Y-37 Exact Match and a Mystery!

[liesahealy]

Hi All,

Just received results of a Y-37 test to verify a relationship between two men with a distinct, fairly uncommon Irish/English surname.

The tests show about a 50-60% chance that these two had a common ancestor within the last eight generations. There is a 99% chance that they are related within 28 generations.

Got it. Here's what I don't get, though.

Subject #1 had a great-grandfather, Richard Deneefe, born in Knocktopher, Kilkenny, Ireland around 1820.

Subject #2 had a great-great grandfather (also named Richard) born in Inistioge, Kilkenny (only 9 miles from Knocktopher) in 1821.

Yet - according to the test results - their MRCA would have lived many, many generations before that.

How is this possible? After all, the two Richards were practically the same age, with the same rare surname. Both hailed from the same fairly small geographical area.

I have yet to find a birth record for one of the Richards, so at this point I cannot establish their relationship via paper trail.

Wouldn't their MRCA be much more recent than a few hundred years?

Can someone please explain this?

Many Thanks,
Liesa

[Frederator]

Congratulations on your match.

But I wonder if an important detail may have falled between the cracks here. I've never gotten a statistically significant match at any level, but what you're describing sounds like a cumulative probability function. It's important to keep some details in mind:

"The tests show about a 50-60% chance that these two had a common ancestor within the last eight generations. "

Meaning that there is a 50-60% chance that your MRCA is NO MORE THAN eight generations ago.

Unfortunately, DNA will never be a super precise genetic clock. There is some normal genetic variation, and a smaller amount of non-normal genetic variation. That's why they necessarily must use what in different contexts would be considered "weasel words". What you describe here looks to me perfectly well within the "normal" range for a common ancestor in the early 1800's.

That said, remember that Irish naming practices in the 19th century were very conservative indeed. Sure, you come across a Toirdhealbhach or Ronan from time to time, but there's a disappointingly high proportion of John's, Edmonds and Patricks. It doesn't seem especially weird to me that you might connect a generation or two before 1820/1821, and that the guys you mention might actually be first, second or later cousins.

But again, congratulations on your match. With any luck, your shared unusual surname and maybe additional matches will take you back further.

Hi All,

Just received results of a Y-37 test to verify a relationship between two men with a distinct, fairly uncommon Irish/English surname.

The tests show about a 50-60% chance that these two had a common ancestor within the last eight generations. There is a 99% chance that they are related within 28 generations.

Got it. Here's what I don't get, though.

Subject #1 had a great-grandfather, Richard Deneefe, born in Knocktopher, Kilkenny, Ireland around 1820.

Subject #2 had a great-great grandfather (also named Richard) born in Inistioge, Kilkenny (only 9 miles from Knocktopher) in 1821.

Yet - according to the test results - their MRCA would have lived many, many generations before that.

How is this possible? After all, the two Richards were practically the same age, with the same rare surname. Both hailed from the same fairly small geographical area.

I have yet to find a birth record for one of the Richards, so at this point I cannot establish their relationship via paper trail.

Wouldn't their MRCA be much more recent than a few hundred years?

Can someone please explain this?

Many Thanks,
Liesa

[vinnie]

After all, the two Richards were practically the same age, with the same rare surname. Both hailed from the same fairly small geographical area.

I'm not familiar with Irish naming practices in the 19th century, but if they were anything like they were in Italy during the same time period, yes, it's entirely possible for these two men to be two different persons.

[507]

one thing i have learned from doing roots research is this. No name is rare. No first name is rare and no last name is rare. We may think that it is but in reality, there are probably 15 people named utychndfyvneuvmehvuke that all live in the same town and drive the same kind of pickup truck.

[liesahealy]

Misunderstanding!

I think a couple of you have misunderstood me. We already know for a fact that these Richards are not the same person.

One emigrated to the US; the other, to Australia.

I think it's very likely that they were far more closely related than the testing shows.

So, I'm trying to understand how it appears that they are not.

[Frederator]

I think my comments are pretty on target, but I'll reiterate if it helps to state it another way.

DNA is not a precise genetic clock. It cannot identify the specific way or point in time at which the two men were related. Not this test, not at this resolution. All that can be said is that the probability is high (50%-60%) that the two donors shared a MRCA WITHIN THE LAST EIGHT GENERATIONS (~240 years, or ~1800 if donors are ~60 give or take).

Not that the donors DEFINITELY share a MRCA born in 1800. Only that there is a 50%-60% chance that their MRCA is no more than 8 generations back. There's an unavoidable uncertainty here. The 50%-40% odds that their relationship is MORE THAN EIGHT GENERATIONS AGO is not neglible, but we're gonna have to live with the uncertainty. It's a fact of life.

