PDA

View Full Version : 67 Marker Test Genetic Distance Questions



J Man
11-20-2016, 01:32 AM
I am creating this thread for general questions about genetic distance for Y-DNA matches at 67 markers. I apologize in advance if a thread like this has already been created. My first question is this.

Generally speaking how far back in time (years) would the most recent common ancestor be for two individuals that match on 61/67 markers (genetic distance of 6 at 67 markers)?

J Man
11-21-2016, 12:01 AM
Bump.

Smilelover
11-21-2016, 12:11 AM
I see this link may help you
But i really also curious for the answer from our experts here ..

https://www.familytreedna.com/learn/y-dna-testing/y-str/two-men-share-surname-genetic-distance-67-y-chromosome-str-markers-interpreted/

There its said .. yes related

I think from my reading of many results
Distance is nearly from 1000 to 1500 yrs but dont take my estimation seriously.

I just give my expectation

J Man
11-21-2016, 12:15 AM
I see this link may help you
But i really also curious for the answer from our experts here ..

https://www.familytreedna.com/learn/y-dna-testing/y-str/two-men-share-surname-genetic-distance-67-y-chromosome-str-markers-interpreted/

There its said .. yes related

I think from my reading of many results
Distance is nearly from 1000 to 1500 yrs but dont take my estimation seriously.

I just give my expectation

Yes thank you I have seen that link before. It gives a good breakdown but it does not list anything in years. It is still good information though.

RCO
11-21-2016, 02:08 AM
Yes, 61/67 markers (genetic distance of 6 at 67 markers) is more or less 1000 years. That's the average result comparing STRs, SNPs Big Y and YFull's calculations.

J Man
11-21-2016, 04:02 AM
Yes, 61/67 markers (genetic distance of 6 at 67 markers) is more or less 1000 years. That's the average result comparing STRs, SNPs Big Y and YFull's calculations.

Thank you! :)

RobertCasey
11-21-2016, 04:59 AM
FTDNA declares a genetic distance of seven as being genetically related (within the surname creation date of around 1,000 years ago). However, the CDYa/b are extremely fast mutating markers and I do not use them for relatedness and allow the maximum of genetic distance of six without them. This works fine for very large clusters but not so well when only comparing two submissions. The statistical variation with a sample size of two will vary widely, so it is very important to build your cluster to get more reliable dating. I also employ another common sense approach for determining if a cluster of submissions are really related - reasonable NPE rates. If the cluster only has two surnames (allowing for one early NPE) and exceeds 50 % of only two surnames, they probably descend from one common male ancestor with that surname (in the case of two surnames - the most common but not always).

There is also the fact of convergence and lack of divergence. Around five to ten percent of matches just randomly overlap. Another five to ten percent do not mutate enough over the last 1,000 years are also false hits. This is why it is critical to YSNP test to reveal these false hits. Here is a YouTube of my recent presentation at the Genetic Genealogy Ireland conference. Just search YouTube for genetic genealogy Ireland 2016 for other great presentations.

https://www.youtube.com/watch?v=3w5DOexfgdg

Afshar
11-21-2016, 08:48 AM
The question with STR is (for me) always how big the deviation is, but in this case it is probably something in between 500-1500 years.

J Man
11-21-2016, 04:07 PM
FTDNA declares a genetic distance of seven as being genetically related (within the surname creation date of around 1,000 years ago). However, the CDYa/b are extremely fast mutating markers and I do not use them for relatedness and allow the maximum of genetic distance of six without them. This works fine for very large clusters but not so well when only comparing two submissions. The statistical variation with a sample size of two will vary widely, so it is very important to build your cluster to get more reliable dating. I also employ another common sense approach for determining if a cluster of submissions are really related - reasonable NPE rates. If the cluster only has two surnames (allowing for one early NPE) and exceeds 50 % of only two surnames, they probably descend from one common male ancestor with that surname (in the case of two surnames - the most common but not always).

