Y-DNA-HAPLOGROUP-I-L Archives

Archiver > Y-DNA-HAPLOGROUP-I > 2007-08 > 1188356175


From: "Ken Nordtvedt" <>
Subject: [DNA-HG-I] Haplogroup I in Europe
Date: Tue, 28 Aug 2007 20:56:15 -0600


The appearance of a large number of different clades and sub-haplogroups of y-haplogroup I in Europe, coupled with the great separations between the founding or modal haplotypes of these clades indicates to me that y-haplogroup I has been throughout Europe at least as far back as 14,000 years or so. I see no evidence that R1b1c* has been in Europe as long, although it may have been. To follow my discussion of the very surprising nature of the I haplotypes in Europe you should look at and perhaps print out a copy of the FoundersTree file at my website http://knordtvedt.home.bresnan.net

The Tree was created by first finding the modal or founding haplotypes for the more robust clades found in y-haplogroup I. 52 of the slowest mutating markers were used from FTDNA and SMGF databases to produce the haplotypes. Some of these clades are defined by various SNPs; other divisions in the Tree have no SNPs yet discovered but are nevertheless well established by unique modalities for the separated haplotype clusters. The most likely number of generations between each pair of clade founding (modal) haplotypes is then determined, and all these "genetic distances" between clade founders makes up our GD matrix. The Tree is a realization of that GD matrix; if you start from any one clade founder indicated by one of the arrowheads and move through the Tree to any other clade founder indicated by another of the arrowheads, the "genetic distance" in generations between those founders can be determined.

Different major clades in y-haplogroup I tend to be separated from each other by about 550 or 600 generations. If we split the difference, that means it is typically about 275 generations further into the past when different clades of haplogroup I had their most recent common ancestors (MRCA). That's 8000 years. Looking at the FoundersTree, that 8000 years is approximately the elapsed time from the location of SNP S31 which occured at the original split up of clade I1* into lines of descendancy leading to I1a, I1b*, I1b1, and I1b2.

These various clades do not have their founder or MRCA until the various arrowheads you see --- typically 8000 years later, more or less, with a couple early interesting exceptions. What happens at these arrowheads? There each clade experiences a star-like explosion of growing variance right up until today. The sub-clades that can be discovered today within those various star-like explosions of variance have much smaller genetic distances between their founders than are these enormous genetic distances between the different major clades of y-haplogroup I.

I think there are two different eras of mankind in Europe being revealed here --- the pre-agricultural, hunter-gatherer era of precarious, almost zero rate population growth, and then the arrival of agriculture.

All these major clades of y-haplogroup I shown in the Tree are European. They generally have different geographical distributions, some differences slight and some very clear. Together they cover all of Europe. Either they (and there are many of them!) moved into Europe individually from the Mid-East or other more eastern location, or their common parent haplogroup I1* was already in Europe before the MRCA of all these clades. If the latter is the case, and I think that most reasonable, then we must add the 8000 years of elapsed time between that initial breakup until the typical arrowheads came on the scene, i.e. the individual clade MRCAs or founders. If we assume the typical clade in haplogroup I such as I1a, for example, is 6000 years old, then I conclude that haplogroup I has probably been in Europe for at least 14,000 years. (Incidently and ironically, we know of no example of an I1* haplotype right now.)

I have seen phylogenetic trees produced for large databases of all individual R1b1c haplotypes. It ends up being one of these star-like explosions of variance, much like what results when the entire database of I1a haplotypes are simiilarly treated. And variance estimates for R1b1c typically come out to be about the same size as I1a variances, or most of the other major I clade variances. There is no great partition of R1b1c into sub-clades separated by several hundred generations of genetic distance as occurs with haplogroup I. There is an Eastern type of R1b, and it would be interesting to know the genetic distance between its founder and that of R1b1c? But we don't know whether that Eastern R1b type is predominately European, Anatolian, or what? If we could estimate the time back to the MRCA between R1b1c and this Eastern R1b, we may still not be able to say with much confidence where that MRCA lived.

Let me describe my interpretation of the particular line of 270 generations of ancestors leading to the I1a MRCA, stretching for 7000 years or so from the MRCA for all of I. It's an amazing picture if one goes where the evidence leads. Life was precarious. The extinction probability for any male's ydna was extremely high (close to 1).
The consequence of an extinction probability extremely close to one is the probability that many generations will go between males in a line who have more than a single son with descendants today. For 270 generations or so father had single son with surviving descendants today, and this son had single son with surviving descendants today, over and over until it finally led to the I1a founder or MRCA. Then descendant populations started to explode, relatively speaking, on up until today. But those 270 generations were not void of producing interesting consequences for the present. First of all, STR mutations continued to accumulate through those generations, so that the eventual I1a founder at the appropriate arrowhead had a largely different haplotype than did the MRCA for all of I 270 generations previously. Secondly, during those 270 generations of father to son transmissions, SNPs could accumulate. If we use a recent conservative estimate of SNP mutation rates, we should have had about 100 SNP mutations on this line which would differentiate I1a from the rest of haplogroup I. Our lab friends have already found 9 of them, and they are listed on my Tree. There are probably many more to be discovered. Where are these 9 SNPs located on that line? They each can be anywhere in those 270 generations (and the same can be said for where I1a's original DYS455 = 8 occured). We will never know unless there were a few of those 270 generations of males who had a second son with descendants today. We know at a pretty good level of precision that any undetected clades emerging from this 270 generation line must be demographically very tiny or hiding where we have not yet looked very thoroughly (like Anatolia, Cacasus, etc.), because we have not found any haplotypes positive for some of these 9 SNPs and negative for others. But we keep looking.

A very similar story could be told about all the other long lines in my FoundersTree. Look at the I1b1-Western line, for example. What a ripe clade for which an SNP should be relatively easy to find --- All those generations since it split from all other clades!

Note: real father/son mutation rates as estimated by Chandler were used for my 52 slow mutating markers. Actually, only the sum of all 52 rates was needed.
No fudge factors here.

This message has gotten maybe too long for our list. So interesting stories about particular clades which I think reinforce my scenario will have to be put off for another day.

Ken



This thread: