be M222. Great, yet again. Very good. The accuracy of the self-reported ysearch samples is very good (at least recently). They compare perfectly well with the accuracy of many genetic studies, and I'm not joking. For example, Berniell-Lee, study of Gabon/Cameroon, by my reckoning, he botched 2% of the results. How 'bout that?

M222 - cis-Brittania

Unfortunately, I found 5 samples that were clearly M222 but were labeled otherwise. That's a 9% error rate (5 out of 53). But even if all the haplogroups also had a 9% error rate, the data would still paint an accurate picture. And I have reason to believe the high error rate for ht35 and M222 is limited to them and not indicative of the other haplogroups. Ireland's rate of M222 is too low (16%), it should be more like 22%. Not a small difference. Scotland's rate is picture perfect. Wales and England's results are also expected.

L21

I'm not going to spoil the fun. See for yourself. Very interesting. Remember to consider things over all of the y-dna, not just as a percentage of R1b. I included the frequencies of R1b for each region in the blue row at the bottom of the graph.


« Last Edit: August 27, 2009, 06:08:09 PM by argiedude » Logged

y-dna: R1b L21
mtdna: U5
argiedude
Old Hand
****
Offline Offline

Posts: 146


« Reply #8 on: August 27, 2009, 06:13:33 PM »

I was waiting for many time a R-L21+ from Italy when the SNP L21 was discovered and I supported the origin also of this haplotype/haplogroup in Italy, and then, that I was convinced that it arose in the Rhine Valley, you are finding all these R-L21+ in North Italy. But the second (I don’t know your surname you haven’t ever communicated), Bonnet, I have demonstrated he is of French origin, not only for his surname, but also because he comes from a town of Waldensians from France.

Yeah, that is a problem. I don't remember how far back he stated his ancestors have been outside of Italy.

Quote
Then, please, if you retain to be of Italian descent, consider you so far the unique Italian R-L21+!

There was also the L21+ P66+ guy from Italy. But the Italian results I just posted, mostly from the southern half, don't have any L21. So at least in the south, it's rare.
Logged

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

Posts: 5023


« Reply #9 on: August 27, 2009, 08:09:14 PM »

. . .

L21

I'm not going to spoil the fun. See for yourself. Very interesting. Remember to consider things over all of the y-dna, not just as a percentage of R1b. I included the frequencies of R1b for each region in the blue row at the bottom of the graph.




For some reason I can't make that image any bigger so I can see it.

I think there should be more R-L21* results in Northern Italy, too. Even though FTDNA began testing for L21 in October, it was as a special-order, stand-alone test. FTDNA did not add L21 to its Deep Clade-R test until late January of 2009.

If we could just get more R1b1b2s in the North Italy Project tested!
« Last Edit: August 27, 2009, 08:09:47 PM by rms2 » Logged

argiedude
Old Hand
****
Offline Offline

Posts: 146


« Reply #10 on: August 27, 2009, 08:14:39 PM »

For some reason I can't make that image any bigger so I can see it.

Right click and select save or view.

Quote
I think there should be more R-L21* results in Northern Italy, too. Even though FTDNA began testing for L21 in October, it was as a special-order, stand-alone test. FTDNA did not add L21 to its Deep Clade-R test until late January of 2009.

When did they include L21 as part of the normal set of SNPs tested (like U106 or U152)?
Logged

y-dna: R1b L21
mtdna: U5
Maliclavelli
Guru
*****
Offline Offline

Posts: 2146


« Reply #11 on: August 28, 2009, 12:53:27 AM »

Argiedude, I appreciate indeed very much your researches, and not only now. I invite you to give a glance to my postings on Dienekes' blog about R-L23- (they are in the thread where I critisized, let's say destroyed as well, the paper on the CMH). I am not able to see your postings on "dna-forums", then, please, let me know an updated spreadsheet of R-L23-. The node about the Italian refugium I think is here, and in the R-L23+/L150-.
I'll post to Ysearch some interesting haploype with DYS385=11-13.2.
Logged

Maliclavelli


YDNA: R-S12460


MtDNA: K1a1b1e

rms2
Board Moderator
Guru
*****
Offline Offline

Posts: 5023


« Reply #12 on: August 28, 2009, 07:13:49 PM »

For some reason I can't make that image any bigger so I can see it.

