R-P312(S116) Variance: Round 2

[Mikewww]

I have a spreadsheet of all the P312 and subclades people I can find in our DNA projects. It has been retooled to exploit Excel 2010 and more RAM. I can easily scrape relevant (downstream only) SNPs from project screens and re-create short haplogroup labels from Krahn's draft tree now.  

I don't do any selecting/filtering of haplotypes to try to create a representative sample.  I just take whatever I can find and calculate straight Sum of the Variance. I do look up Ysearch IDs and try to decipher MDKA location information as best as possible to classify to the Country/Sub-Region and Dept/County/Shire level.

Based on various discussions on STRs, I throw out all the multi-copy STRs and 425 because it can be null.

Also I do a run where I throw out all but the slowest 35 of the first 67 from, all but 25 from 37 length haplotypes and all but 17 from 25 length haplotypes.

I won't display the actual Sum of the Variance numbers themselves but I'll divide by a base so that everything is shown in relative terms with 1.0 as equivalent to the base. This just makes the numbers easier to compare.

I have found that the relative positioning tends to be erratic for haplotype groups of less than 15 or so.

All of the haplotypes are posted at the P312 and L21 Yahoo groups in the FILES section.

[Mikewww]

Here are the major subclades and their relative variance. This is only the 35 slowest non-multi-copy/non-null out of the first 67.

The GD's are at all 67 markers and include using infinite allele methods on the multi-copy markers and for null. I even try to adjust for recLOH in the GD's but they are not perfect.

R-U152 All_____: Var=1.13 @Slow35; AvgGD=15, MaxGD=30 @67 (N=389)
R-U152/L2______: Var=1.03 @Slow35; AvgGD=16, MaxGD=31 @67 (N=201)
R-U152/L2/L20__: Var=1.12 @Slow35; AvgGD=14, MaxGD=21 @67 (N=42)

R-Z196 All_____: Var=1.07 @Slow35; AvgGD=14, MaxGD=25 @67 (N=180)
R-Z196/L176.2__: Var=1.02 @Slow35; AvgGD=13, MaxGD=25 @67 (N=162)
R-L176#/SRY2627: Var=0.95 @Slow35; AvgGD=13, MaxGD=27 @67 (N=140) [# under Z196]
R-Z196/M153____: Var=0.34 @Slow35; AvgGD=6, MaxGD=11 @67 (N=5)

R-L238_________: Var=0.48 @Slow35; AvgGD=7, MaxGD=9 @67 (N=5)

R-L21 All______: Var=0.91 @Slow35; AvgGD=15, MaxGD=27 @67 (N=1800)
R-L21/L513_____: Var=0.65 @Slow35; AvgGD=11, MaxGD=25 @67 (N=77)
R-L21/P314.2___: Var=0.59 @Slow35; AvgGD=9, MaxGD=17 @67 (N=20)
R-L21/L513/L193: Var=0.41 @Slow35; AvgGD=6, MaxGD=14 @67 (N=42)
R-L21/M222_____: Var=0.40 @Slow35; AvgGD=7, MaxGD=16 @67 (N=371)
R-L21/L144_____: Var=0.40 @Slow35; AvgGD=7, MaxGD=22 @67 (N=12)
R-L21/L159.2___: Var=0.34 @Slow35; AvgGD=7, MaxGD=25 @67 (N=68)
R-L21/L226_____: Var=0.24 @Slow35; AvgGD=6, MaxGD=11 @67 (N=62)
R-L21/L371_____: Var=0.17 @Slow35; AvgGD=5, MaxGD=10 @67 (N=5)

We'll go into Z196 later, hopefully when more results are available.  I'll run through U152 and L21 by Old World geography.  The numbers start to shrink so I'll have to use some shorter haplotype calculations.

[Mikewww]

I have about 4200 P312 haplotypes now with a large number of them being 67 STRs or more.

Here is the country/region view of R-P312 All. This includes all subclades.

67 STR Deep Clade Tested P312 haplotypes only. Relative variance for non-multicopy markers. GD's are over 67 to the modal for each subgroup.

France_________: Var=1.14 @1copy50; AvgGD=15, MaxGD=27 @67 (N=117)
East/Cent Euro_: Var=1.04 @1copy50; AvgGD=17, MaxGD=30 @67 (N=79)
England________: Var=1.00 @1copy50; AvgGD=15, MaxGD=28 @67 (N=420)
Low Countries__: Var=0.99 @1copy50; AvgGD=14, MaxGD=23 @67 (N=31)
Germany________: Var=0.98 @1copy50; AvgGD=15, MaxGD=25 @67 (N=127)
Iberia_________: Var=0.98 @1copy50; AvgGD=15, MaxGD=27 @67 (N=87)
Ireland________: Var=0.98 @1copy50; AvgGD=15, MaxGD=27 @67 (N=727)
Italy__________: Var=0.93 @1copy50; AvgGD=15, MaxGD=22 @67 (N=36)
Scotland_______: Var=0.93 @1copy50; AvgGD=15, MaxGD=26 @67 (N=362)
Nordic Area____: Var=0.90 @1copy50; AvgGD=14, MaxGD=24 @67 (N=54)

The Low Countries are Benelux. East/Cent Euro is anything east and southeast of Germany, Austria and North Italy.

I find the East/Central Europe numbers interesting when you start looking at the GD's. It is a wide area so maybe it is not a fair comparison.

[Mikewww]

We have quite a few Z196 results so I feel that we can reliably show relative variance for it. The below is based on the slowest 35 markers out of the first 67.  The sample sizes are pretty high except as noted.

Z196 looks to be every bit as old U152. L21 is the younger brother.


Z196____________ 1.14
_L176.2(S179.2)_ 1.06
__SRY2627(M167)_ 0.95
__L165(S68)_____ 0.64
_M153___________ 0.33
U152(S28)_______ 1.12
_L2(S139)_______ 1.00
__L20___________ 1.07
L21(S145)_______ 0.91
_L513(DF1)______ 0.67
__L193(S176)____ 0.44
_P314.2_________ 0.59
_DF21___________ 0.56
_M222___________ 0.40
_L159.2(S169.2)_ 0.35
_L226(S68)______ 0.22
L238(S182)______ 0.40
P312*(S116*)____ 1.03

Notes: Sample size of only 5 for M153 and only 7 for DF21

[alan trowel hands.]

We have quite a few Z196 results so I feel that we can reliably show relative variance for it. The below is based on the slowest 35 markers out of the first 67.  The sample sizes are pretty high except as noted.

Z196 looks to be every bit as old U152. L21 is the younger brother.


Z196____________ 1.14
_L176.2(S179.2)_ 1.06
__SRY2627(M167)_ 0.95
__L165(S68)_____ 0.64
_M153___________ 0.33
U152(S28)_______ 1.12
_L2(S139)_______ 1.00
__L20___________ 1.07
L21(S145)_______ 0.91
_L513(DF1)______ 0.67
__L193(S176)____ 0.44
_P314.2_________ 0.59
_DF21___________ 0.56
_M222___________ 0.40
_L159.2(S169.2)_ 0.35
_L226(S68)______ 0.22
L238(S182)______ 0.40
P312*(S116*)____ 1.03

Notes: Sample size of only 5 for M153 and only 7 for DF21

Been looking that way for a while.  L21 seems to take off at the end of the early burst stemming from Z196 and U152.  Or you could look at it an one continuous period of early expansion by S116 with L21 one of the last manifestations of this first burst.

I would hazard a guess that there would be a correlation between variance of clade and how westerly the average longitude of ancestral origins for that clade (down to time of L21).  It certainly seems that way given that L21 and SRY 2627 are the youngest of the group of early S116 clades.  Kind of supports the notion that the S116 movement moved from east to west/north-west.  Not clinching the deal maybe but suggestive.  

I dont think that would tie in with any simple beaker origin model as the chronological/geographical sequence doesnt fit the sequence of spread of beakers as suggested by RC dates.  It also is hard to fit into a pure LBK based early Neolithic model  Might fit the middle Neolithic model of Myres though. 

[GoldenHind]

We have quite a few Z196 results so I feel that we can reliably show relative variance for it. The below is based on the slowest 35 markers out of the first 67.  The sample sizes are pretty high except as noted.

Z196 looks to be every bit as old U152. L21 is the younger brother.


Z196____________ 1.14
_L176.2(S179.2)_ 1.06
__SRY2627(M167)_ 0.95
__L165(S68)_____ 0.64
_M153___________ 0.33
U152(S28)_______ 1.12
_L2(S139)_______ 1.00
__L20___________ 1.07
L21(S145)_______ 0.91
_L513(DF1)______ 0.67
__L193(S176)____ 0.44
_P314.2_________ 0.59
_DF21___________ 0.56
_M222___________ 0.40
_L159.2(S169.2)_ 0.35
_L226(S68)______ 0.22
L238(S182)______ 0.40
P312*(S116*)____ 1.03

Notes: Sample size of only 5 for M153 and only 7 for DF21

I find it very interesting that M222 and L238 have the same variance.  Due to its very strong STR signature, I have thought L238 was relatively young, perhaps as young as M222. This is also one of the reasons why I believe there is an as yet undiscovered SNP (or possibly more ) between P312 and 238, just as there at least one (L21) and probably more between P312 and M222.

[Mikewww]

I find it very interesting that M222 and L238 have the same variance.  Due to its very strong STR signature, I have thought L238 was relatively young, perhaps as young as M222. This is also one of the reasons why I believe there is an as yet undiscovered SNP (or possibly more ) between P312 and 238, just as there at least one (L21) and probably more between P312 and M222.

M222 has a very strong signature too. It's also quite hard to break it down further into clusters. To me its an amazing story.. a lot of people closely related. Something big was going on in Ireland while the Anglo-Saxons were mucking around in Britain.

L238 and M222 probably both have multiple SNP's between them and their current next levels up, but I think that is likely in most of these subclades.