We get Y with a little help from our friends
With all the attention paid to autosomal DNA testing these days — with good reason, since it’s the newest, most complicated, most difficult (and most fun!) DNA test to work with — we sometimes forget that there are other, older, somewhat simpler DNA tests out there that can sometimes answer our questions.
And then again maybe not.
Today’s reader question is one of those maybe-not situations.
Reader Barb DeMarco is trying to understand YDNA results. And, she asks, “Can knowing the genetic distance from a Y-DNA match, or say 35/37 markers, help to identify at what generation the link is made? What does a 37/37 match mean?”
Let’s start by defining some terms:
• YDNA, remember, is the kind of DNA that only men have — it’s contained within the gender-determinative Y chromosome that’s passed down through the generations from father to son to son. So only men can take this test and it provides information on the direct paternal line only.1
• The particular DNA sequences being looked at in YDNA testing are called short tandem repeats (STRs): patterns of DNA code that form sequences and where the sequences are repeated a number of times at those particular locations on the Y chromosome.2
• The markers are the genes or DNA sequences with known locations within a chromosome that make them useful for comparing one person’s DNA with another.3 In the case of YDNA testing, it’s the STRs — those short tandem repeats — in a number of locations (12, 25, 37, 67 or 111 if you’ve tested at Family Tree DNA) that are the markers being reviewed.
• Genetic distance is a term used to describe the differences, if any, between the markers reported in one person’s DNA results and those reported in a second person’s results. “A genetic distance of zero means that there are no differences in the two results and there is an exact match.”4 If the genetic distance is small enough, the two people are considered a match, meaning they share a common ancestor.
But which ancestor? How far back? That’s the $64,000 question,5 isn’t it? We’re all out to identify that wondrous, sometimes-mythical-seeming creature called the Most Recent Common Ancestor (or MRCA) — the most recent individual from which those tested are directly descended6 — and to figure out how many generations back that person lived, called the Time to Most Recent Common Ancestor (or TMRCA).7
So Barb’s problem involves looking at the results of a test where 37 markers were looked at, and trying to understand what it means if there’s a 37-for-37 match (which means a genetic distance of zero) and what it means if there’s only a 35-for-37 match (which means a genetic distance of two). And whether that difference can help narrow down the MRCA or the TMRCA.
Family Tree DNA — where most YDNA testing takes place because of its wide range of surname projects — has a really good Frequently Asked Questions section which helps explain what it means to match another person in a YDNA test where the two men share a surname in 12 markers, 25 markers, 37 markers, 67 markers and 111 markers.
For 37 markers, here’s what Family Tree DNA says about genetic distances of 0, 1 and 2, for men who share a surname:8
|0||Very Tightly Related||
A 37/37 match between two men who share a common surname (or variant) means they share a common male ancestor. Their relatedness is extremely close with the common ancestor predicted, 50% of the time, in 5 generations or less and over a 95% probability within 8 generations. Very few people achieve this close level of a match.
All confidence levels are well within the time frame that surnames were adopted in Western Europe.
A 36/37 match between two men who share a common surname (or variant) indicates a close genealogical match. Very few people achieve this close level of a match, and it is within the range of most well-established surname lineages in Western Europe.
It’s most likely that they matched 24/25 or 25/25 on a previous Y-DNA test, and the mismatch will be found within DYS576, DYS570, or CDY.
A 35/37 match between two men who share a common surname (or variant) means they share a common male ancestor. The mismatch is likely within the range of most well-established surname lineages in Western Europe.
It is most likely that you matched exactly or closely on previous Y-DNA tests and the mismatch is within DYS439 or DYS385, DYS389i, 389ii, DYS458, DYS459, DYS449, DYS464, DYS576, DYS570, or CDY.
See the problem? It’s a statement of odds. And there are a whole lot of variables that affect the odds.
First and foremost, we need to understand that markers change — the changes are called mutations, but the changes are usually benign, not the three-eyes-two-heads type of mental image we tend to get when we hear the word mutation. And they can change in a single generation: a father may match his own son in only 36 of 37 markers.
Second, we need to understand that some markers change more quickly than others. That’s why you see the references in the chart above to mismatches in certain specific markers. They’re the ones known to change faster than others, and where you’re most likely to see a difference even between father and son.