Right click and select save or view.

Quote
I think there should be more R-L21* results in Northern Italy, too. Even though FTDNA began testing for L21 in October, it was as a special-order, stand-alone test. FTDNA did not add L21 to its Deep Clade-R test until late January of 2009.

When did they include L21 as part of the normal set of SNPs tested (like U106 or U152)?

I finally saw your chart on Rootsweb. Nice work, but some of our guys have not updated their YSearch entries to show that they are R1b1b2a1b5 and a number of them have no YSearch entries at all.

The answer to your question is late January of 2009, when L21 became part of FTDNA's Deep Clade-R test.

I think the downturn in the world economy has hampered testing for L21. Remember that U106 and U152 were discovered in 2005, in good economic times, and thus have had nearly a four-year headstart on L21, most of it, as I said, in good economic times.

L21 was discovered (for all practical purposes) just as the world economy went down the toilet.
Logged

argiedude
Old Hand
****
Offline Offline

Posts: 146


« Reply #13 on: August 29, 2009, 03:47:24 PM »

Argiedude, I appreciate indeed very much your researches, and not only now. I invite you to give a glance to my postings on Dienekes' blog about R-L23- (they are in the thread where I critisized, let's say destroyed as well, the paper on the CMH). I am not able to see your postings on "dna-forums", then, please, let me know an updated spreadsheet of R-L23-. The node about the Italian refugium I think is here, and in the R-L23+/L150-.
I'll post to Ysearch some interesting haploype with DYS385=11-13.2.

Thanks, I included a link to those L23- graphs in rootsweb. Same exact post, same exact images.

If you want, I could ask in dna-forums if banned members can be allowed to rejoin with only the option of viewing threads, not posting.

I finally saw your chart on Rootsweb. Nice work, but some of our guys have not updated their YSearch entries to show that they are R1b1b2a1b5 and a number of them have no YSearch entries at all.

It doesn't matter. That's the good thing about purposefully selecting samples that were tested after January. It avoids exactly that kind of problem. Post-January L21+ samples shouldn't have any reason to indicate their haplogroup as R1b1b2a1b.

Quote
The answer to your question is late January of 2009, when L21 became part of FTDNA's Deep Clade-R test.

Thanks, that was the best possible answer I could've hoped for. Expect a massive update to my chart next week! I think I'll be able to get even more detailed this time, for example, regions within England, or dividing Germany into at least north/south and east/west, as well as Italy.

Hmmm, if FTDNA introduced L21 as part of the standard test in late January, what would this mean more specifically? Would samples ordered in early January but still in the testing process when the announcement was made be admitted into the new policy and have L21 tested?

How fast is FTDNA's turnaround, currently? If you receive the kit and send it back today, when should you expect to see your SNP-tested result posted?
Logged

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

Posts: 5023


« Reply #14 on: August 29, 2009, 06:23:25 PM »

. . .

Hmmm, if FTDNA introduced L21 as part of the standard test in late January, what would this mean more specifically? Would samples ordered in early January but still in the testing process when the announcement was made be admitted into the new policy and have L21 tested?

There were people who did not quite make the cut who were not getting L21 results as part of their old-version Deep Clade-R even as the first of the Deep Clade-Rs that included L21 were rolling in. So, no, some of them did not make the new L21-inclusive version of the Deep Clade-R.

There are still people in YSearch and in FTDNA projects who are listed as "R1b1b2a1b" who have never been tested for L21.

How fast is FTDNA's turnaround, currently? If you receive the kit and send it back today, when should you expect to see your SNP-tested result posted?

It varies. For L21 as a stand-alone test, it can be very fast. I haven't kept an actual record, but it seems to me some have come in in as little as three to four weeks. Some take longer, but a month or two seems about average.

The same goes for the Deep Clade-R, although it seems to take a bit longer than just L21 as a stand-alone.
« Last Edit: August 29, 2009, 06:24:09 PM by rms2 » Logged

Maliclavelli
Guru
*****
Offline Offline

Posts: 2146


« Reply #15 on: August 30, 2009, 02:54:27 AM »

Argiedude writes:” If you want, I could ask in dna-forums if banned members can be allowed to rejoin with only the option of viewing threads, not posting”.

