Y-37 Exact Match and a Mystery!

[507]

My question is, how did these other people even know to join our surname project to begin with? It's MY last name that is the name of the surname project and there are probably 20 people in my group or "cluster" . I apologize for having so many questions but I cannot get any help or answers from our so called " co ordinator" at all. The majority of the people in the cluster have the same last name as me but the two people I have the exact 37 marker matches with are different last names than me. But they are in the project too and they are in my cluster. But I have 5 matches on 36 of 37 markers and they have the same last name as me.

[507]

It's not that your trees are garbage. It's just that there is an error in who a father was and it could have been hundreds of years ago. First of all, your grandfather is the first generation, because he is the ancestor that you have in common with your first cousin. Count back from there. Also consider certain circumstances, such as what happened in a family I know. This man born a hundred years ago, found out shortly before he died that the woman he thought was his mother was actually his grandmother - and the woman he thought was his sister was actually his mother. His sister had gotten pregnant out of wedlock and the grandmother told everyone that she was pregnant and claimed him as her child. So all generations following him have the wrong surname. And in other cases , a woman may have gotten pregnant by a husband who died and she remarried before the child was born - and the child only knew the second man as his father - and that was the name he carried. It's possible if she was barely pregnant when she remarried that she wasn't even sure who the father was . All generations following would have the wrong surname. There are many scenarios like this that could have happened. Just realize that you ARE all related .

Why start with my grandfather ? wouldn't Myself or My Dad be the first generation ?

[507]

How much of the "most recent common ancestor" is opinion and how much of it is fact ? HOw much better is the y67 test than the y37 test ? Is it worth it to upgrade ?

[nathanm]

To get back on track here, this tool makes sense to me in terms of what the probabilities mean and why the relatedness per number of generations decreases if the paper trail can verify no relatedness within X number of generations. But I do have a few questions in terms of how the percentages are calculated. Are they haplogroup specific and does it make any difference that one haplogroup is very commonly represented in population samples while others are observed with much less frequency? In other words, does a smaller statistical sample cause an increase or decrease in these percentages, or does it even factor at all? Also, are the faster mutating markers taken into account in the equations used to tabulate these percentages?

The calculation is independent of haplogroup, so you need to make sure those match before even comparing. As far as how common the various haplogroups are, I think that would change substantially if more men from different parts of the world were tested. R1b is far over-represented in all the databases, just because the vast majority of men tested are Western European.

Dr McDonald's TMRCA calculator uses an infinite alleles model, which assumes markers mutate at equal rates. (There's an older, but still informative description of that and a stepwise mutation model here.) It should get you in the ballpark, but it's only as good as those overall average mutation rates reflect actual behavior. It should work better if you differ on only one marker and use the observed mutation rate for that specific marker.

FTDNA used to state in the FAQ that their calculations are based on the individual mutation rates, so it should be more accurate (they changed the FAQ, so I can't find it right now). However, their calculations aren't at all transparent. I've tried matching their results with Dr McDonald's, and even with perfect matches or one off, the curves don't fit. They're using a different model.

As far as sample size, the TMRCA calculator is always n=2. It can only compare two individual haplotypes against each other. For doing comparisons among groups, a different tool is necessary.

[ekc123]

The calculation is independent of haplogroup, so you need to make sure those match before even comparing. As far as how common the various haplogroups are, I think that would change substantially if more men from different parts of the world were tested. R1b is far over-represented in all the databases, just because the vast majority of men tested are Western European.

Dr McDonald's TMRCA calculator uses an infinite alleles model, which assumes markers mutate at equal rates. (There's an older, but still informative description of that and a stepwise mutation model here.) It should get you in the ballpark, but it's only as good as those overall average mutation rates reflect actual behavior. It should work better if you differ on only one marker and use the observed mutation rate for that specific marker.

FTDNA used to state in the FAQ that their calculations are based on the individual mutation rates, so it should be more accurate (they changed the FAQ, so I can't find it right now). However, their calculations aren't at all transparent. I've tried matching their results with Dr McDonald's, and even with perfect matches or one off, the curves don't fit. They're using a different model.

As far as sample size, the TMRCA calculator is always n=2. It can only compare two individual haplotypes against each other. For doing comparisons among groups, a different tool is necessary.

Thanks, I'll take a look at Dr. MacDonald's calculator. That helps a lot to point me in the right direction. I had wondered about the rather large data set of samples from Western Europe (particularly England) myself and whether the percentages could change substantially as folks tracing ancestry to other areas are increasingly tested. On the my haplogroup matches page, it shows something like 25,000 samples tracing back to England very low numbers for other areas.

[nathanm]

How much of the "most recent common ancestor" is opinion and how much of it is fact ? HOw much better is the y67 test than the y37 test ? Is it worth it to upgrade ?

Neither fact nor opinion are really the proper way to look at it. The calculations are based on the probability two separate Y-DNA lines would randomly mutate to the same haplotype, and match each other when they're not truly related within a recent time frame. It's entirely possible you could have an exact match at 111 markers, and not share a common ancestor in 1000 years, but the odds of that happening are so slim that the alternative explanation, that you are closely related, is a better bet.

For situations like yours, I think upgrading could help answer some of your questions. You might be close matches at 37 markers, but have a much higher GD at 67. It will at least narrow the window for your possible MRCA.