Third, we need to understand that the marker changes can be random: they can occur in the same place in the YDNA but in different generations or different lines of descent. In my Baker family, for example, there is one marker out of 37 — a fast-changing marker — where two-thirds of our group show one result but fully one third of the men tested have the same different result. You might expect that they’d all trace back to the same son or grandson of our earliest known ancestor. And four of them do descend from a particular son. But the fifth one doesn’t. He descends from a different one.
Finally, whenever we’re dealing with statistics, we have to remember that changing one variable will change the odds. For example, my Uncle David’s YDNA is a 37-for-37 match with another man of the same surname. Assuming that we know they don’t share a common father (one generation), the chances are 89.48% that they share a common ancestor within five generations. Assuming that we know they don’t share a common grandfather (two generations), the chances drop to 83.49% that they share a common ancestor within five generations. And assuming that we know they don’t share a common great grandfather (three generations), the chances are only 74.10% that they share a common ancestor within five generations.9
The odds are different if the match is 35-for-37 markers. Again using my Uncle David and one of his same-surname matches, if we know they don’t share a common father (one generation), the chances are 41.65% that they share a common ancestor within five generations. If we know they don’t share a common grandfather (two generations), the chances drop to 40.13% that they share a common ancestor within five generations. And if we know they don’t share a common great grandfather (three generations), the chances are only 35.76% that they share a common ancestor within five generations.10
So the answer to Barb’s question is a complicated one. We know for sure that a 37-for-37 marker match means the two men are closely related from a genetic perspective. A 35-for-37 match means they are related but not as tightly. But can it tell us anything about how far back they may find the common ancestor?
As long as we understand that it’s all a matter of the odds.
- See ISOGG Wiki (http://www.isogg.org/wiki), “Y chromosome DNA tests,” rev. 7 Dec 2013. ↩
- ISOGG Wiki (http://www.isogg.org/wiki), “Short tandem repeat,” rev. 20 July 2013. ↩
- See Emily D. Aulicino, Genetic Genealogy: The Basics and Beyond, Kindle edition (Bloomington, IN : AuthorHouse, 2014), Appendix F: Glossary, “marker.” ↩
- ISOGG Wiki (http://www.isogg.org/wiki), “Genetic distance,” rev. 9 Nov 2013. ↩
- Wikipedia (http://www.wikipedia.com), “The $64,000 Question,” rev. 24 Jan 2014. ↩
- ISOGG Wiki (http://www.isogg.org/wiki), “Most recent common ancestor,” rev. 27 Oct 2013. ↩
- Ibid. ↩
- “If two men share a surname, how should the genetic distance at 37 Y-chromosome STR markers be interpreted?,” Frequently Asked Questions, Family Tree DNA (http://www.familytreedna.com/faq/ : accessed 25 Jan 2014). ↩
- FTDNATiP analysis on known kits, names withheld for privacy, Family Tree DNA (http://www.familytreedna.com : accessed 25 Jan 2014). ↩
- Ibid. ↩
Thanks for clarifying these points. People often seem to have great difficulty understanding probabilities. However, in addition to considering the odds we are finding that it is increasingly important to consider the SNP testing status of a match. There have been cases of seemingly close matches at 37 markers, even with the same surname, where the two people have turned out to be in different subclades and hence not related in a genealogical time frame. This is a particular problem with haplogroup R1b1a2. See the ISOGG Wiki article on convergence for the background on this problem:
Very good point, Debbie, and thanks for adding your voice!
Family Tree DNA provides a TiP report – (orange button on matches list) that calculates the TMRCA. This report takes into account differences in marker mutation rates, and the number of markers that differ between those in your match list. It can be a useful tool.
Hmmm… I probably should have made it clearer that that’s what I used for the comparison of my uncle’s 35-for-37 and 37-for-37 matches — but it’s in the source footnotes!!!
I have a surname back 5 generations (3rd great) on my father’s direct line. One reason I wanted to have my brother test was to possibly identify other surnames that may have been taken when Jews were forced to take surnames about 1812. I did the 67 marker. Was that totally misguided? I sincerely doubt there will be other Schwarzbaums (uncommon name) that have tested so there will not be a surname project.
I am completely confused as to how one does research with patrilineal names. If you don’t yet know the name of the father I am not sure how you can discover who is the son of the father. But that is another topic.
Carolyn, doing a Y test is a good way to check possible other surnames adopted in populations that took surnames late. My cousin Dick, whose paternal line is Ashkenazi, has NO matches to his own surname — but a couple of good matches to one other, which may point the way back to a particular location. The problem with the Ashkenazi population in particular is endogamy (intramarriage within the group). So you’re always dealing with so many more matches than you might in less endogamous populations. The leads can be overwhelming.
Thank you. I know. I have so much intermarriage in my own family that I do not even understand when someone says they are only doing their direct line. Even among my mother’s German Lutherans, cousins md cousins.
But it is good to know I did not waste my money and the results may be helpful.
I have used the new tool dnagedcom (Don Worth) (http://www.dnagedcom.com/) which allows you to see much more information on a larger group than using the tools at FTDNA. My sister and I have a different group of 2-4th cousin matches on autosomal results which indicates how important it is to test everyone you can.
I can’t wait to see my brother’s results and see which “deck of cards” he was dealt.
Hello, we are trying to find the father of my nephew. Ancestry.com did not turn anything up. He has now done the Y chromosone test and some exact matches turn up. After reading your comments, how if ever can we find who is his dad?? His mom is no help.
You’re going to need to do some studying to understand how to integrate DNA results with traditional paper trail genealogy. You might want to read through the methodology info at DNAAdoption.com.
Sorry to be completely anal Judy, but there is a typo in your references. Item 5 should read the $64,000 question.
You’re right — and it’s fixed. Thanks for the heads-up!
Although having a paper trail back to 1801 with my last name and good research on most cousins, I have no relatives of my surname. I have 2 people at 1 distance on Y37 both living in different countries and my family left there 200 years ago. Should I take the y111 test?
The only possible answer is, it depends. Do the two people you DO match have anything in common with each other, even if they don’t match your surname? Have those two people tested beyond the 37-marker level? Do you have a common or uncommon haplogroup?
Glad to hear the reference to a match where the surname is not the same. I have a 37/37 match with a fellow with a different surname. My kin have eight generations on this side of the Atlantic (a fairly complete Tree) and his about five also a good Tree. He says that the answer is in England and I believe that it is a non-paternal event. We will likely never know.
Warren, one of the best treatments I’ve found of the different-surname scenario can be found here at the Bowes One-Name Study website.
Thanks for the explanation and it was helpful. I know every situation is unique and complicated by migrations and odds, etc. I have a situation whereby an ancestor matches 25 (genetic distance of 1 or 2) and 37 (genetic distance 3 or 4), but they are all from England. My ancestor was from central, rural Poland. Even a 67 marker person (genetic distance 6 and 7) is English and the another German. Everything I have read states I would likely share a common ancestor with the 37 marker matches, even with a genetic of 3 or 4. So far I don’t really match anyone from Poland.
Would this possibly mean that this person may have originated in England, perhaps 500+ years ago.
Sometimes more data is even more confusing…Thanks
It could mean the common ancestor may have been in England 500+ years ago — or anywhere else in Europe or even around the world! Remember that YDNA changes very slowly, so that you, the German, the Englishman and others could all have a common ancestor from (just as one example) Russia even farther back than 500+ years.
My grandfather Smith, the bookie in early 20th-century Cincinnati, would have loved DNA testing – it’s all about the odds!
My maiden name is Smith, and my cousin (my paternal uncle’s son) volunteered to Y test for us. More than any surname, you would think we’d have plenty of Smith matches, especially given the enormous size of the Smith surname project. Nope… not a one. Looks like I’m a Hunter! My cousin matches one of their lines 67/67. The divergence from Smith is not just with my cousin himself – whatever he is, I am, too, since we match autosomally. Maybe Grandpa Smith, the bookie, in an attempt to evade the law, changed his name? No, we know his father was, at least in name, also a Smith. But there was this old family story kicking around about someone way back running off to work in a carnival and changing their name to Smith. Don’t know about the carny part, but looks like the name change part is true! And now we know where to search for the change – with the Scottish Hunters!
Just FYI, the link to the Bowes One-Name Study in your Jan 27 post reply regarding different surnames in Y matches no longer works. I don’t know if it’s the same, but there is a Bowes One-Name Study site now at https://sites.google.com/site/bowessurnames/home. However, I was unable to locate within it a page about the different-surname scenario.
