PDA

View Full Version : Verifying Two Branches Descend From Same Ancestor (18th Cent)



royaljoker
10-12-2019, 09:26 AM
Hi guys, Im new to this forum,

I wanted to ask if there is anyway to confirm Patrilineal descent between 10th Cousins. I belong to Junior branch of a family of some historical importance (nobility). I was looking to confirm our descent through a genetic test with the main branch. The thing is though that we branched off from the main branch around mid 18th Century. And there are now 10 generations including me, after this common ancestor. I've done a geographic 2.0 test out of general curiosity and am currently in the process of transferring it to ftdna. But I've been told that it only shows ancestors 5 generations back so I'm not entirely sure how or which test I can use to verify our lineage.

For more clarity my family tree is something like this (excluding names for confidentiality):
1) Common Ancestor (Should be the MRCA with the 10th cousin if the family history is correct)
2) Male
3) Male
4) Male
5) Male
6) Male
7) Male
8) Male
9) Male
10) Male
11) Me (Male)

My 10th Cousin is at the same level as me, and also a male line descendant. At most I could get an Uncle to test himself in order to reduce the generational gap, but that's still a test between 9th Cousins once removed. I have a Haplogroup from my Geno 2.0 test but perhaps thats too vague to narrow it down. I know a number of other noble families that also have the same haplogroup.

I would really appreciate any help.
Thanks

royaljoker
10-12-2019, 12:10 PM
Hi guys, Im new to this forum,

I wanted to ask if there is anyway to confirm Patrilineal descent between 10th Cousins. I belong to Junior branch of a family of some historical importance (nobility). I was looking to confirm our descent through a genetic test with the main branch. The thing is though that we branched off from the main branch around mid 18th Century. And there are now 10 generations including me, after this common ancestor. I've done a geographic 2.0 test out of general curiosity and am currently in the process of transferring it to ftdna. But I've been told that it only shows ancestors 5 generations back so I'm not entirely sure how or which test I can use to verify our lineage.

For more clarity my family tree is something like this (excluding names for confidentiality):
1) Common Ancestor (Should be the MRCA with the 10th cousin if the family history is correct)
2) Male
3) Male
4) Male
5) Male
6) Male
7) Male
8) Male
9) Male
10) Male
11) Me (Male)

My 10th Cousin is at the same level as me, and also a male line descendant. At most I could get an Uncle to test himself in order to reduce the generational gap, but that's still a test between 9th Cousins once removed. I have a Haplogroup from my Geno 2.0 test but perhaps thats too vague to narrow it down. I know a number of other noble families that also have the same haplogroup.

I would really appreciate any help.
Thanks

I would appreciate if anyone could recommend any test. Especially if anyone knows if a standard consumer test like Myheritage or ancestry, If done by both me and my cousin, would allow us to verify our descent from the same person.

Thanks

Pylsteen
10-12-2019, 01:15 PM
If both you and your far cousin are from the same paternal line, the best thing would be take a Y-chromosome test; you can opt for such a test at Family Tree DNA or Yseq.

asquecco
10-12-2019, 01:47 PM
I would appreciate if anyone could recommend any test. Especially if anyone knows if a standard consumer test like Myheritage or ancestry, If done by both me and my cousin, would allow us to verify our descent from the same person.

Thanks

If the paternal line is involved, you need a YDNA test.
By the way, I tested a distant paternal cousin, with genealogical common ancestor born in 1606 (exactly 11 levels in both lines, like in your case).

royaljoker
10-12-2019, 02:18 PM
If the paternal line is involved, you need a YDNA test.
By the way, I tested a distant paternal cousin, with genealogical common ancestor born in 1606 (exactly 11 levels in both lines, like in your case).

Which test did you use ? Me and this cousin should share a common Y-DNA . This common ancestor was a guy who was born at the start of the 18th Century and died mid 18th Century. We then descend from different sons.

Anyway how many SNP markers do you guys think I should check. And do you guys know how accurate it would be. My results with ftdna are pending transfer, but once they're complete do you guys think that ftdna family finder and its features like "Chromosomal Browser" and "Advanced Matches" would be at all useful?

asquecco
10-12-2019, 02:48 PM
Which test did you use ? Me and this cousin should share a common Y-DNA . This common ancestor was a guy who was born at the start of the 18th Century and died mid 18th Century. We then descend from different sons.

