That and $2.29 = coffee
It’s another one of those things that comes up all the time in genetic genealogy.
Someone, somewhere, will tell you that you don’t really need to do YDNA testing — the kind of DNA test that looks only at the male gender-linked chromosome and defines the genetic characteristics of our father’s father’s father’s line1 — or mitochondrial DNA (mtDNA) testing — the kind of test that looks at the type of DNA we all inherit from our mothers and that defines the genetic characteristics of our mother’s mother’s mother’s line.2
And, that person will often insist, you can get “the same” information from doing an autosomal test with 23andMe — that’s the kind of DNA test that works across gender lines and helps identify cousins who share bits and pieces of DNA you both inherited from common ancestors3 — since, after all, you will get your haplogroup if you do.
A haplogroup, in case you were wondering, is “a genetic population group of people who share a common ancestor on the patrilineal or matrilineal line. Haplogroups are assigned letters of the alphabet, and refinements consist of additional number and letter combinations.”4 In plain English, it’s the general branch, sometimes even the twig, on the human family tree where you and your ancestors — male or female — can be found roosting.
So, if you know your haplogroup from testing with 23andMe, you don’t really need to shell out for the specialized YDNA or mtDNA tests, right?
Um… wrong.
For two reasons.
First and foremost, the haplogroup is just the general location on the overall human family tree where your ancestral line can be found. By itself, it’s not enough information to help you compare your results to another person’s results in anything more than the most general way that will be even remotely meaningful in genealogy.
I don’t want to understate the haplogroup. It’s absolutely true that if you are, say, mtDNA haplogroup U5, then you do not descend from the same common female ancestor as someone else whose mtDNA haplogroup is K. You may surely be cousins, but not along your direct female line: you’d need to look somewhere else in your ancestry for the common ancestor you both share.
The same is true for the YDNA haplogroup. My Robertson ancestors were haplogroup J. That means we don’t share a direct male ancestral line with all those other Robertsons who’ve tested and turn out to be haplogroup I or R.
But when you turn it around — say, we found another Robertson who was J — that by itself just isn’t enough information to say he’s kin to our Robertsons. Right now, in the Clan Donnachaidh DNA project, there are three separate — and totally unrelated — groups of Robertsons, each of which is haplogroup J, but whose DNA is so different from the other groups that there’s no realistic chance we descend from the same man.
To be able to quickly and easily compare one set of haplogroup J Robertsons to another set of haplogroup J Robertsons, we need some information about DNA markers called short tandem repeats, or STRs, that are patterns in the DNA.5 And to get the STR markers, you need to do dedicated YDNA testing.
Don’t misunderstand, please: it’s a really good thing that 23andMe gives people their haplogroup predictions when they do autosomal testing with 23andMe. I sure don’t want 23andMe to stop providing those haplogroup estimates. It’s just not the same thing, and not as genealogically useful for direct line male research, as having the STR data you get from YDNA tests.
The same problem exists with the mtDNA haplogroup predictions from 23andMe: they’re nice to have, but not as complete and not as useful as you can get from dedicated mtDNA testing. My mtDNA haplogroup prediction from 23andMe, for example, is H3: a fairly large branch on the female haplogroup family tree. Dedicated mtDNA testing can break it down further, to a smaller sub-branch of H3g.
And since mtDNA haplogroups change so very slowly over so many hundreds, even thousands, of years, even that smaller sub-branch is too broad to be really helpful. In reality, especially for women like me in haplogroup H, only the very highest level of mtDNA testing is really useful for comparing my mtDNA to that of another individual.
Telling me I’m H3 tells me I’m very much like a huge percentage of women of European descent. Telling me I’m H3g is better, but still not enough. But telling me exactly what my results are, for me individually, with a careful look at every location within my mtDNA, is the kind of detail I can actually use in genealogy.
So what good is it to have a haplogroup, by itself, without more?
Except in the most general way, not very much.
It’s good enough to buy you a cup of coffee… at least if you’ll chip in $2.29 in cash.
SOURCES
Image: Haplogroup J (Y-DNA) by Rafy, Wikimedia Commons, CC BY 3.0
- See ISOGG Wiki (http://www.isogg.org/wiki), “Y chromosome DNA tests,” rev. 5 March 2014. ↩
- See ISOGG Wiki (http://www.isogg.org/wiki), “Mitochondrial DNA tests,” rev. 9 July 2014. ↩
- See ISOGG Wiki (http://www.isogg.org/wiki), “Autosomal DNA,” rev. 26 July 2014. See also Judy G. Russell, “Autosomal DNA testing,” National Genealogical Society Magazine, October-December 2011, 38-43. ↩
- See ISOGG Wiki (http://www.isogg.org/wiki), “Haplogroup,” rev. 26 July 2014. ↩
- See ISOGG Wiki (http://www.isogg.org/wiki), “Short tandem repeat,” rev. 20 July 2013. ↩
Have taken both the autosomal and mtDNA test with Family Tree as well as the autosomal test with Ancestry. Eventually I want to get around to taking a test with 23 and Me. I did solve a family mystery when I took the mtDNA test and matched up EXACTLY with a lady in Canada. We had to go back 9 generations but all roads led back to a Quaker ancestress. It was rather exciting. Had my first cousin take the Y-DNA test to see if we could solve a mystery as to who our great great grandfather was on my family’s paternal side. DNA is a very useful tool and taking as many tests as possible only increases a person’s chances for solving family lineage/mysteries.
