Archiver > Y-DNA-HAPLOGROUP-I > 2008-03 > 1205856226

From: "Ken Nordtvedt" <>
Subject: [yDNAhgI] New y Haplogroup I Paper
Date: Tue, 18 Mar 2008 10:03:46 -0600

A paper "New Phylogenetic Relationships for Y-chromosome Haplogroup I: Reappraising its Phylogeography and Prehistory" has appeared as a book chapter in late 2007. Authors are Underhill, Myres, Rootsi, Chow, Lin, Otillar, King, Zhivotovsky, Balanovsky, Pshenichnov, Ritchie, Cavalli-Sforza, Kivisild, Villems, Woodward. I only have hard copy and someone asked for a summary. I decided that I could only invest in a summary if it was a commentary as well as summary.

I start by discussing their own mentioned "seven main topics".

1. They find three "new" bifurcating SNPs within Haplogroup I. Two of these are actually identical with the SNPs S31 and S33 which Ethnoancestry brought to market a couple years ago. M438(S31) unites all of I(xI1a), while M436(S33) unites I1b2* with I1b2a (M223+). Another way of stating the function of M436(S33) is that it separates I1b2* (the haplogroup with 10 at DYS455 and 12 at DYS454...) from I1b1 and I1b*. (I will use the present ISOGG haplogroup names here) The really new SNP is M423. Their paper finds I1b1-Dinaric of Eastern Europe positive for M423, but the M26+ I1b1a-Sardinia subhaplogroup being negative. So M26+ branched off early enough in the I-Tree to escape the later M423+ mutation. A number of people on this list are presently testing M423 at Ethnoancestry to see how the I1b1-Isles and I1b1-Western clades are for M423. My expectation is that Western will be found negative and Isles could go either way, although probably positive.

2. The formally revise the nomenclature for haplogroup I in basically the same way as Hammer and FTDNA will soon do. Most of the renaming is due to P38 being demoted to equivalence with P19 and M258 as defining all of I. Two haplotypes which were purported to break that equivalence have disappeared for various reasons. I suspect that a deep study of haplogroup I in Turkey and beyond could very well resurrect P38 as non-equivalent with P19 and M258, requiring a Tree revision in the reverse direction in the future.

3. They reinterpret some historic scenarios of the origins of sub-clades. Their main revision seems to concern I1a. "The availability of new and more comprehensive ... results from Western Europe (including England, Ireland, Denmark) ... with larger sample sizes from Germany and Netherlands allows us to reconsider the suggestion that I1a and M223+ dispersed from France (They actually promoted southern France in 2004). The new data supports a slight adjustment of .... the initial expansion zone .... more toward Denmark."

4. They report a "new" SNP which unites y haplogroups I and J after separation from parent F haplogroup and before I and J obtain their separate defining SNPs. Ethnoancestry has brought to market two such SNPs a couple years ago. No "IJ"(xI and xJ) haplotypes have ever been found. This intermediate haplogroup could be extinct.

5. They discuss possible source areas for "I" and for various sub-groups. Origin of "I" is suggested as southeast Europe which I tend to agree with, although I'd put it as the place where "I" began and quickly broke up into its branch lines which eventually (many thousands of years later) lead to the MRCAs for the various sub-populations in I we see today.

6. They add new data for UK, Denmark, and Germany which they say will permit evaluation of the potential role of Doggerland (now under the North Sea) landbridge to the Isles. Remarkably they did not cover Scotland in the new paper. As a result their 200 or so UK dna samples again lead them to not find any M284+ examples. The 2004 paper did not sample the UK and also missed M284+. M284+ is actually a very robust haplogroup from the UK and primarily Scotland. Ethnoancestry has been testing for M284 for a couple years and is finding large numbers of positives whenever suspect haplotypes, easily recognizable, are sent to them.

7. Their final point I don't understand. It deals with pre-Greek toponyms in the Palaeolithic heritage.......

The paper makes some age estimates for sub-populations within y haplogroup I, using the variance method. But they continue to use the so-called "effective" mutation rates for their STRs and as a result get age estimates which are mostly much too large in my view.

Using the concept that a haplogroup will show greatest diversity where it has been the longest (origin), they attempt to locate such places. However, their sample sizes become very tiny, and I suspect they end up over interpreting their results.

I'm sure I missed some interesting points being made in the paper. Hopefully it will be available online one of these days. The publisher of the book in which the paper appears as a chapter wants to sell books, so online access may not come to pass soon, however. I have not yet seen any supplementary data sheets for their enlarged collection of haplotypes. They use more STR markers for some of them than in the original "I" paper of 2004.

They have some maps showing the distribution of I1a, I1b1-Dinaric M423+, and M223+ in Europe. But I don't care for the interpolating software that jumps oceans and mountains in ascribing distribution and which uses darkness levels in suspect ways. Nevertheless, maps are maps and usually useful to look at. There is really nothing seen in these distributions, however, that we probably did not already know from the many other maps in various papers.


This thread: