Definition of mtDNA Haplogroup I

[britzkrieg]

I am confused about the defining markers for mtDNA Haplogroup I (a.k.a. Clan Iris or Isha).

The Family Tree DNA site for this haplogroup says that they are:

16129A, 16223T, 16391A

However, I have just visited Bryan Sykes' website http://www.bloodoftheisles.net, and it seems that he lumps people without the 391 mutation into Haplogroup I. More confusingly, on Mitosearch.org, I see two people with the same HVR1 mutations in different haplogroups:

EW5C5 N* 129A,223T,519C
D66HX I 129A,223T,519C

Granted, the "I" individual had his/her HVR2 region tested, while the N* individual did not. Maybe the "I" individual's 391 mutation flipped back?

Still, there are no HVR2 results in the mtDNA PDF showing the profiles Sykes has collected from the British Isles. How do you know if someone is Hap I without either the 391 mutation or an HVR2 sequence? Does Sykes just define Hap I differently than others do?

Thanks in advance!

britzkrieg

[cacio]

britzkrieg:

Ian Logan has a page, with a I haplogroup tree:
http://www.ianlogan.co.uk/discussion/hap_N.htm

It seems all his I sequences have 16391, except one quirky one (number 25), which simply was not tested for HVR1 and HVR2 (it was tested only for the coding region) - so it is likely that it also has 16391.

As for FTDNA, they claim that at some point they started checking selected coding region mutations. So one logical possibility is that they did check one coding region mutation representative of I and found it in one of the sequences missing 16399. Or simply they made a mistake in assigning the haplogroup (it happens). But who knows, I think one has to ask them directly, they seldom say what they really tested.

Don't know what Sykes is doing, though presumably you are right, he's just lumping everybody together, I, N1 and what else.

cacio

[vineviz]

britzkrieg:

Ian Logan has a page, with a I haplogroup tree:
http://www.ianlogan.co.uk/discussion/hap_N.htm

Ian has a phylogenetic tree that shows the differences from CRS for haplogroup I:

http://www.ianlogan.co.uk/discussion/gifs/N_gif.htm

There's a good one at Argus as well:

http://www.argusbio.com/tools_docs/world_tree7.3.pdf

It looks like I is actually defined by 13780, in conjunction with other upstream SNPs.

Generally, HVR-I results in conjunction with the coding region test done by FTDNA is enough to get an accurate haplogroup assignment.

[vraatyah]

I am confused about the defining markers for mtDNA Haplogroup I (a.k.a. Clan Iris or Isha).

The Family Tree DNA site for this haplogroup says that they are:

16129A, 16223T, 16391A

However, I have just visited Bryan Sykes' website http://www.bloodoftheisles.net, and it seems that he lumps people without the 391 mutation into Haplogroup I. More confusingly, on Mitosearch.org, I see two people with the same HVR1 mutations in different haplogroups:

EW5C5 N* 129A,223T,519C
D66HX I 129A,223T,519C

Granted, the "I" individual had his/her HVR2 region tested, while the N* individual did not. Maybe the "I" individual's 391 mutation flipped back?

Still, there are no HVR2 results in the mtDNA PDF showing the profiles Sykes has collected from the British Isles. How do you know if someone is Hap I without either the 391 mutation or an HVR2 sequence? Does Sykes just define Hap I differently than others do?

You are correct. All sequences expected to fall into I (aka N1I in recent nomenclature) should be checked for coding region RFLPs, either 10237HphI or 10032AluI, particularly if 1) they lack the transition at 16391 required for "diagnosis" of I, or 2) the HVS2 sequence was not obtained.

In data published on Sykes' website http://www.bloodoftheisles.net
99 sequences were classified as I without any information about whether the samples were ckecked for I with RFLPs. 22 out of 99 lack 16391, although the authors sequenced the HVS1 region between 16024 and 16400. Most sequences have mutations 129-223 only, so should be considered ambiguous, although in Europe such variants usually fall into I. However, some types without 16391 are firmly in I, for instance

129-172-223-311,

and the contrary, these ones were obviously misclassified:

#3387 223-278-318-390 (typical L2c1)
#A8291 223-278-318-390 (typical L2c1)
#A8226 223-278-318-390 (typical L2c1)
#A2756 129-223-291-298 (likely "Gypsy" M5a)

Valery

[vraatyah]

There is another possibility: 16391 was overlooked at least in several samples, due to some damage of the forward trace border.

[vineviz]

In data published on Sykes' website http://www.bloodoftheisles.net
99 sequences were classified as I without any information about whether the samples were ckecked for I with RFLPs. 22 out of 99 lack 16391, although the authors sequenced the HVS1 region between 16024 and 16400.

I'm not aware of any commercial labs that routinely use RFLP to diagnose mtDNA haplogroups except for FTDNA.