There is also the fact of convergence and lack of divergence. Around five to ten percent of matches just randomly overlap. Another five to ten percent do not mutate enough over the last 1,000 years are also false hits. This is why it is critical to YSNP test to reveal these false hits. Here is a YouTube of my recent presentation at the Genetic Genealogy Ireland conference. Just search YouTube for genetic genealogy Ireland 2016 for other great presentations.

https://www.youtube.com/watch?v=3w5DOexfgdg

So would you say then that generally two individual men who are a 61/67 marker match (genetic distance of 6 at 67 markers) most likely share a most recent common ancestor somewhere in the range of 500-1500 years ago?

GoldenHind
11-21-2016, 08:02 PM
So would you say then that generally two individual men who are a 61/67 marker match (genetic distance of 6 at 67 markers) most likely share a most recent common ancestor somewhere in the range of 500-1500 years ago?

I would generally agree with that, but only if the two have the same terminal SNP. This is necessary to rule out false matches due to convergence.

Also STR markers mutate at different rates. Some mutate at a comparatively fast rate (such as CDY, which I also generally ignore), others are very slow. The time to the MRCA will be influenced by the number of fast vs. slow markers which differ between the two individuals in question.

J Man
11-21-2016, 08:10 PM
I would generally agree with that, but only if the two have the same terminal SNP. This is necessary to rule out false matches due to convergence.

Also STR markers mutate at different rates. Some mutate at a comparatively fast rate (such as CDY, which I also generally ignore), others are very slow. The time to the MRCA will be influenced by the number of fast vs. slow markers which differ between the two individuals in question.

So if the two individuals that match 61/67 do not have the same terminal SNP then the most recent common ancestor could be much further back in time?

GoldenHind
11-22-2016, 12:14 AM
So if the two individuals that match 61/67 do not have the same terminal SNP then the most recent common ancestor could be much further back in time?

Correct, if one has actually tested positive and the other negative for the SNP in question. If the difference in terminal SNPs is due to the fact only one of them has tested for it, obviously that wouldn't apply, as without testing, there is no way of knowing whether the other person has the same terminal SNP or not.

In the former case, the MRCA would date to the period between the origin of the last SNP common to both, and the origin of the SNP only one of the two carries.

J Man
12-04-2016, 04:33 AM
Recently a friend of mine had his Y-DNA 67 marker results come in at Family Tree DNA. At 67 markers he has a 61/67 match with a man who has a different surname. They both come from Montenegro but from different regions and like I said they have different surnames. They belong to Y-DNA haplogroup J2a-M92. Anyway they have a 61/67 marker match so that is a genetic distance of 6. I am not sure exactly on which markers they differ but here is their TiP results at 67 markers. I am just going to list their initials.


AS vs. MP:


COMPARISON CHART
Generations Percentage
4 3.34%
8 29.38%
12 64.39%
16 86.44%
20 95.76%
24 98.85%



I know that this is all probabilities but based on this how far back in time do they most likely share a MRCA?

RobertCasey
12-04-2016, 03:51 PM
COMPARISON CHART
Generations Percentage
4 3.34%
8 29.38%
12 64.39%
16 86.44%
20 95.76%
24 98.85%


This summary is the average probability over a large sample size (think 100 to 1,000). What this summary does not show is the statistical variation which can be very large. I have first cousins in my genetic cluster that have a genetic distance of three and yet also have seventh cousins with 67/67 matches (even with CDY values) in the same cluster, so it can vary a lot.

Here is my updated presentation on the relationship of YSNPs and YSTRs which includes examples of convergence (false matches that can not share common cousins earlier than 4,500 years) - convergence. It also shows a newly discovered "lack of divergence" over the last 1,000 to 2,000 years where genetic distance of 1 to 4 can not be related earlier than 1,000 to 1,500 years.

The best YDNA presentations can be found from the recent Genetic Genealogy Ireland 2016 conference. Just search YouTube for "Genetic Genealogy Ireland 2016" to find a list of all the great presentations that will really help understand what latest trends in YDNA testing:

https://www.youtube.com/watch?v=3w5DOexfgdg

(https://www.youtube.com/watch?v=3w5DOexfgdg)

jova
10-24-2017, 03:14 PM
according to FTDNA my 67 marker matches with a GD of 6 have a 99.7% probability of having a common male ancestor within the last 24 generations...which would be within the last 700 years and a 98.68% probability of having a common ancestor within the last 20 generations.

are the FTDNA TIP calculations inaccurate ? Should we assume the relationships are not within the last 700 years ?