Isn’t one perfect match enough?
Some of those who’ve been following along with The Legal Genealogist‘s quest for a definitive answer to the identity of an elusive fourth great grandmother have had a question.
Since there are only two candidates for that particular slot in the family tree chart, and we descend in an unbroken female line from only one of those two, and we already know we have a perfect mitochondrial DNA match (mtDNA for short) to descendants of one of them, why are we bothering chasing candidates to test the other line?
Good question.
Let’s do a quick recap.
My 2019 research goal is to put paid to one particular genealogical question: who was the mother of Margaret (Battles) Shew?
Margaret is my 3rd great grandmother. My line comes down from Margaret’s daughter Martha Louise, to Martha Louise’s daughter Eula, to Eula’s daughter Opal, to Opal’s daughter Hazel, who was my mother.
The problem is that Margaret’s father, William Battles, was married twice, first to Kiziah Wright in Oglethorpe County, Georgia, in 1818,1 and, second, to Ann Jacobs, in St. Clair County, Alabama, in 1829.2 Since William ran off with Ann during his marriage to Kiziah,3 and Margaret’s birth year can’t be narrowed down closer than a range of years that spans both relationships, it’s entirely possible that either wife could be Margaret’s mother.
We have a perfect mtDNA match to a documented descendant of wife #2, Ann Jacobs. So… why isn’t that good enough? Since we don’t think that Kiziah actually left any descendants, why am I spending time trying to find someone else who would share her same mtDNA signature — a documented female line descendant of Kiziah’s mother, whom we now know was Lucy (Jones) Wright Williford?4
After all mtDNA for short is such a powerful testing tool, giving solid evidence of our mother’s mother’s mother’s line,5 surely having a perfect match to Ann’s descendants means she was Margaret’s mother, right?
Well, let’s ask a few more questions:
Q: What exactly does it mean to have a perfect mtDNA match to someone else?
A: It means that somewhere back in time that person and I (or, more accurately, my maternal aunt since she’s my family’s baseline tester) have a common female ancestor.
Q: How far back in time was that common female ancestor?
A: We have no idea.
Q: Why not?
A: Because haplogroup H has been around for 25,000 to 30,000 years, and haplogroup H3 for about 7,000 years or more. But we don’t know how long our specific subclade has been around or how stable it’s been over time.
Q: What’s your specific subclade?
A: We are H3g.
Q: How common is that?
A: We have no real idea, just some guesstimates based on who’s tested.
Q: What does the testing show?
A: Well, haplogroup H is the most common among women of European ancestry, showing up in 40-50% of all those tested. Subclade H1 makes up about half of those (so about 20-25% of women of European ancestry) and subclade H3 is two to three times less common (so about 8-10%).6
Q: And…? What about subclade H3g?
A: We have no real idea, just some guesstimates based on who’s tested.
Q: C’mon. What does the testing show?
A: There are 12,439 people who’ve tested their mtDNA and who are members of the Mitochondrial DNA Haplogroup H & HV Project at Family Tree DNA. Of those, 7,895 have done the full mtDNA sequence test, so we have complete results for them. They’ve been grouped into 853 groups and the vast majority are grouped as H or H1. Some 15 are grouped as H3g.7
Q: Well, then, 15 out of 7,895 — that’s less than two-tenths of one percent of all those who’ve done full mtDNA testing and are in haplogroup H. Isn’t that strong evidence?
A: We have no real idea.
Q: You’re being annoying with this “no real idea” bit.
A: Tell me about it.
Q: What’s your problem now?
A: First off, the project numbers don’t tell the full story because we know not everybody who’s tested belongs to the project. My aunt doesn’t just have 14 mtDNA matches — she has 33 at a genetic distance of 0. And she has many more matches at a genetic distance of 1.
Q: Well, you can disregard all those who aren’t perfect genetic-distance-of-0 matches, right?
A: Nope. Because anybody with what’s called a heteroplasmy (“the presence of a mixture of more than one type of … mitochondrial DNA … within a cell or individual”8) is recorded as a genetic distance of 1 — and that includes both my aunt’s niece (me) and her daughter (my first cousin).9 So we can’t just toss those out as meaningless.
Q: Sigh… this is complicated.
A: Tell me about it. And there’s another problem too. We really don’t know how common H3g is overall in the population — we only know how common it is in this one subset of this one database.