P.S. Here in Louisville, KY, our genealogical society just had our annual fall seminar last Saturday with the wonderful Christine Rose as our main speaker. Can’t wait for October 17, 2015, when it will be you, Judy!
Darn it, that Bowes site has been completely revamped. But you can find the item via the Wayback machine here.
Debra, my step-grandfather left his wife (my grandmother) and ran off with a neighbor girl and his boss’ chickens. He left Cincinnati and he was found in Pennsylvania several years later. Had changed his name to Smith and his many descendants may not know to this day they are really “Adams” (or…did he change his name prior to fathering 2 children with my grandmother? hmmm…)
I need a clarification. Is it possible to confirm the common ancestor approximately 30 to 35 generations ago.There are 4 surnames slightly varies from each other, but suspects to be separated from a common ancestor 950 years ago.
It’s possible to confirm that all are related. Without a paper trail, it isn’t possible to specify exactly how they’re related (a father and son will look as much alike in a YDNA test as an uncle and nephew).
Concerning the “sharing a common surname” between two matches. In my case, my Y-DNA matches will likely never share a common surname, since my paternal line is Swedish….and the naming custom was “patronymic”…..whereby the surname changed every generation, with the father’s first name used as the root of the son’s surname. My Y-DNA 12 marker test shows six matches with zero genetic distance. Without any of these matches showing the same surname (due to the patronymic naming system), is there a chance that these matches do not share a common ancestor with me? I went ahead and paid extra money for the 25, 37, and 67 marker tests (kept getting ads from my testing service recommending these additional tests)…..but after spending all this extra money, I ended up with zero matches in the 25, 37, and 67 marker tests! So I am stuck with only the results of my 12 marker test, and no matches sharing my surname. Was the money I spent on all these Y-DNA tests a complete waste? Do the results tell me anything definitive? As I said, I show 6 matches with zero genetic distance, and another 4 matches with 1 genetic distance. Some of these matches have swedish sounding surnames, just none that are the same as mine. I’d appreciate any input to help me understand these results and what they tell me. Thanks.
Europeans in general test in far fewer numbers than their American counterparts, and many of us who have recent immigration in our families will have few if any matches. At this point, you’re better off trying to do targeted testing (coming up with candidates you believe you’re related to and trying to convince them to test) rather than just relying on matches showing up in your match list.
Is there a tool or a site that would help the mathematically challenged figure things out regarding Y DNA? My surname is very uncommon and was geographically very distinct in Co Mayo Ireland and I’m slowly getting males of non-related lines to test to prove we are related, even if a long time back. Records prior to 1800 simply are few and far between for our folks. But it would be nice to put some sort of timeframe on things
John – b 1954
Jim b 1956
Tony b 1941
12 Y DNA Genetic Distance John, Jim and Tony all match at 0
25 Y DNA Genetic Distance John and Jim are a match at 0. Tony is a match with Jim and John at 1
37 Y DNA Genetic Distance James and John are a match at 2. Tony is a match with Jim and John at 3.
67 Y DNA Genetic Distance Jim did not test at this level. Tony is a match with John at 3
As we slowly add more testees, I need help with both understanding and interpreting it from the perspective of : “hypothetically – This group as a hole shares a common GrX grandfather about the year XXXX. Jim and John share a Gr X grandfather from about XXXX
Is that possible? Is there a calculator somewhere or a group or anything that we can use?
Thank you so much! My ministry is that my grandson’s grandfather (deceased) was adopted. My grandson’s 111-marker Y-DNA gave him many matches, that one is actually another 111-marker test with a distance of 1. I’m still confused as to what this might indicate, since two today adoption of his grandfather, we don’t have “real” names to work with. I would appreciate your input, if you have time.
That’s a very close genetic match, but it doesn’t necessarily mean close in time. You need to work with your matches to see if you can put your grandson’s grandfather into the right place at the right time to be the child of some member of the matches’ family. It isn’t an easy process but it is doable. Working with others who work with adoption cases can be helpful. There are resources such as DNAAdoption.com and the Facebook group DNA Detectives that may help.
