U106 'the wander years'

[alan trowel hands.]

One reason why I like the U106-Lusatian link is it  was a wealthy trading culture in an otherwise fairly impoverished late Bronze Age in the northern European plain.  Its relationships are complex though and its hard to pin down.  I remember reading about it years ago when it was usually called Lausitz culure and it seemed an impressive culture of the hierarchical trading elites type with fingers in many pies.  It seems to be thought of as IE but exactly what seems mysterious.  It seems to have been linked into both central Europe (and even the Med. via that) but also the Nordic Bronze Age too.

[Mikewww]

Two things as I understand. Busby selected a set of STRs that had a better linearity (D)

I forgot to add the specific STRs that were used in Busby's study but 389i was not used in my estimates.

DYS19   DYS389I   DYS389b   DYS390   DYS391   DYS392   DYS393   DYS437 DYS438   DYS439   DYS448   DYS456   DYS458   DYS635   Y-GATA-H4

Ouch. We've been through this so I don't want to re-bore people to death. I wouldn't assume Busby's STR set is a good for comparing variance between populations. According to their own methodology the majority of these aren't even linear for the timeframe. According to Marko Heinilla's linear duration analysis, his work supports these STRs aren't very good as well.  To be fair to Busby, I don't trust the Myres or Balaresque usage of STR diversity (much of the data is common) either which also means when I cited the Myres results in this thread they too are really shaky.

These STR marker sets are just way too limiting.  I don't find consistency until I get up to using a very high number of markers, like 49 of FTDNA's 67, or Marko's longest linear 36 out of FTDNA's 67.... but that's a whole another topic. I can show long haplotype STR variance that I think is good but that will bring on complaints about FTDNA's lack of representativeness across geographies.  Still, it's the best we've got as far as I'm concerned where we do have decent numbers. There are lots of holes. geographically, in the data from FTDNA, although there are holes in the geographical distribution of the academic data as well.

[Jean M]

One reason why I like the U106-Lusatian link is it  was a wealthy trading culture in an otherwise fairly impoverished late Bronze Age in the northern European plain. ... It seems to be thought of as IE but exactly what seems mysterious.  

Lusatian or Lausitz is just a type of Urnfield found in Poland and East Germany. We cannot be certain of the dialect of IE spoken there, since of course the region has seen subsequent waves of migration bringing Germanic and then Slavic, both very late types of IE. It might have been a bit different from the Celtic that was developing in the adjoining North Alpine region and the Baltic developing to the east. Both were pretty close to PIE in their early stages. So if we just imagine something fairly close to PIE, we won't go far wrong I imagine.

[stoneman]

U106 is at least 6000 ybp. Variance is unreliable. The subclade Z156 is a Bronze age SNP originating in the Isles.

[Mkk]

U106 is at least 6000 ybp. Variance is unreliable. The subclade Z156 is a Bronze age SNP originating in the Isles.

What are you relying on for that 6000 ybp? Variance? Most of the amateur calculations come out as about 4000 years.

[Jarman]

Lusatian or Lausitz is just a type of Urnfield found in Poland and East Germany.

A more general question.  When did cremation replace inhumation in the now Polish regions? . . . Scandinavia? . . . northern Genrmany?  Is cremation connected to arrival of IE?  Thanks.

[Mark Jost]

MikeW,

I just ran U106 variance at 50 markers and got some surprises using your latest Spreadsheet.

Edit: Added column for counts.