[vinnie]

My question is, how did these other people even know to join our surname project to begin with? It's MY last name that is the name of the surname project and there are probably 20 people in my group or "cluster" . I apologize for having so many questions but I cannot get any help or answers from our so called " co ordinator" at all. The majority of the people in the cluster have the same last name as me but the two people I have the exact 37 marker matches with are different last names than me. But they are in the project too and they are in my cluster. But I have 5 matches on 36 of 37 markers and they have the same last name as me.

My response assumes that all of you share the exact same subclade. If your surname matches the majority of people in the project, and these two don't, these two people are descended from the same common male ancestor that you and the others are, but because they don't share your surname at sometime in the past, there may have been a name change, or an adoption, or a birth out of wedlock, or any other scenario that could result in a male child not having his birth father's surname. It could go back as far two different branches of the same family adopting different surnames at the time when people started using surnames. This is a great example of how DNA testing can provide new information on a family's heritage that might never have been found through traditional genealogy.

[Yaffa]

My response assumes that all of you share the exact same subclade. If your surname matches the majority of people in the project, and these two don't, these two people are descended from the same common male ancestor that you and the others are, but because they don't share your surname at sometime in the past, there may have been a name change, or an adoption, or a birth out of wedlock, or any other scenario that could result in a male child not having his birth father's surname. It could go back as far two different branches of the same family adopting different surnames at the time when people started using surnames. This is a great example of how DNA testing can provide new information on a family's heritage that might never have been found through traditional genealogy.

This person keeps asking the same questions over and over.Its been explained in several different ways by several people. Either this person is just not accepting the truth and not reading all the posts which explain everything or they are trolling the forum.

[507]

I do not understand all the abbreviations and words you guys use. Makes it very difficult to folliow

[vinnie]

I do not understand all the abbreviations and words you guys use. Makes it very difficult to folliow

Did you try reading the information at the links I provided you with? If you did and you still don't understand, give us a list of abbreviations and words we're using that you don't understand, or give the FTDNA office a call tomorrow and ask for help from one of the reps.

[efgen]

I've been following this thread and it appears that it keeps going in circles Maybe we need to go "back to basics" here.

507, please read through this carefully.

There are two major factors at play here:

A) DNA Mutations.

These happen randomly, they're part of our biology. Nothing we can do to change that. They happen between a parent and a child, but they definitely do not happen every generation.

While father and son *usually* are an exact match on the Y chromosome, these random mutations do happen periodically. If mutations didn't happen ever, then the entire human population would be clones. So a father and a son are typically an exact match, but there are some father/son pairs who mismatch on one marker due to these random mutations. Rarely will you see a father/son who mismatch on more than one marker, but it's not an impossibility.

Now on the flip side, since mutations happen randomly and do *not* happen every generation, and may not happen for several generations -- two people could be descended from the same ancestor several generations back and still be an exact match.

So that's why nobody can give you the *exact* number of generations that you keep asking for. The timeframes are given in probabilities because of the random nature of these mutations. So we can't say that you are definitively related X generations ago based on your match. However, what we can say is that 50% of the time, or 90% of the time, or 95% of the time, two people with a certain match level will be related within a certain number of generations.

So using the chart in this FAQ:

http://www.familytreedna.com/faq/answers.aspx?id=4#565

We can say that for any *exact* matches that you have at 37 markers, there's a 95% chance that you share a common ancestor within the past 7 generations (meaning from 1 to 7 generations). But there's still a 5% chance that the common ancestor is beyond 7 generations back. This 5% is not completely negligible -- it can't be ignored as though it was 0%.

If you mismatch on two of the 37 markers, then (based on the chart) we can say that there's a 95% chance that you share a common ancestor within the past 14 generations. But still, there's a 5% chance that the common ancestor is beyond 14 generations.

These numbers aren't simply *guesses* -- rather, they're estimates based on studies of father/son pairs.

B) Surnames.

There are a number of reasons why two people with different surnames might match at 37 markers. Here are some of the most common ones:

1) Adoption
2) Deliberate surname change
3) Surnames inherited recently (less than ~300 years ago)
4) Patronymics used as surnames, which change each generation (ie, John Jacobson is the son of Jacob Williamson)
5) Infidelity/misattributed paternity
6) Child took mother's maiden name

There are probably others.

Now, in your situation, given all the information that you provided, I would say the following:

1) You said that you match a bunch of other people with your own surname, at the 37-marker level. Assuming these are distant cousins from other branches of your family tree (ie, not just your own father and close family), then this is EXCELLENT news. There's no reason to believe that your family tree is incorrect. If it was, you wouldn't be matching so many people with your own surname.

2) Two of your matches have different surnames. THEY are the odd ones out, so THEY have had one of the above situations happen to cause their different surname. I would guess that these two guys don't match anyone with their own surname.

3) You asked how these two people joined your surname project? I can only guess, but they probably contacted the project administrator to ask if they could join, given that they matched so many people with your surname. FTDNA doesn't stop someone from joining a project with a different surname. It's up to each project administrator to decide how to manage their own project and who to allow to join the project.

Does this all make sense?

If you have more questions, please ask them, but please be specific. It doesn't help for you to write emotionally-charged posts and outright reject the DNA results and what we're telling you. You're not going to learn and understand if you've already made up your mind that the DNA results are useless.

Elise

[507]

Elise thank you so much for explaining. I understand better now.

[507]

You didnt have to do that. I appreciate it.