Oxford Ancestors mtDNA test (the one that Sykes used) certainly does not sample the coding region at all and includes only a small subset of the control region.

[vraatyah]

I'm not aware of any commercial labs that routinely use RFLP to diagnose mtDNA haplogroups except for FTDNA.

Oxford Ancestors mtDNA test (the one that Sykes used) certainly does not sample the coding region at all and includes only a small subset of the control region.

Nevertheless, Sykes for some unknown reasons assigned U haplogroup for some portion of their HVS1-CRS sequences. How is it possible to guess what it is without any hint such as RFLPs or HVS2? Using witchcraft, I suspect.

[britzkrieg]

Thanks, everyone, for all the responses so far. I really appreciate it.

I'm curious about this because I am trying to identify a native European with my particular HVR1 motifs. Mitosearch.org shows no matches for me. I have three (known) matches in the FTDNA database, but all three women trace their maternal ancestry back to the same region of the United States that I trace mine. (In one case, down to the county!) For genealogical purposes, I'd like to know, as specifically as possible, where my mtDNA originated. (I hope that doesn't make this post too off-topic for this forum.)

My HVR1 motifs are as follows:

16129A 16171G 16223T 16293C 16311C 16391A 16519C

I remember reading somewhere that the 16519C mutation is not useful for genographic purposes, since this position is volatile and prone to flipping back and forth. I don't know if that makes it safe to ignore for genealogical purposes, though.

Anyway, most of my other motifs are characteristic of Haplogroup I or subclade I1. Two, however, are more interesting -- 16171G and 16293C. I was very excited to find a single sample in Sykes' PDF showing these two motifs. The sequence is the same as mine except that it lacks the last two motifs, 16391A and 16519C. The sample is categorized as Haplogroup I.

Thus my original question about 16391A.

Could I be maternally related to the person who provided this sample, within the genealogical time frame? I guess that's the real question here.

Thanks,
britzkrieg

[vineviz]

Nevertheless, Sykes for some unknown reasons assigned U haplogroup for some portion of their HVS1-CRS sequences. How is it possible to guess what it is without any hint such as RFLPs or HVS2? Using witchcraft, I suspect.

My money was on the Magic 8 ball.

[britzkrieg]

Oh, addendum to my last message: The 293 mutation I found in the profile in question is actually 2931. I have no clue what that extra "1" is supposed to mean; for all I know, it's a scanning artifact. Let me know if you have any insights. Thanks.

[vraatyah]

>My money was on the Magic 8 ball





>Could I be maternally related to the person who provided this sample, within the genealogical time frame? I guess that's the real question here.

unfortunately, the answer is likely to be "NO". There are 3 published matches, all from different countries:

Dupuy 1996, Norway
129-171-223-293C-311-362A
(sequenced between 16110 and 16362)

Richards 1996, Welsh
129-171-223-293C-311
(16069-16365)

SWGDAM, USA Caucasian
129-171-223-293C-311
73-199-204-250-263-315.1C
(16024-16365, 73-340)

[vineviz]

I remember reading somewhere that the 16519C mutation is not useful for genographic purposes, since this position is volatile and prone to flipping back and forth. I don't know if that makes it safe to ignore for genealogical purposes, though.

For genealogical purposes, you should not ignore even the hottest of mtDNA hotspots. But remember that a perfect match does not guarantee a genealogical link, nor do mismatches guarante that such a link does not exist.

Thus my original question about 16391A.

Could I be maternally related to the person who provided this sample, within the genealogical time frame? I guess that's the real question here.

Without more data on the Sykes sample, there is no way to say for sure. Really, unless you are comparing two full mtDNA sequences this question is nearly impossible to answer.

[vraatyah]

Oh, addendum to my last message: The 293 mutation I found in the profile in question is actually 2931. I have no clue what that extra "1" is supposed to mean; for all I know, it's a scanning artifact. Let me know if you have any insights. Thanks.

in most papers .1 stands for an insertion which is really possible at this position in the middle of a C-tract. However, the above 3 sequences show 16293C - the same as you wrote before. Anyway, from general knowledge, insertion at this point would be several times more likely if accompanied by a transversion 16293A>C: in this case we get a chain of 7C.

[britzkrieg]

Thanks, everyone, for all the great information. I've learned a lot from this thread. The conclusion is disappointing, though not unexpected. Evidence of a Welsh origin would have really sharpened my genealogical research, but I figured it was too good to be true.

I have had my HVR2 region sequenced as well. Now I just have to wait for that European exact match to pop up.

Thanks again. You folks are great!

[vraatyah]

I'd like to correct my post about the feasible connection in the last centuries.

The answer is likely to be "NO" unless you limit your search to the same location where your line was first documented. Of course, local matches with the same rare mutations could be related to your line, moreover, some of them must be related.