N=   U106   Variance
n=   U106   Variance      Gen   Age @30
6     15.90   IS Eng North East      143.1   4,292.3
17    15.46   IS Sco East      139.1   4,172.5
6     15.23   EW Fra North & Central      137.1   4,112.4
4     14.58   IS Ire Connacht      131.2   3,936.9
14    14.10   IS Sco North      126.9   3,807.6
32    13.84   Poland      124.6   3,737.0
6     13.67   Ukraine      123.0   3,689.4
5     13.30   EW Fra Northeast      119.7   3,590.5
29    13.17   IS Eng South East      118.5   3,555.3
147    13.08   Ireland      117.7   3,529.8
49    13.01   IS Ire Ulster      117.1   3,513.2
11    12.95   EE East Mediterranean      116.6   3,497.2
50    12.89   EW Low Countries      116.0   3,479.0
147    12.81   Scotland      115.3   3,458.6
24    12.80   France      115.1   3,454.2
74    12.60   EE East Cont Euope      113.4   3,401.5
13    12.51   IS Sco West & Central      112.6   3,377.9
36    12.51   EW Ger Middle      112.6   3,376.6
5     12.50   Hungary      112.5   3,374.5
7     12.24   Spain EW Iberian Peninsula      110.1   3,303.8
11    12.18   Norway      109.6   3,287.0
14    12.11   IS Eng West Midlands      109.0   3,269.7
122    12.02   Germany      108.1   3,243.7
10    12.01   Finland      108.1   3,242.5
37    12.01   NO Baltic Sea      108.0   3,241.5
3     12.00   EW Fra North Atlantic      108.0   3,239.5
14    11.92   Denmark      107.3   3,218.7
21    11.83   EW Ger South      106.5   3,194.5
56    11.78   IS Eng South West      106.0   3,181.2
13    11.75   IS Wales      105.7   3,170.9
4     11.33   Czech Rep      102.0   3,059.5
17    11.31   IS Ire Leinster      101.8   3,054.6
216   11.30   Is Eng All      101.7   3,051.8
10    11.20   IS Ire Munster      100.8   3,023.5
22    11.14   EW Alpine & Cisalpine      100.3   3,008.6
22    10.94   IS Sco South      98.4   2,953.2
9     10.75   EW Iberian Peninsula      96.7   2,902.1
17    10.54   IS Eng East Midlands      94.9   2,845.6
31    10.45   IS Eng East      94.1   2,821.7
21    10.38   IS Eng Yorkshire      93.4   2,803.1
17    10.22   EW Ger North      91.9   2,757.9
27    10.21   Sweden      91.8   2,755.0
17    10.02   IS Eng London      90.2   2,704.6
14    9.65   Switzerland      86.8   2,605.5
6     8.42   Belarus      75.7   2,272.2
2     7.50   EW Aquitaine & Pyrenees      67.5   2,024.7
4     7.33   Slovakia      66.0   1,979.7
10    7.22   Lithuania      65.0   1,949.7
5    6.30   Russia      56.7   1,700.7

MJost

[Jarman]

MikeW,

I just ran U106 variance at 50 markers and got some surprises using your latest Spreadsheet.

Variance   U106
20.1   IS Wales
15.9   IS Eng North East
15.5   IS Sco East
15.2   EW Fra North & Central
14.6   IS Ire Connacht
14.1   IS Sco North
13.8   Poland
13.7   Ukraine
13.3   EW Fra Northeast
MJost

Does it matter if DYS390 modal is considered 23 or 24?

[Please forgive if you answered this previously.]

[Mikewww]

MikeW,

I just ran U106 variance at 50 markers and got some surprises using your latest Spreadsheet.

Variance   U106
20.1   IS Wales
15.9   IS Eng North East
15.5   IS Sco East
15.2   EW Fra North & Central
14.6   IS Ire Connacht
14.1   IS Sco North
13.8   Poland
13.7   Ukraine
13.3   EW Fra Northeast
MJost

Thanks, Mark. This will probably make Jaska happy because he doesn't like STR diversity by region. This is fairly granular. How big are the sample sizes?

[Mark Jost]

Does it matter if DYS390 modal is considered 23 or 24?

[Please forgive if you answered this previously.

Only as it relates to any U106, not any subclades which I didnt not consider using. Only Regions.

MJost

[Mark Jost]

MikeW,

I just ran U106 variance at 50 markers and got some surprises using your latest Spreadsheet.