Again, congratulations on your match. It proves pretty conclusively that you're related. And it is far, far better than many testers ever achieve. You may hope to identify ancestors further generations back and I hope you eventually do. But DNA can never prove out the precise relationship or timeframe. That will depend on a strong paper trail.

I think it's very likely that they were far more closely related than the testing shows.
.

[nathanm]

I think a couple of you have misunderstood me. We already know for a fact that these Richards are not the same person.

One emigrated to the US; the other, to Australia.

I think it's very likely that they were far more closely related than the testing shows.

So, I'm trying to understand how it appears that they are not.

Fred's right, the probabilities mean the MRCA is within so many generations. When a mutation happens in a paternal line, that father and son end up having a GD of 1. How close is the actual match, i.e. how many markers match out of how many tested?

[Jim Barrett]

The tests show about a 50-60% chance that these two had a common ancestor within the last eight generations. There is a 99% chance that they are related within 28 generations.

I don't know where you are coming up with your numbers. I checked two pair of exact 37 marker matches and both show a 83.49% probability of a common ancestor within 4 generation. This is much closer than what you show.

How is this possible? After all, the two Richards were practically the same age, with the same rare surname. Both hailed from the same fairly small geographical area. Many Thanks, Liesa

The key word PROBABILITY. These estimates are based on averages. What if your two men aren't average? If you drop 10 coins, on the average 5 will land heads up and 5 will land tails up. Give it a try. Drop 10 coins 10 times and record the number of heads and tails each time. How many times did you get 5 heads and 5 tails?

[sjadelson]

I don't know where you are coming up with your numbers. I checked two pair of exact 37 marker matches and both show a 83.49% probability of a common ancestor within 4 generation. This is much closer than what you show.

Depends on what values you want to assume for mutation rate. The McDonald calculator says 62.4% at 4 generation, 90% at ~9 generations. You can disagree with his assumptions, but he does know what's he's talking about.

[vinnie]

Also remember that: mutations can happen at any time; there are some studies that show certain families have higher mutation rates; depending on the haplogroup, some DYS markers may be more variable than they are in other haplogroups. These two men could be related much more recently than estimates based on average mutation rates.

[507]

I don't know where you are coming up with your numbers. I checked two pair of exact 37 marker matches and both show a 83.49% probability of a common ancestor within 4 generation. This is much closer than what you show.



The key word PROBABILITY. These estimates are based on averages. What if your two men aren't average? If you drop 10 coins, on the average 5 will land heads up and 5 will land tails up. Give it a try. Drop 10 coins 10 times and record the number of heads and tails each time. How many times did you get 5 heads and 5 tails?

Where is the proof for this MRCA timeframe we are givin? I have 2 exact matches at 37 markers and we have no clue how we could be related. Our family trees are as separate as night and day.

[Daveb]

Cousins, Maybe!

To a researcher my cousin and I would cause similar problems, we have identical names, live about 40 miles apart and tracing our fathers takes you back to the same Devonshire town and to a common grandfather. However as I was an adopted child the DNA would fail to show us being in the same family tree, I have to use two trees,one for the paper trail shown by all legal records and one for the facts. That is part of life's rich tapestry I was told. You may have the reverse situation of a true ancestor with a false paper trail.
Good luck
Dave

[on the edge]

I don't necessarily trust the numbers: I have two relatives who share a common ancestor within eight and ten generations, both established through rigorous paper trails. Neither of them match me or each other closely because they both have (different) multi-step mutations. This throws both of them off the scales and they do not show up at any matching level. As far as the calculations are concerned they are are not related at all, despite sharing the same unusual name and geographical area. I also do not rely on the accepted generation calculation. The standard appears to be based on an average of 25 years but in my case is more like 34 years. This will also wildy obscure the 'reality' of certain situations. I know this is all about probabilities, but for some (possibly more than we know) the improbabilities are perhaps much more likely.

Also remember that: mutations can happen at any time; there are some studies that show certain families have higher mutation rates; depending on the haplogroup, some DYS markers may be more variable than they are in other haplogroups. These two men could be related much more recently than estimates based on average mutation rates.

[canaryislandgirl]

new Y page format problems

Is anybody else having problems reading the matches on the new Y matching page ? The sections are so squished together at the 37 marker match that I can't even scroll to see the name. And you can forget getting the 67 marker or the 111 marker area to show - the sections are nearly touching with no space between them to see the name. I hope the IT department fixes this - this is TERRIBLE. Why in the world would they go to this type of formatting ? It was fine before.

[canaryislandgirl]

Y page glitch

Well, I logged out and logged back in and now the Y page reads correctly. I guess it was a wierd glitch in the system.