Fluxus Network Diagrams for K Subclades

From MitoSearch at 1000 Entries


The five charts discussed below were produced using Fluxus-Engineering Network software. The more complicated ones were run through the Reduced Median then the Median Joining algorithms. In each chart, KROOT represents the ancestral K haplotype with HVR mutations 16224C, 16311C, 16519C, 73G, 263G and 315.1C. If one of those six mutations appears otherwise on a chart, that would represent a back mutation. If any other mutation appears more than once between KROOT and the end node, that would also represent a back mutation. However, the latter type of back mutation may just represent a placement decision by Fluxus which may or may not reflect reality. In the interest of sanity, I have not marked these back mutations on the charts; but I have discussed many of them below. The six charts include one each for three of the major K subclades, K1b, K1c and K2, plus two from the large K1a subclade. Most subclades of K1a, which is about 60% of K, may only be determined from coding-region mutations, not from HVR mutations. All of the position 524 insertions have been normalized to the most common FTDNA 524.1C, 524.2A, etc., scheme.


K1a plus 195C: This chart includes all K haplotypes with 497T, the defining mutation for the large subclade K1a, plus 195C. This group not only includes the Ashkenazi subclade K1a9 defined by 16524G, but also a large cluster defined by 16048A, to which I've given the temporary name K1a10. In addition, I've given the temporary names Pre-K1a10 and Pre-K1a9 to clusters of haplotypes with and without pairs of 524 insertions. K1a10 has 524 insertions, while K1a9 does not. For this chart all mutations are used. The position 309 insertions are important since they appear in Pre-K1a9 and Pre-K1a10, but never in K1a9 or K1a10. Although this K1a/195C group was only 8.5% of the sequences upon which Dr. Doron Behar's K chart was based, it included 19% of the K entries on MitoSearch.


Beginning at KROOT, the line goes up via 497T and 195C to the first large node QXGNG which represents five matching sequences with the modal value of Pre-K1a9. There are 10 other sequences in a mostly star pattern around that node. As I have said before, I'm not sure the ones with 114T belong here, since that mutation is usually found in the K1a1+ lines. Note that all the remaining sequences follow the line via 16093C to branching point mv1 and beyond. That mutation is a heteroplasmic one, often called a hotspot, so the diagram may or may not reflect reality. The branch up from mv1 is K1a9, defined by 16524G. The modal node G987N is the largest on the chart, representing 16 sequences. The two-sequence node PAKP6 shows a back mutation on 16093C. Behar's K chart shows 16093C below 16524G and K1a9. Both sequence orders may have actually occurred. If K1a9 appears to branch off the main line of the chart, it also branched off geographically, since it is found mostly in Eastern Europe. The other three main groups are found mainly in Western Europe, even K1a9’s "parent" Pre-K1a9.


The line to the right from mv1 moves through the first pair of position 524 insertions to 3PH3A, the beginning of Pre-K1a10. This cluster includes five sequences in a star pattern, three more at 2KJR3 which have dropped (or never had) 16093C, then a couple more off that node. From 3PH3A past a second pair of 524 insertions leads to AGZK3, which has branches to many more in Pre-K1a9. The large four-sequence node NGKZG and five other singletons are without 16093C.


From AGZK3 to mv4 and beyond is K1a10, defined by 16048A. The first branch up from mv4 is via two new mutations and a dropped 16093C to a three-sequence node 4WAZU. But it's from mv4 via 16291T which leads to the 11-sequence node 4FK85, the modal for K1a10. Especially interesting is that Fluxus has shown the sequences in this cluster with only two pairs of 524s - the three-sequence GAEWJ and 3CD9P - as back mutations from the larger two-pair modal node. I've suggested before that this might have been the proper order of events, but I'm a little surprised that I was able - not on purpose - to produce a diagram representing it. The solution was to include K1a10 as a branch of Pre-K1a10, which, of course, is part of the theory I've suggested. In other words, the original 16048A mutation occurred in a person in Pre-K1a10 who already had two pairs of 524s. The most distant node KHSKB, with a third pair of 524s, is the record-holder in K with 19 differences from the CRS, including six non-heteroplasmic mutations since the founding of K.


K1a1b1a: A vastly less complicated chart than the one above, this chart is of the largest Ashkenazi subclade defined by 16234T. The modal node R369N also includes 114T. A second large node picks up 16223T. One sequence has a rare back mutation on 16519C. Note that none of the Ashkenazi subclades have 524 insertions.


K1b: K1b is defined by a coding-region mutation, but two of its lower subclades may be identified by HVR mutations. From KROOT directly to the right and to the lower right of mv4, there are several sequences which do not have defining HVR mutations. These are shown as K1b+, but most are probably K1b1. Note that all levels of K1b have sequences with and without 524 insertions. Up from mv4 is K1b1a, defined by 16319A and 152C. Most of those also have 16463G. There are sequences with and without 16093C. Also, many have back mutations at position 524, or never had them. To the upper right of mv4 is K1b2, defined by 146C and 195C. This subclade has a second large node with 16129A, but without 524s – shown here by back mutations. K1b2 has the greatest number of variants in K of the 524 insertion pairs, from zero to three pairs.


K1c: With another simple chart, K1c is defined mainly by 498-, a deletion only found in haplogroups K and L0. All have 146C and all but one has 152C, which in reality is probably a back mutation for CGD4E. From the modal node K4EKU, there is a classic star pattern. Most of these are probably in K1c1 and lower subclades, but those are defined by coding-region mutations. To the right, K1c2 is defined by 16320T, again with a large modal node and a star pattern. One sequence has a rare back mutation on the key K mutation 16224C. Note that none of the sequences here has 524 insertions, although two related persons in the K Project do have them in an otherwise unusual sequence. In fact, node 7F7U3 has a pair of deletions at 522 and 523 which in effect represent “below zero” in the line of C nucleotides which include position 524.


K2: This subclade starts with 146C; to the right are many nodes in a tree-like structure. The group is marked K2+ because some of them may be in lower subclade K2b which is defined by several coding-region mutations. K2c is partly defined by HVR mutations, but none have shown up in MitoSearch or the K Project. K2a adds 152C, with the large node PYDTF surrounded by smaller nodes in a mostly star pattern. There are several with the 522 and 523 deletions discussed under K1c above. Also like K1c, there are no examples with 524 insertions. The group is marked K2a+, since are several lower subclades. Adding 512C leads to the third Ashkenazi subclade K2a2a. The descendants of the founder of that subclade mostly moved on to Eastern Europe with the other Ashkenazi subclades, while the others in the parent K2a subclades mostly stayed in Western Europe. K2a2a has little variation, with a few having the insignificant 309.1C, demonstrating the relative youth of the subclade.


©2007 William R. Hurst

Administrator, mtDNA Haplogroup K Project