Page 1 of 1

Y-DNA Haplogroup G Project

PostPosted: Tue Apr 10, 2012 2:59 am
by soulblighter
The haplogroup G project is run by Ray B, with more than 3000 samples(to date) from FTDNA.
FTDNA Project Website:
Group project website:
Monthly Newsletter:

Re: Y-DNA Haplogroup G Project

PostPosted: Mon Oct 22, 2012 6:57 am
by Palisto
soulblighter wrote:The haplogroup G project is run by Ray B, with more than 3000 samples(to date) from FTDNA.
FTDNA Project Website:
Group project website:
Monthly Newsletter:

I took all STR111 individuals from this and all other projects to create a tree of haplogroup G (N=139).
More details here: ... -tree.html


Re: Y-DNA Haplogroup G Project

PostPosted: Sun Sep 29, 2013 6:21 pm
by palamede
Last letter of Ray Banks today with very interesting subjects : ... essages/12

Yesterday, I got my result CTS4803+
I expected this result because all tested Z726+ are CTS4803+ except Wadsworth and I am near of CTS4803 modal for STRs.

3 years, I tried a classification of G2a3b1-dys388=13 by STRs with some difficulties to find caracterized STR values.
But for Wadsworth I found a group with
439(I)=13 389u(L)=30 459b(O)=10 449(U)>31 H4(AA)=10 576(AF)=17

At this time, the members of the group were :
Gahm, Roscoe, 2 Morrow , Townsend, Conley, 2 Lawrence, 2 Wadsworth.
Every members of the group (except the 2 Wadsworth) were also 385a(E)=13

Re: Y-DNA Haplogroup G Project

PostPosted: Sat Dec 14, 2013 1:18 pm
by palamede
Yesterday, Ray Banks published 2 texts in facebook

The first text :

13 Dec 2013
The G Tree at ISOGG which is the unofficial clearinghouse for proven Y categories is now changed dramatically
This is due to a consensus arising that well vetted SNPs from whole sequencing could be allowed. I am the first to take advantage of this. And this will result in the addition in total of perhaps 700-800 SNPs to the G tree, now the one with the most SNPs listed. Yet to be added are the SNPs for G2, G2a and general G which are actually the largest in number. One will note that there are new subgroups listed under L91, L166, P16 and what used to be called L177. There are better SNPs available and L177 and P18 have been withdrawn.

For the newletters, I don't see a newletter of this month. In more, Ray seemed to do a mix-up between newletters of haplogroup G and haplogroup T.

The second text :

just posted this simultaneously in the ISOGG closed discussion group:

New genetic findings regarding the 5300-year-old Iceman mummy, Oetzi.

Studies earlier this year supplied data suggesting two different calculations for how many years are represented by each SNP mutation in the higher quality, mappable areas of Y-DNA. I have come across new data that takes sides with one of the two studies. Oetzi the Iceman is perhaps the best test for checking the validity of the proposed dating because he is reliably dated to 5300 yrs by carbon dating and was the only ancient person ever to have full Y sequencing.

Francalacci in a sequencing study of Sardinia suggested that a SNP mutation in the described chromsome area occurs every 205 yrs +/- 50 yrs. Poznik in a study of a much smaller number, but more geographically diverse, men did not state such a figure for his men, but it can be deduced it was probably half the number. Francalacci calibrated his calculations to the first settlements in Sardinia, and Poznik to the first peopling of the Americas.

Oetzi was reported in a study as belonging to the G-L91 subgroup. But the raw data available in his public data file (BAM file) indicates he is positive more specifically for the G-L166 mutation which characterizes L91’s subgroup.

We received this week high-coverage (50x) YDNA sequencing of the sample of the man of German ancestry in whom L166 was first discovered. This was ordered through Full Genomes Corp. There are other L166 sequenced samples available for comparison, both Asian and European, and – of course – Oetzi’s.

Oetzi’s sample is not perfect, and that problem makes it difficult to put together a study involving my findings for publication. The sample is also unfortunately in the old Build36 version of the chromosome location mapping, and that makes it difficult to simultaneously compare with all the other samples, a process which can improve the quality of reads for him and possibly bring out some additional positive results. There are about 600 or so known shared G SNPs earlier than the L166 subgroup which Oetzi would be expected to share, but he has an unusual number of “no reads” comparatively. Oetzi was sequenced with low-coverage sequencing, but others with similar sequencing had fewer no reads for expected mutations.

In this context, the newly sequenced L166 has 36 quality SNPs unreported in other L166 men. He is the closest man so far to Oetzi, sharing with him the only L166 SNP negative in the Asian L166. There may be about 10 L166+ men in the Francalacci data. L166 results were not reported for them, but their status can be inferred by the constant sharing of SNPs with proven L166 men. These 10 were not reported also for FGC5672 which is negative in the Asian contingent and present in the men just tested and in Oetzi.

So multiplying the 36 SNPs from the newly tested man that are unique to him by the 200 yrs suggested by Francalacci, we come up with 7200 yrs to the nearest shared common ancestor, which would be Oetzi. Poznik, on the other hand, would indicate this calculates to about 3600 yrs. The latter clearly cannot be the case because Oetzi is 5300 yrs old.

Oetzi could share a few more SNPs with the newly tested man which could not be read. That would reduce Francalacci’s time calculation to perhaps 6800-7000 yrs ago to the shared male ancestor with Oetzi and change Poznik’s to perhaps 3300 to 1500 yrs ago. This makes Poznik’s time calculation less tenable.

All this just suggests that the 200 yrs per SNP listed in Francalacci is the most valid. Without more good reads and higher coverage sequencing of Oetzi, my observations do not prove anything, just suggest it.

In our prepublication paper involving 1000 Genomes Project samples and a few other samples, ... t.pdf+html
we point out that the Amazonian Indian sample has an unusually large number of these unshared (singleton) SNPs. This might indicate some branches of Ameridians have ancestors that predate the peopling of the Americas. Poznik wrote: “Because the M3 mutation appears to be specific to the Americas (20), it likely occurred after the initial entry.” He used only two Maya Indian samples to make this calculation. If the M3 divergence occurred earlier, Poznik would have to reset his yrs/SNP calculation closer to Francalacci’s.

Francalacci’s dating at 200 yrs/SNP would place the emergence of the G branch from F (the first divergence within the G to T haplogroups) at perhaps 100,000 yrs ago. Various researchers initially were listing the age of G’s divergence at abt 21,000 yrs ago based on the formula they use as applied to the number of STR marker mutations, and a calculation in the earlier study (in sync with Poznik) suggested the divergence of the G through T haplogroups from F took place about 50,000 yrs ago.

I personally have noted some inconsistencies in the number of SNPs per haplogroup in the 1000 Genomes data. This has not been fully explored. At times these haplogroups were apparently nothing more than a small family group existing for millennia. Perhaps longer haplogroup-specific longevity of males or a culture that allowed older men to have more wives might be involved. Or one haplogroup’s ancestors camped on a uranium deposit? So this time calculation does not seem a simple problem to me.

Ray Banks ... t.pdf+html