Variance   U106
20.1   IS Wales
15.9   IS Eng North East
15.5   IS Sco East
15.2   EW Fra North & Central
14.6   IS Ire Connacht
14.1   IS Sco North
13.8   Poland
13.7   Ukraine
13.3   EW Fra Northeast
MJost

Thanks, Mark. This will probably make Jaska happy because he doesn't like STR diversity by region. This is fairly granular. How big are the sample sizes?

You know I thought of that more than half way through, so no I didnt but these are Tested U106's from your latest spread sheet. I will get the info.

MJost

[Mark Jost]

Thanks, Mark. This will probably make Jaska happy because he doesn't like STR diversity by region. This is fairly granular. How big are the sample sizes?

ok posted the counts in the detail post
http://www.worldfamilies.net/forum/index.php?topic=11059.msg142049#msg142049

MJost

[Peter M]

MikeW,

I just ran U106 variance at 50 markers and got some surprises using your latest Spreadsheet.

Edit: Added column for counts.

N=   U106   Variance
13   20.07   IS Wales
6    15.90   IS Eng North East
17   15.46   IS Sco East
6    15.23   EW Fra North & Central
4    14.58   IS Ire Connacht
14   14.10   IS Sco North
32   13.84   Poland
6    13.67   Ukraine
5    13.30   EW Fra Northeast
29   13.17   IS Eng South East
49   13.01   IS Ire Ulster
11   12.95   EE East Mediterranean
50   12.89   EW Low Countries
24   12.80   France
74   12.60   EE East Cont Euope
13   12.51   IS Sco West & Central
36   12.51   EW Ger Middle
5    12.50   Hungary

MJost

Mark (or anybody for that matter), what conclusions would you draw from this data ?? This IS a serious question.

BTW, let's assume for the sake of argument, that throwing all branches of a SNP like U106 on a single heap will produce anything but utter nonsense and let's see what this data as such will lead us to.

[Mikewww]

Mark (or anybody for that matter), what conclusions would you draw from this data ?? This IS a serious question.

Peter, I think the sample sizes for many of the regions are quite small so it may be a little too granular of a geographic analysis to be very helpful (for making conclusions) yet.

[Mark Jost]

ok, Wales was incorrect, one of the haplotypes (f147174 Edwards) had a zero in 391. I remove 0** and 0's as a standard but must have missed this one. Thus dropping the variance to 11.4. But I rechecked every count and every variance amount and reposted them.

MJost

[Mark Jost]

Mark (or anybody for that matter), what conclusions would you draw from this data ?? This IS a serious question.

BTW, let's assume for the sake of argument, that throwing all branches of a SNP like U106 on a single heap will produce anything but utter nonsense and let's see what this data as such will lead us to.

I need to understand if it is incorrect to use combined variances (throwing all branches of a SNP like U106 on a single heap will produce anything but utter nonsense)

Nordtvedt states:
"Self-variances of entire clade C, or of separate subclades A and B, will give average age estimates (AAE) younger  than node n1, or n2 and n3, respectively.  " (Note: clade C (say U106), subclades A (Z38) and B(Z18), ect.)

http://knordtvedt.home.bresnan.net/Different%20Variances.ppt

We are using 67 markers haplotypes improving the confidence. Ken method uses variance (both Whole and Sample population) to calculate generation ages and an increased variance is an increased age to a common ancestor.

MJost

[Arwunbee]

Mark (or anybody for that matter), what conclusions would you draw from this data ?? This IS a serious question.

To me, this information is more confirmation that U106 played a considerable role in the Volkswagen Wanderings out of the SW Baltic.  Perhaps behind that it is also indicating that there have been a series of minor VW roadtrips from the Schleswig-Holstein region pre Roman period.  I'd be surprised if variance ever gives us the origin of U106 to within a 500 mile radius accuracy.

[Mark Jost]

Mark (or anybody for that matter), what conclusions would you draw from this data ?? This IS a serious question.

To me, this information is more confirmation that U106 played a considerable role in the Volkswagen Wanderings out of the SW Baltic.  Perhaps behind that it is also indicating that there have been a series of minor VW roadtrips from the Schleswig-Holstein region pre Roman period.  I'd be surprised if variance ever gives us the origin of U106 to within a 500 mile radius accuracy.