Dear Argiedude,
I thank you for your proposal, but I don’t know if you are speaking by yourself or for others’ mandate. I had a similar proposal, I wrote this mail and I had no answer. Now I say to you that I’ll search for your postings and I’ll consider always your opinions, but with Van Laar & Co. I’ll desire nevermore having anything to do. From here to eternity.
Kinds regards, Gioiello Tognoni del Badia. R-L23+/S136+. K1a1b1.

(9 March 2009)
Dear Rick,
I thank you for your kind words and I ask you to pass my posting to Mr Van Laar.
Really I don’t desire to come back and write on the forum. There isn’t an unanimous agreement on this, even though what has happened in these days has surprised and moved me and sincerely it was unexpected, but perhaps, if Ruth hasn’t started with her greetings, no one would have thought on me.
Really I don’t desire to write on the forum as I am very different from your multicultural and multiracial world. I was born in Tuscany three years after the end of WWII, and in my thorp we were all Tuscans, from thousands of years we were only Tuscans, and also a man from a different region was a different one. I am the last of this world, being my wife from Sicily and then my sons a little bit “Greeks, Arabs, Jews, Berbers, Normans, Albanians and others” as Maddi said.
I ask only to Van Laar to can participate to the forum as a reader, to have a permanent moderation. Why who is banned can’t either read the postings? On “Genealogy-dna”, also after my banishment, I have always had  access and I could subscribe to “dna-forum” with a false name for having this. But I have only a name: mine.
I am a polemist by nature and I’ll be to the end. Now that Gaza War has finished and the papers on Jewish and Arab ancestry in Spain  and the Capelli’s second time have been published, I hope you can allow me willingly what  I could have  got  by fraud.
 
King Regards, Gioiello Tognoni


Logged

Maliclavelli


YDNA: R-S12460


MtDNA: K1a1b1e

rms2
Board Moderator
Guru
*****
Offline Offline

Posts: 5023


« Reply #16 on: August 30, 2009, 08:37:47 AM »

Argiedude -

Can you explain the figures on your chart a bit? What exactly do they mean? Are they percentages of the total R1b entries for a country?

I understand from a friend of mine your chart is being used by some misguided individuals on another forum (one I don't frequent) to claim that L21 originated in the British Isles, a shortsighted claim that overlooks a number of facts, but one that pops up from time to time because it is so convenient for genealogical purposes.

Has anyone pointed out to them the overwhelming number of persons of British Isles ancestry in YSearch and in dna testing overall?

Have they looked at the number of men in the British Isles categories of the other haplogroup projects, like the U106 Project, or the number of those in Colonial categories with British surnames? Even though testing for U106 has been going on since 2005, British Isles U106+ still outnumbers all the other categories by a pretty good margin (I haven't gone to that project to count, but I know this is true).

Will they argue that U106 first arose in the Isles?

Much is made of Tim Janzen's Rootsweb postings, in which he used figures from several months ago to compare a tub full of British L21 haplotypes to a thimbleful of continental haplotypes and yet came up with but 140 years difference in variance, hardly a statistically meaningful figure (a fact that was pointed out at the time by Vince Vizachero and a number of others). Yet Janzen - Lord knows why - has mused on Rootsweb that L21 might have arisen in Scotland!

Best to ignore such silliness, I guess, which is one more reason I don't frequent dna forums.

« Last Edit: August 30, 2009, 08:40:01 AM by rms2 » Logged

rms2
Board Moderator
Guru
*****
Offline Offline

Posts: 5023


« Reply #17 on: August 30, 2009, 01:51:57 PM »

Looking at argiedude's chart (http://img504.imageshack.us/img504/2466/r1bestimatesusingunknow.gif), here are some of my thoughts, especially with regard to R-L21*.

The two in YSearch who list Italy for their most distant y-dna ancestor were left off the chart.

David Faux has nine Iberian U152+ (eight Spaniards and one Portuguese) listed on his U152 web site: http://www.davidkfaux.org/R1b1c10_Data.htm. Four of them do not have YSearch entries, so they're not showing up. Of course, that is a YSearch entry problem and not argiedude's fault. It is just one of the limitations of using YSearch for this kind of thing. We have the same problem with L21+ guys who never create YSearch entries.