Anyway how many SNP markers do you guys think I should check. And do you guys know how accurate it would be. My results with ftdna are pending transfer, but once they're complete do you guys think that ftdna family finder and its features like "Chromosomal Browser" and "Advanced Matches" would be at all useful?

In my case, just a simple YDNA STR 37 test was enough to confirm our genealogical paternal connection.
Family Finder will not give you any final answer on the paternal connection.

msmarjoribanks
10-13-2019, 08:32 PM
If you aren't interested in Y-DNA for any other reason, I'd start with Y-37. If you have a match, you can then test deeper to try to confirm how close/identify subclades. You can also see who the other matches are and explore from there.

royaljoker
11-22-2019, 11:07 PM
Thanks guys for all your help. I recently did a Y-37 test. I'm still waiting on my cousin to do it but i've got my results in the meantime. I wanted to ask a couple of things. There are two guys who's results I can also see on a FTDNA research page and who I've been matched with at the 12 STR level. Basically the genetic distances that I've worked out manually, would for the other level would be:
Person A:
12 STR:0
25 STR:-5
37 STR:-6

Person B:
12 STR: 0
25 STR:-5
37 STR:-5

Is it likely at all that I'm related to these two guys ? (they seem to have a lot more in common with each other though)

Also while going through FTDNA pages I cam a cross a guy who seemed to be quite close to me however he didn't come up as a match. He scored the following:

Person C:
12 STR: -4
25 STR: -3
37 STR: -5

For the other levels of testing he seems to match me much better. Is there any possibility that he could be related to me and do you guys think I should do further testing like 67 or 111 to find out. Of course all the above guys belong to the same branch of the Haplogroup as me.

RobertCasey
11-23-2019, 02:41 AM
If you upgrade one of the Y37 to Y67, you could run your YSTRs through the YSNP Predictor NEVGEN. Around 10 to 15 % of the time, Y37 may work. Around 5 % of the time you may have to upgrade to Y111 if there is significant convergence with other haplogroups. It would also help if you post three things about your test results:

1) Surname of your oldest proven ancestor with place and date
2) The FTDNA ID
3) Some project that you have joined (this assumes that your data is public and the project has public YSTR reports).

Generic advice is not always the best compared to having more information that you could provide which would enhance the possible advice.

Also, YDNA is your only solution as atDNA tests at ten generations would be an extremely long shot in dark to give you any information.

I no longer recommend Y37 as the first test for the following reasons:

1) False hits go down from 40 to 60 % to 10 to 30 % with Y67 (the FTDNA match lists are almost always too optimistic).
2) With Y67, you can almost always get reliable YSNP prediction around 50 % of the time (there are just too many
haplogroups to analyze and some are so small that prediction would be problematic). This prediction is for haplogroup
R, the rarer the haplogroup, the less likely that it has been analyzed for YSNP prediction.
3) Eventually, somebody that belongs to your haplogroup may create a chart via SAPP or manually. Y67 is required for
charting as Y37 is extremely low accuracy for charting. Here is an example of a chart for R-L226:

http://www.rcasey.net/DNA/R_L226/Haplotrees/L226_Home.pdf#Page=26

royaljoker
11-23-2019, 12:35 PM
Well, honestly I was hoping to keep things private. But I'm gonna need advice so here it goes.

I'm a minor Indian prince of a north Indian muslim dynasty. Nawab Ali Muhammad Khan of Rohilkhand (https://en.wikipedia.org/wiki/Ali_Mohammed_Khan) is my earliest confirmed ancestor.

The thing is though, that he is supposed to have been adopted by his afghan chief foster father. This fact wasn't known by my family, and so my branch which branched of quite early and became quite isolated and had forgotten much of our history other then the royal lineage had always believed until recently that we were Pathan (Pashtun). Very early records do indicate that he was adopted but there have been various claims including that he was Ahir, Jatt or etc.To be honest the Ahir and Jatt theories have mostly been espoused by enemies of the dynasty. Around the end of the 18th century the main branch of the dynasty started claiming Alid Quraishi lineage, saying that he was adopted from a Quraishi family. Which wasn't known to us until recently. Why or how they came to that conclusion I honestly don't know and maybe it was valid.

Now I've done the Y-37 test. I'm R-L657, the same as the royal family of Arcot (also Quraishi but of a non-Alid line). In my matches at the 12 STR level I haven't matched with any Ahir or etc, I mostly seem to be matched with Quraishi Arabs . Members of the R1a Arabia. I haven't been matched with anyone at the 25 STR and 37 STR levels.