I dated the variances and posted them in the previous post.


They were wonders at that.

MJost

[Peter M]

Mark (or anybody for that matter), what conclusions would you draw from this data ?? This IS a serious question.

BTW, let's assume for the sake of argument, that throwing all branches of a SNP like U106 on a single heap will produce anything but utter nonsense and let's see what this data as such will lead us to.

I need to understand if it is incorrect to use combined variances (throwing all branches of a SNP like U106 on a single heap will produce anything but utter nonsense)

Nordtvedt states:
"Self-variances of entire clade C, or of separate subclades A and B, will give average age estimates (AAE) younger  than node n1, or n2 and n3, respectively.  " (Note: clade C (say U106), subclades A (Z38) and B(Z18), ect.)

http://knordtvedt.home.bresnan.net/Different%20Variances.ppt

MJost

I guess it depends on what you are using them FOR. Ken is using Variance to estimate the AGE of a whole clade and/or of sub-clades. It appears to me, you are using variance as a tool to estimate ORIGIN, or possibly migration paths, which, I guess, is something completely different.

I have a pet theory about migration paths of SNPs (I call it "red cap watching") and I hope to be able to explain it on the L257.org web site (if the hackers give me a chance, that is). The key thing is, one can only talk about the migration path of, say, U106 as long as it behaves as a single entity. Exactly this is what the various branches of U106 put together have NOT done and therefore one cannot treat them as one big group. If you don't believe me, look at the geographical spread of Z156 wrt L48. What is the meaning of any quantity derived from those two taken together ??

But still, I find it interesting to hear what conclusions you draw from the data you've posted.

[Mark Jost]

N=8   R1b-Z372
N=55   R1b-L257
PeterM

SNP                              YBP         +OR-YBP
R1b-Z372  GAFounder = 2,256.1   780.4
R1b-L257  GBFounder = 1,754.3   688.2

Interclade
YBP    +OR-YBP
1,782.0   693.6

Looks very close to your 500AD as you have sugggest on your website.

MJost

[Mikewww]

The key thing is, one can only talk about the migration path of, say, U106 as long as it behaves as a single entity. Exactly this is what the various branches of U106 put together have NOT done and therefore one cannot treat them as one big group.

I get what you are saying and I agree it is a problem. The essence of the problem is that all of the U106 people in a single geography are not related to a single MRCA (Most Recent Common Ancestor) when compared to another geography. There was may have been and probably were exchanges between the geographies as well as interventions of U106 people from other geographies not in the comparison.

However, still, there is no need to throw the baby out with the bath water. If the STR diversity of a clade is very high in geography A (say .9) and very low in geography B (say .1) and the populations of the clade are significant in both A and B, then I think there is meaning which would increase the probabilities of which population was older. Now if you say the diversity of U106 in A is .89 while in B is .84 that is insignificant for something so imprecise as STR diversity.

So the data should be analyzed in terms of relevance to other known data and to significance in variation. I don't think this is all so hard to grasp. No data is perfect but that does not mean one should not arbitrarily disregard data.

To the point of comparing clades only at an aggregate level, this is an infinite problem. I absolutely agree we need to break the data down to the most granular subclades as possible, so yes, Z156 should be looked, Z381 should be looked at, L48, etc., etc. but the counter balance is the loss of sample size to a point of no statistical validity.  I just think you have to look at the onion as a whole, and you have to peel it back and look at the pieces, then look at it all along with the other facts.

Sometimes I think we are searching for the silver bullet. There is no such thing. There are many bullets, some misleading, some right on, and then there are shotgun shells and everything other kind of thing. We must look at them all individually and together.

I'd be surprised if variance ever gives us the origin of U106 to within a 500 mile radius accuracy.

Me too, at least for a while to come.  However, I'd be ecstatic if we could get within 500 miles. It looks like over 500 miles from London to Bern, over 500 miles from Bern to Rome, over 500 from Bern to Berlin, over 1000 from Berlin to Odesa, over 700 from Bern to Sarajevo and over 700 miles from Sarajevo to Istanbul.  If we could figure out which city U106 originated closed to as well as P312 and L11, wow!