Now on to France. R-L21* shows up as 20% of the R1b in France on argiedude's chart. I wonder if the out-of-the-Isles crowd realizes what an astounding figure that is (even given the limitations of YSearch) for such a relatively recently-discovered SNP.

France is roughly 60% R1b. If L21 is even merely 20% of French R1b (I suspect the true figure will be higher than that), that means 12% of all Frenchmen are L21+! If, as I suspect, the true figure is closer to between 30 and 40%, then 18-24% of Frenchmen are L21+.

Think about that. About 63 million people live in mainland France. Probably at least 30 million of them are male. Compare that with Ireland. Although Ireland currently looms large in y-dna databases - and don't get me wrong, I am part Irish and love the Irish - it is an island of 5 million people (the Paris metro area alone has about 12 million people in it). Of course, there are considerably more men of Irish descent in the Irish diaspora (I think) than in the old home island itself. But, even so, 20% of the French R1b is a considerable - even an amazing - number.
Are we supposed to believe 20% or more of French R1b came from the British Isles originally?

I would venture to say, although I have not tried to collect the exact figures, that perhaps 50-60% of the R1b in Northern France is L21+. Yes, that is right, that is what I am guesstimating based on my experience as admin of the R-L21 Plus Project and what I have been seeing in the results of those of French y-dna ancestry who have been tested for L21. L21 tapers off rapidly to the Southwest in France but does very well in the North, Northwest, and East.

To Germany. The figure of 4% surprises me, especially since R-P312* is listed as 23%. I suspect something is wrong there. Either 1) there are quite a few German R-P312* who are in YSearch who have not joined the R-P312 and Subclades Project; 2) there are Germans listing themselves as R1b1b2a1b who have not actually been tested for L21; 3) there are German R-L21* still listed in YSearch as something else; 4) there are German R-L21* without YSearch entries; or a combination of all of those factors.

Nevertheless, let's think for a minute. L21 was discovered just last October, as the world economy was going down the toilet. It did not become a part of FTDNA's Deep Clade-R test until late January of this year, 2009, as I said, during rotten economic times.

Yet L21 is already 4% of the German R1b entries in Ysearch, despite the limitations of Ysearch and the huge headstart - in good economic times - subclades like U152 and U106 have enjoyed. (That makes the French figures that much more astounding.)

No need to go over how populous a country Germany is compared with, say, Ireland (again).

I am not a bean counter by nature. I was good in math in school, but it's not something I have kept up with or do for grins. So, I haven't made an exact tally, but I know from experience that the British Isles component in commercial dna testing is overwhelming. It creates a fog, especially when it comes to R1b, that makes understanding the evidence difficult. That is especially true for an SNP like L21, which has done very well for itself in the British Isles and that, at least in my opinion, originated on the European Continent.

Take for example M222. One would think it was one of the very most frequent R1b SNPs, if not the most frequent. Look at the R-M222 Project. That son-of-a-gun has over 500 members! Yet in all probability R-M222 is actually a fairly small proportion of total R1b. Its significance has been magnified many times by the level of British Isles bias in the commercial genetic genealogy databases.

I guess this post is long enough already, but I would urge everyone to keep in mind how new an SNP L21 is, how under-tested the European Continent is (especially for L21) compared to the British Isles, and how much the current economy has had an adverse effect on dna testing in general and Deep Clade/SNP testing in particular.
« Last Edit: August 30, 2009, 02:04:43 PM by rms2 » Logged

Mike Walsh
Guru
*****
Offline Offline

Posts: 2963


WWW
« Reply #18 on: August 30, 2009, 03:44:54 PM »

Looking at argiedude's chart (http://img504.imageshack.us/img504/2466/r1bestimatesusingunknow.gif), here are some of my thoughts, especially with regard to R-L21*.

.....
France is roughly 60% R1b. If L21 is even merely 20% of French R1b (I suspect the true figure will be higher than that), that means 12% of all Frenchmen are L21+! If, as I suspect, the true figure is closer to between 30 and 40%, then 18-24% of Frenchmen are L21+.

Think about that. About 63 million people live in mainland France. Probably at least 30 million of them are male. Compare that with Ireland. Although Ireland currently looms large in y-dna databases - and don't get me wrong, I am part Irish and love the Irish - it is an island of 5 million people (the Paris metro area alone has about 12 million people in it). Of course, there are considerably more men of Irish descent in the Irish diaspora (I think) than in the old home island itself. But, even so, 20% of the French R1b is a considerable - even an amazing - number.
Are we supposed to believe 20% or more of French R1b came from the British Isles originally?
.......
 So, I haven't made an exact tally, but I know from experience that the British Isles component in commercial dna testing is overwhelming. It creates a fog, especially when it comes to R1b, that makes understanding the evidence difficult. That is especially true for an SNP like L21, which has done very well for itself in the British Isles and that, at least in my opinion, originated on the European Continent.

Take for example M222. One would think it was one of the very most frequent R1b SNPs, if not the most frequent. Look at the R-M222 Project. That son-of-a-gun has over 500 members! Yet in all probability R-M222 is actually a fairly small proportion of total R1b. Its significance has been magnified many times by the level of British Isles bias in the commercial genetic genealogy databases.
.....
I'm not sure about the math, but I can say that I'm having a hard time keeping up with new R-L21* people being identified.  It feels like there are a lot of L21+ people out there.  Don't forget, all M222+ people are L21+.

RMS2 makes a good point about the population in France and the low testing level there.    Although not quite the same point, I have a similar feeling about England.   We tend to look at Ireland and get mesmerized by all of the L21 folks there.  However, the general population is so much higher in England that I wouldn't be surprised if there weren't more L21 folks in England than Ireland.
« Last Edit: August 30, 2009, 03:46:11 PM by Mike » Logged

R1b-L21>L513(DF1)>L705.2
rms2
Board Moderator
Guru
*****
Offline Offline

Posts: 5023


« Reply #19 on: August 30, 2009, 04:04:35 PM »

. . .   Although not quite the same point, I have a similar feeling about England.   We tend to look at Ireland and get mesmerized by all of the L21 folks there.  However, the general population is so much higher in England that I wouldn't be surprised if there weren't more L21 folks in England than Ireland.


I agree, and they seem to be all over England, too, not just in the West or Northwest.

But probably the proportion of L21 in the total male population of Ireland will be greater than just about anyplace else, except for maybe Wales, which may equal that of Ireland.

But, of course, that says almost nothing about where L21 originated.

Personally, and I know I will be criticized for saying this, I think, for the most part, there is a direct connection between L21 and Celtic languages and culture. I am not excluding the part played by other subclades (like U152) and even other haplogroups, but the subject of this discussion is L21.

Celtic language and culture got to the British Isles somehow, and they spread from the Continent to the Isles and not the other way around.

Right now haplotype variance has told us little. Back several months ago, when Janzen compared a boatload of British Isles 67-marker haplotypes to a very small number of 67-marker continental haplotypes, the difference in variance amounted to a paltry 140 years. No help there. So let's look at SNP diversity. Where is there greater P312+ diversity, in the British Isles or on the Continent? I think the answer is pretty obvious.

How likely is it that P312 spun off its other "children" on the Continent but got to the Isles before giving rise to L21? And then we are supposed to believe L21 doubled back to the Continent in sufficient strength to show up in quantity there - and just since the historical period, as some believe?

To make those last points clearer, consider the age estimates for P312 and its subclades, or for the big P311+ family, for that matter. Things were happening fast. P312 is not much older than its offspring. Where did P312 first arise? I wouldn't look too far from there for the origin of its subclades, including L21.
« Last Edit: August 30, 2009, 04:06:53 PM by rms2 » Logged

NealtheRed
Old Hand
****
Offline Offline

Posts: 930


« Reply #20 on: August 30, 2009, 10:23:51 PM »

Right now, I think it's hard to pin down P312's point of origin, but I'm guessing somewhere in the Caucasus or Balkans.

I think that L21 was exactly what people needed to connect Celtic populations in the British Isles with their forebears in Southern Germany and Switzerland. With the discovery of U152, it was easy for people to dismiss the idea of a large Celtic migration to Britain and Ireland, but not anymore. I don't see why people still have to argue about it; it's as plain as day when you look at the distribution of L21.


Logged

Y-DNA: R-Z255 (L159.2+) - Downing (Irish Sea)


MTDNA: HV4a1 - Centrella (Avellino, Italy)


Ysearch: 4PSCK



Pages: [1] 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 0.144 seconds with 19 queries.