My father’s Y-DNA has me puzzled we had the 37 markers tested and he has no matches on the 37 nor on the 25 and on 12 MARKERS Only- 676 MATCHES and only one surname matches him on the 12 markers. If you can Please help….Thank you
Your father likely comes from a population where few men have tested. My own brothers match each other — and no-one else at more than 12 markers. Our father was born in Germany where few people have DNA-tested at all.
Is frequency of surname more important then genetic distance? Or are both important for different reasons. I’m helping my father locate his bio father or family. I notice he has many of the same frequency with two surnames. The one surname show up with 4- 5 matches and the other 3-4 depending on which marker. All 1’s and 2’s for each surname. One surname comes up more on 12, 25, and 37. However, there is a 0 match that is a different surname that’s on 12 and that same person is a 1 on 25 but then doesn’t show up at all on 37. Thanks for advice. 😉
Matches at 12 markers are very likely to be well outside the range of genealogical time and therefore not verifiable. If the match doesn’t continue to exist at 25 or 37 (or more) markers, it can safely be disregarded. A match that exists at 25 markers but disappears at 37 needs to be checked: if the person tested to 37, then it may be that several difference appear in the next 12 markers making it outside the range of matching. If the person didn’t test to 37 markers, you can’t draw any conclusions.
I tested with Family Tree Y DNA up to 67 markers. My test came back with 243 12 marker matches, 11 25 marker matches, 0 37 marker and 1 67 marker. None of those surnames are my surnames. of those I have one who matches 3 markers 12,25 & 67 and 5 who match 12 & 25 markers. Am I assuming that the others who only match 1 marker aren’t closely related? Also the 5 who match 2 markers aren’t but the one who matches 3 markers could be related??
There is some good help available in the Family Tree DNA Learning Center on understanding these numbers. For 12 markers, here. For 25 markers, here. For 37 markers, here. For 67 markers, here. Not sharing a surname complicates the issue, and you might want to read Roberta Estes’ blog “We Match… But Are We Related?“
if you can tolerate my ignorance, may I rehash the question about 0 1 2 …..and its relevance? I have a person that on 12 markers shows 0 and on 25 shows 1. So, what would be the likely time span between generations or has this been a waste of time because it’s impossible to say? Your comments are eagerly saught.
You will get a statistical estimate by using the TiP calculator in your YDNA results. Under the names of each of your matches, you will see the little icon with the letters TIP. Click on that and it will produce a calculator where you can indicate (if you know it) whether you and the match are related within a certain number of generations and it will tell you what the odds are that the match is within X number of generations. It’s purely statistical, which means it’s not set in stone, but it’ll give you some idea.
Very interesting information-thanks. So I can better understand my issue, I have matched 36 out of 37 makers with a person that I believe his grandfather to be my Father’s Father ( MY father never knew who his was). The maker that is different by 1 if 459b. Any insight you could provide would be great
A genetic distance of 1 from a potential first cousin (if the match’s grandfather is your grandfather, you’re first cousins) is well within the realm of possibility, regardless of the marker. But it’s not proof by itself: it simply suggests that you share a common male ancestor at some point in time. But if you are in fact first cousins an autosomal DNA test (the Family Finder test at FTDNA) will show you as a close match. You should both do that test if you haven’t already.
My brother took a DNA-37 test. At 25 markers, he has 6 matches–all different surnames. I contacted the top match (GD-0) who lives in Lithuania and tested DNA-111. We exchanged several match lists and DNA certificates. Interestingly, his 25 marker list contained the exact 6 matches as my brother’s list and the GD were identical. This top match drops off our 37-marker list as do 4 other names but one match continues to show at 37-markers (moved down from GD-2 to GD-4). What’s more, he still shows up on the Lituanian’s 67-marker list (moved from GD-2 to GD-7) but all the other names are gone from his list. My Dad just submitted a DNA-67 test so will see whether this shared match still matches my Dad at 67. I don’t know what to make of the fact that someone can fall off one’s match list yet continue to share another common match.
You have to be matching on the same markers for all of you to stay matching. Using small numbers to help explain this: If you and persons A and B all match on markers 1-5, you’ll be declared a match. If you then test to 10 markers, and person A matches you only on 1-6 and matches person B on 1-10, he might be declared a match to person B but not to you any more.
great article .. thank you .. I just got my Y results from FamilyTree for 12 markers .. I have an exact match with only one man there .. I have male line records back to Adam Lober born in Plauen Germany in 1489 .. he knows his fathers back to Johan Engstom born in Sweden in 1788 .. what can you tell us about that? and .. there is only one other fellow tested as an exact match .. he does not know the identity of his father.. what can we tell him ?