STR diversity AND SNP diversity could be a great hope. In fact, we don't have much better hope than in these measurements for Y lineages. Frequency has more of a relationship to the end of the trail than to the beginning.

Something else to keep in mind is that STR diversity reflects SNP diversity. Typically, areas with great STR diversity have great SNP diversity. The subclade diversity drives the STR diversity.  You could say that still it is just an admixture, but the presence of brothers, 1st cousins, 2nd cousins, etc. is meaningful, at least in comparison to their lack of presence.

Does that mean that a high diversity area is a launch point or a pooling point/destination?  Stand-alone, we don't know. That's why it is important to look at the archeology, history, etc. .... but areas of signficantly lower diversity can probably be (almost) ruled out as origin points.

[stoneman]

Posting Removed

Terry

[Peter M]

The key thing is, one can only talk about the migration path of, say, U106 as long as it behaves as a single entity. Exactly this is what the various branches of U106 put together have NOT done and therefore one cannot treat them as one big group.

I get what you are saying and I agree it is a problem. The essence of the problem is that all of the U106 people in a single geography are not related to a single MRCA (Most Recent Common Ancestor) when compared to another geography. There was may have been and probably were exchanges between the geographies as well as interventions of U106 people from other geographies not in the comparison.

Exactly. And that's precisely what happened with the U106 subtree. Consider how U106 entered the British Isles. It did so on multiple occasions (more than one or two). There were different groups on different moments in time with a different mix of sub-clades that entered the Isles. All taking along a certain unknown variance. It therefore appears to me, talking about variance of a whole sub-tree that entered the Isles, or the remnants of it that survived to date, is relatively meaningless. "The variance of U106 in Ireland/Scotland/Wales" is a meaningless figure, as far as I'm concerned !!

Sometimes I think we are searching for the silver bullet. There is no such thing. There are many bullets, some misleading, some right on, and then there are shotgun shells and everything other kind of thing. We must look at them all individually and together.

I'm sure these questions are solvable, but not in this way. The Genographic Project once showed diagrams on their web site describing M343 entering Europe. I haven't heard anybody describing things in those terms ever since. Now people are describing migrations in terms of U106. That's not going to work either. The granularity is wrong.

If there's something I learned from R-Z18, then it's that these type of investigation should be done on the LOWEST level possible; not the highest. If you look at the lowest level, you'll start to see meaningful clusters and the geographical spread of these clusters will start to tell stories. Afterwards, one could compose, say, a U106 story by taking together the stories of the sub-clades. But this story will only be relevant for a relatively short period in time after U106 emerged. After that, it rapidly got far too diverse to be discussed as a single entity.

[gtc]

If there's something I learned from R-Z18, then it's that these type of investigation should be done on the LOWEST level possible; not the highest. If you look at the lowest level, you'll start to see meaningful clusters and the geographical spread of these clusters will start to tell stories. Afterwards, one could compose, say, a U106 story by taking together the stories of the sub-clades. But this story will only be relevant for a relatively short period in time after U106 emerged. After that, it rapidly got far too diverse to be discussed as a single entity.

An excellent point.

(edit) ... but the branch needs to be pretty long/deep (i.e. recent) in order for the clusters to truly "cluster" in a meaningful way.

[Richard Rocca]

Peter, to your point and as I've mentioned before, some variance numbers are skewed because of a greater or lesser number of migrations into and out of a region that may or may not have occurred! To that point, Italy seems to not have been impacted by a major L48+ migration because unlike in most areas of Europe, L48 is much less common. We tested the following L48- kits a few months ago and all are now U106+ L48- Z156-   

179540   Giacomo Zeni, b. 1630 Verona, Italy
141915   Berardo Cesaroni, b.c. 1575, Cartoceto, Pesaro and Urbina, Italy
N12646   Veturio Cesaroni, b.1880, Terni, Umbria, Italy