AS OF MARCH 8, 2006


This is my fourth survey of the mtDNA haplogroup K entries on FamilyTreeDNA’s MitoSearch. The previous survey, at 300 entries, may be found at K300 Survey, which contains a link to the first two. The K300 survey featured a subclade chart, which was soon outdated by the January publication of Dr. Doron Behar’s paper which included a K subhaplogroup chart which seems to have become the standard His subgroups are based primarily on coding-region mutations beyond the control-region mutation in HVR1 and HVR2. Therefore, this edition will not discuss subclades. I should point out that MitoSearch, and therefore my surveys, have, for various reasons, a greater representation of British haplotypes than Behar’s paper.


The new feature of the K403 (I missed the 400 mark; the numbers quickly skipped from 399 to 403) survey is the publication of an Excel CHART, which is an extracted, sorted and annotated version of the high-resolution-only K entries from MitoSearch. After taking five months for the entries to go from 100 to 200, it took only three months to go from 200 to 300 and again from 300 to 400. Perhaps the leveling off of the rate of increase may be due to the holiday period involved.


Of the 403 total entries, I removed all of those which have tested only the low-resolution Hypervariable Region 1 (HVR1), leaving only those which have also tested HVR2. So the chart contains 182 entries, of which two are obvious duplicates. So 45% have high-resolution results, somewhat down from the almost 50% at 100 entries. A big factor in the decrease is the number of new entries through the National Geographic Society’s Genographic Project. That project only offers the HVR1 test, so those transferring their results to FTDNA have to take the mtDNARefine test to add HVR2. Many of those starting at FTDNA would have ordered originally the mtDNAPlus test which combines HVR1 and HVR2.


A few interesting facts may be gleaned from the full 403 entries. Nineteen of them claim to be “CRS,” or matching the Cambridge Reference Sequence from which all mtDNA results are measured. CRS results are usually found in haplogroup H, and occasionally in others, but not in K. None of those listed HVR2 results, so they do not appear on the chart. 192 list their HVR2 results as “Not tested,” while 29 say “No mutations.” I don’t think a real K would have no HVR2 mutations – three being the usual minimum – so all those were removed. 33 entries, 11 of which were removed, have a pedigree chart included in the entry.


Concentrating from this point on the 181 HVR1 + HVR2 high-resolution entries there are 133 different haplotypes, for a diversity percentage of 73%. That’s down from 78% in my last survey. There are two haplotypes with seven entries, two with six, two with five, one with four, four with three, and seven with two entries. All those numbers have gone up, which lowers the diversity. I won’t list the mutations for these common haplotypes this time, since they are easily found on the chart. There are 115 singletons or one-of-a-kind haplotypes.


I should first explain the headings shown on the chart. I have sorted the entries first by HVR2 then by HVR1. I think that sort order best shows the connections between the entries. So the first three columns are the MitoSearch number, HVR2, and HVR1. The next column is the country of origin as entered by the person tested. I have made no attempt this time to categorize the countries. The “Tested with” column shows that 173 tested with FTDNA. Three show Genographic Project. The previous survey had none from that project; I had been told that after the mtDNARefine test was taken by those starting with the project, the new result would be listed as from FTDNA. If the three only represent those who entered their results by “hand” or who added HVR2 results to an existing entry, the actual number who actually started through the Genographic Project is unknown. One each listed Sorenson, Oxford Ancestors, and DNA Heritage. I commented on those in my previous survey. There are also three from just “Other.” There are still none from Relative Genetics, although I have been informed that they have recently started giving haplogroup assignments to their mtDNA results. There are five K-looking results from Relative Genetics on MitoSearch under the Unknown haplogroup category.


There are also five entries under the U5 category which have perfect or near-perfect matches in K. Those all contain the HVR1 mutations common to K and the 16270T mutation which is common to U5. Three of those have HVR2 results, which are also consistent with K. In K, there are also five entries with 16270T; the three with HVR2 results are on the chart. I understand that FTDNA will eventually determine the status of these 10 entries by looking at their coding-region mutations. So K could gain or lose up to five entries.


Continuing the discussion of columns, if “Show” is listed under Pedigree, there is an uploaded GEDCOM for that entry. The last column “Notes,” is one I added. I mention the large haplotypes, mutations which seem to be missing because of typographical errors or back mutations, and the duplicates. Also noted are the haplotypes with the largest number of mutations – 16, with 12 in HVR2 – and the least number – 6. I marked the two large haplotypes which are identified as usually being of Ashkenazi origin in Dr. Behar’s paper. There are also seven odd entries near the bottom of the chart, which appear to be Ashkenazi, but barely look like K’s. They may have five back mutations. The ancestor of one of them married a rabbi, so my Ashkenazi guess might be correct.


Two entries, 9AJ6S and AN4E6, appear to be duplicates; but after a close reading of the entries’ comments, these may be triangulated haplotypes of descendants of two daughters of the same ancestor.


I have added colors to mark several different mutations. Blue in HVR2 marks 497T, which Dr. Behar uses to define subhaplogroup K1a. Yellow marks 16320T in HVR1 and 498- (a deletion at the 498 position) in HVR2; these mutations are often found in British K’s. Note that there is no overlap between 497T and 498-. Green marks the 16223T and 16234T mutations in HVR1, which Dr. Behar has identified as Ashkenzaki. Note that two of my “odd” haplotypes, as discussed above, have 16234T. Orange marks 16270T; those are the ones which match U5 haplotypes.


A new development since my last survey is the institution of haplogroup projects at FTDNA. I launched the mtDNA Haplogroup K Project on January 9, 2006. We now have 75 members. Although K is shown as less than 10% of the records in FTDNA’s database according to their pie chart, we are at last count the largest mtDNA haplogroup project. Please see our project website. For information on other mtDNA projects, etc., see World Families Network here. To join the K project or your appropriate mtDNA or Y-DNA project, use the Join button on your FTDNA personal page.


(Additional information relating to geographical considerations of the above is now available here.)


William R. Hurst