Q: In other words…
A: We have no real idea just how many women of European descent might actually share the exact same mtDNA, inherited from one common female ancestor who lived 7,000 years ago.
Q: But surely the mtDNA would have changed over time and nobody alive today would have the exact same mtDNA as somebody who lived 7,000 years ago!
A: We have no idea. There just isn’t any data. The best you come up with in the literature are statements like “In general, mtDNA transmission is stable across many generations”10 — but how many generations? We have no idea.
Q: Which means that…?
A: Right. Kiziah could have been H3g too.
Which is a long way of saying that it’s not enough to match Ann’s mtDNA. We have to be sure that we don’t also match Kiziah’s mtDNA simply because she and Ann were both women of European descent and shared a common female ancestor somewhere back in the mists of time.
Cite/link to this post: Judy G. Russell, “Why chase that mtDNA,” The Legal Genealogist (https://www.legalgenealogist.com/blog : posted 23 June 2019).
SOURCES
Image: Distribution of mtDNA haplogroups H1 & H3 in Europe, North Africa and the Middle East, Eupedia.com
- Oglethorpe County, Georgia, Marriage Book 1: 61, Battles-Wright, 12 December 1818; “Marriage Records from Microfilm,” Georgia’s Virtual Vault (http://vault.georgiaarchives.org/cdm/ : accessed 23 June 2019). ↩
- St. Clair County, Alabama, Marriage Record 1: 53, Battels-Jacobs, 25 Dec 1829; digital images, “Marriage records (St. Clair County, Alabama), 1819-1939,” FamilySearch.org (https://familysearch.org : accessed 23 June 2019). ↩
- Blount County, Alabama, Circuit Court Minutes B: 373-375 (1829); Circuit Court Clerk’s Office, Oneonta, Ala. ↩
- See generally Judy G. Russell, “Finding Margaret’s mother, part 2,” The Legal Genealogist, posted 18 May 2019 (https://www.legalgenealogist.com/blog : accessed 23 June 2019). ↩
- ISOGG Wiki (https://www.isogg.org/wiki), “Mitochondrial DNA tests,” rev. 28 Jan 2019. ↩
- Maciamo Hay, “Haplogroup H (mtDNA),” Eupedia (https://www.eupedia.com/ : accessed 23 June 2019). ↩
- “Project Statistics,” Mitochondrial DNA Haplogroup H & HV Project, Family Tree DNA (https://www.familytreedna.com/groups/h-mt-dna-haplogroup/ : accessed 23 June 2019). ↩
- ISOGG Wiki (https://www.isogg.org/wiki), “Heteroplasmy,” rev. 26 Nov 2017. ↩
- generally Judy G. Russell, “The myth of the GD0,” The Legal Genealogist, posted 26 Nov 2017 (https://www.legalgenealogist.com/blog : accessed 23 June 2019). ↩
- V. Cortellini, et al., “Evaluation of mtDNA stability across the maternal line: A study on three generations in two families,” FSI Genetics 5 (Dec 2015): e19-e20. ↩
In general, how many matches are people getting on the FMS test? Is it common to have absolutely no matches at the genetic distance of 0 and no more than about a half dozen at the genetic distance of 1?
My sister has 2 mutations I don’t. So she has zero 0GD matches and I have around 300.
Judy, thank you for such a clear explanation. I didn’t see the problem until you explained it either. Do I have this right? You have a connection to descendants of one wife through mtDNA but you don’t know if the other wife also shares that same DNA through a potential common ancestor to both of them?
The stats are more in your favor once you look at the full database at FTDNA.
https://www.familytreedna.com/public/mt-dna-haplotree/H;name=H3g
There are 99 out of 150,000 records that are just plain H3g. You don’t belong to any of the further subclades, so that rules out quite a number.
Thanks Judy
Ah, if only mtDNA was always conserved generation to generation. See figure 4 in http://dx.doi.org/10.1101/222109 (free) for a pedigree that has siblings with 2 different sequences with one of those siblings having 3 children with 3 different sequences. If you only look at haplotype specific positions (T2e in this case), there are no changes but a more detailed look shows changes do occur at potentially every generation. In fact, the mitochondrial disease families (Table 3), show substantial changes in a single base frequency that leads to the disease.
Heteroplasmies by definition are not useful for genealogy.