Twelve markers are basically just enough to confirm that you’re both male and descend from (in your case) European stock. You’d need to test many more markers to be able to say anything more with any degree of confidence. The problem is that 12 markers just isn’t enough to be sure that you’re related within a time frame where there’s any chance at all of figuring out how you’re related — and not even enough to be 100% sure that you’re related at all.
12 markers markers understood; what about 25?
Hi Judy. Being adopted at birth, I have no idea what my father’s surname was so I don’t have the benefit of matching surnames with any Y DNA match. I tested out to 111 markers but only have matches out to 67. My 49 matches at 67 are all 61/67 or 60/67. At 12 markers the genetic distance is 0 with all 81 matches. At 25 they are all 0 to 2 in terms of distance for 109 matches. 49 matches at 37 are 36/37 to 34/37. I have no matches at 111. I understand the odds percentages presented in Tip reports but really need some clarity on the best way to zero in on likely most recent common ancestors before I start messaging anyone with questions of my paternity. Any guidance from you will be greatly appreciated!
If you don’t see any cluster of surnames in your YDNA matches, then YDNA isn’t going to give you much help, particularly since you have no particularly close matches at 67 markers. You need to add autosomal DNA to the mix, and start building out the trees.
I have a question regarding one of my father’s Y-matches. I had my father Y test, to find out where in Germany my family originated from. My ancestors appear on the UK censors records as having been born in Germany, despite of having a British surname, namely ‘Lewis’ Lewis is also a common last name amongst Jews, where it is an anglicization of various Levitic surnames. Our ancestor came to Britain as a journeyman baker and confectioner with his wife and his son. Autosomal tests with various genetic testing companies and with Gedmatch have ruled out that they were of Jewish descent and werer very likely of Northern European descent, as the autosomal showed very Scandinavian and British-like results, where Gedmatch also shows a relatively inflated Eastern Euro input. Recently, my father got his first close match at 107/111 markers. Our match has paternal ancestry from Bad Ischl, Austria with the surname Grabner/Grebner. We have gotten in touch with him and it seems he does not match us autosomally. We also do not have autosomal matches with his particular last name. The majority of our close autosomal matches are either from Britain, which is my father’s British ancestry which are all rather well documented and Scandinavian, East German/Silesian matches, which I presume are from our immigrant ancestor. How informative is our 111 distance 4 match? At the 67 level, he matches us distance 3. What does this match prove? Our haplogroup in question is G-L497, which is a stricly Alpine haplogroup, although out of all the G-L497 subgroups, we seem to be in a group which also has some presence in Sweden and Northern Europe in general, as we are in the same subgroup as the famous Bure-kinship of Swede, although we share a most recent common ancestor 1100 BC. I would love to hear your thoughts and speculations regarding the origins of my paternal ancestor. Thank you in advance.
FamilyTreeDNA provides an explanation for the potential relationship of persons testing at various levels of YDNA markers in its Learning Center. For 67 markers, see here. For 111 markers, see here. The problem of course is that YDNA is stable enough that the “origins of [your] paternal ancestor” may be lost back in the mists of time. Your Alpine haplogroup chap may have had itchy feet… at a time when no records were kept.
How many markers Y-DNA we need to test to figure out if we have a common paternal lineage ancestor born before 1600, with another man. We both share the same surname, we both have well researched and documented family trees into 1600, and both of our families lived in the same region. Our paternal lineages do not connect by 1600 and the church records do not go any deeper in the past. Thus, I guess the Y-DNA could help us, to prove the obvious probability.
Also, how to deal with a highly speculative surname scenario without paper trail, for a possible common ancestor born before 1200.
Thanks for your help.
There’s no way to answer either of your questions easily. If the two men share a very uncommon haplogroup, then 37 markers could be enough to be reasonably confident that they descend from a common male line. If they share a very common haplogroup (R-M269 for example), you might need a Big Y test. As far as going back before or beyond the time frame of genealogical records, there’s no methodology that I know of.