Genetic genealogy geek here:
maternal = J2c3a1
paternal = R-CTS4466 (A151)
I’ll buy the coffee next time I see you and we can share “another one of those things.” Something else is bound to come up, and I’ll set the science aside and share stories about someone I know.
Judy, THANK YOU for this. It explains exactly the position I face daily as an Admin of multiple projects continually challenged by 23andMe’s proclamation of their belonging to such and such a Haplogroup and therefore FTDNA or my knowledge is at fault. (Or as occurred in the past, Ancestry’s previous tests before these were removed).
When my brother tested with 23andMe (as did I), as a newbie I was confused by the apparently different classifications (we both started with 23andMe until I discovered FTDNA and then I tested us both specifically for mtDNA FMS plus FF plus Y-111 for him) UNTIL I went searching seeking answers and it all came clear. That is, both labs used different naming systems – and continue this today.
I would like to simply now quote your article for such future queries – so again THANK YOU!
Gail Riddell
Glad to help, Gail!
Also confusing is that I got *different* haplogroup results from 23&me (R1b1b2a1a1) and FTDNA (R1b1a2a1a1a) a couple of months apart. Heard different explanations such as different chip, changing definitions, error etc. Finally, it seems NatGeo & FTDNA later reduced me to R-Z2, presumably,giving up on the other end of the haplotree.
Haplogrouping seems to be a pretty fluid term.bob
Hi Judy,
I recently had the opportunity to get a free test at 23andMe by joining a study group that is sponsored by Pfizer Pharmaceutical co.
I have not had any other tests done, but for now I was happy with the results in my autosomal test. It matched up very well with my 20+ years of research on my family.
A few items needed some research on my part, especially on my mothers side, C1b. It seems like anyone who’s family descended from indigenous people, whether from North or South America, or the Caribbean will be part of this haplogroup, lsted as East Asian and Native American. Understanding how these people migrated helped me understand why my mother was in this group. I’m fairly certain that her Puerto Rican ancestry will have gene segments from the Taino or other Caribbean tribes.
The strangest result was from the list of DNA relatives. Out of 982 matches, number one was a woman with polish and Russian ancestors and no surnames in common with me, but was described as a 2nd to 3rd cousin with 2.09% shared and 7 segments. Then I find the third person from the bottom of the list that was listed as a 3rd to distant cousin with only 0.23% shared and 1 segment was actually my 2 cousin once removed, who I had found through my own research years ago.
Could the common ancestor from this Polish/Russian woman be as close as Two or three generations back? How would you proceed to find out? Any suggestion would be very welcome.
Regards,Joe
My example to the beginning person is to compare two dogs.
If I have a German Shepherd and a Poodle, they CAN’T be cousins.
If I have two German Shepherds, they MIGHT be related.
So comparing haplogroups is similar. If the haplogroups don’t match, they can’t be related. But if the haplogroups do match, they might be related.
This logic seems to work for those people who are a little afraid of some of the genetic vocabulary that is used.
You can also bounce it up to two Swedes vs a Swede and an Italian etc.
Good analogy!
FTDNA says I’m HV4, and 23andMe says I’m RO. What caught me by surprise though was when FTDNA said I’m basically Italian and Middle Eastern, and 23andMe said I’m Italian and British/Northern European. How did that happen? Could it be contamination?
R0 is up the mtDNA ladder from HV4, so what happened is, 23andMe gave you its best guess on fewer markers that it reads to give predictions of mtDNA haplogroups, and FTDNA gave you your exact position down the ladder from R0. You can see this on this chart: http://www.phylotree.org/tree/R0.htm
Okay, that makes sense. Thank you for this chart. Then I can help family and friends who have these questions too. What do you suppose happened with my paternal (non-Italian) side showing such drastically different ethnicity estimates on 23andMe? I’m thinking of telling my dad that he should just test with FTDNA instead of 23andMe because of this. My mom, on the other hand, got accurate information with both companies.
Each company makes its ethnicity estimates based on its own reference populations and they can be drastically different. Mine, for example, are all over the map (see: https://www.legalgenealogist.com/2017/04/16/still-not-soup/). In general, it appears that 23andMe is the most accurate these days, but that can change any time.
Oh wow! Okay, thank you for sharing this with me. I was worried that something had happened to my sample along the way, and my parents were really upset too. I feel better seeing this. Now I’m going to relay the message to my family.
Glad I could help. I wish people understood better that the ethnicity estimate part of DNA testing is the least accurate part — no matter what the TV ads suggest! 🙂