However manually combing through the R1a research page I've found another person separate form persons A,B and C, a Bahraini Quraishi. and the distances between me and him would be:
12 STR: 0
25 STR:-5
37 STR:-4

Would it be likely that this guy is related to me.? And hence would it confirm that at least in the last 600 years or so that we do indeed descend from a dynasty that 600 years ago claimed Quraishi lineage.
I'm hesitant to post my STRs and would rather keep them private.
Thanks

Dave-V
11-23-2019, 01:15 PM
95% of the time two people who share common descent and a genetic distance of 4 at 37 markers are related within 28 generations. That’s not very precise, but genetic distance is only a vague estimator of relationships.

It is possible with STRs to have an effect commonly called “convergence”, which means that STRs have randomly mutated along one line to adopt similar enough values as another line that the two lines look more closely related than they really are. This effect is commonly seen at Y37 especially in heavily-tested haplogroups like R1b whose branches have not heavily diverged from each other. But it can happen in any haplogroup and so you can’t rule out that your apparent close match with this Bahrainian Quraishi is coincidental.

If you had a small group of matches you can usually rule out convergence by examining the actual STRs that differ among the group and confirm that the mutation history makes sense (statistically and based on any known genealogy). In your case since you (so far) really only have this one possible match, you may need to upgrade to higher numbers of markers or an advanced SNP test to confirm the relationship. Convergence is MUCH less of an issue even at Y67 levels although it’s still possible.

Bear in mind that while you CAN confirm a close relationship Y-DNA will never confirm the exact number of generations back to your common ancestor with a match; it is all statistical with wide error margins and especially imprecise when dealing with small numbers of kits.

RobertCasey
11-23-2019, 04:11 PM
Dave - There are two forms of convergence that leads to false positive matches:

1) The traditional "convergence" as you described is much less common than we originally believed. When predictable haplogroups overlap each other due to drifting into each other's signatures or never drift away for the modal values of the older haplogroups, this is what you are describing which is turning out to be less than five percent of all testers. Predictable haplogroups are between 1500 to 2500 YBP but the FTDNA match list is around 1000 YBP.
2) Within predictable haplogroups, there is another form of convergence which I have labeled "lack of divergence" which is a much more serious of an issue. This causes an average of 10 to 30 % false positives vs. the 5 % that FTDNA implies. For some testers, the false positives can reach over 90 %. This has been verified across at least first 20 haplogroups that I have analyzed in depth. Almost every haplogroup analyzed has this issue (but some are worse than others).

The FTDNA model for YDNA match lists and 95 % confidence intervals with the Tip tool are just too optimistic about matches and make the vast majority of testers appear to be closer than they really are in reality. These FTDNA models assume that our genetic genealogy data are normal distributions. There is a measurement of normalcy called kurtosis. I have ran several haplogroups for this measurement and they are either very close 2.0 (normalized kurtosis) or just above 2.0. According to statistics, any model based on normal distributions must be rejected if the models regularly produce kustosis of 2.0 (normalized where 2.0 is subtracted). So any statistician would reject the FTDNA models since they routinely fail the Kurtosis test for normalcy.

However, statisticians are purists and the models can be used if you add some common sense via genetic distance filters. These filters can be based on two pieces of information - 1) surnames of testers (Irish and English) exceeding a 50 % NPE rate; 2) confirmed haplogroups that are incompatible that exceed 50 %. This issue is covered in my presentation:

https://www.youtube.com/watch?v=pmv7RyrVTa4

Dave-V
11-23-2019, 05:09 PM
Dave - There are two forms of convergence that leads to false positive matches:

1) The traditional "convergence" as you described is much less common than we originally believed. When predictable haplogroups overlap each other due to drifting into each other's signatures or never drift away for the modal values of the older haplogroups, this is what you are describing which is turning out to be less than five percent of all testers. Predictable haplogroups are between 1500 to 2500 YBP but the FTDNA match list is around 1000 YBP.
2) Within predictable haplogroups, there is another form of convergence which I have labeled "lack of divergence" which is a much more serious of an issue. This causes an average of 10 to 30 % false positives vs. the 5 % that FTDNA implies. For some testers, the false positives can reach over 90 %. This has been verified across at least first 20 haplogroups that I have analyzed in depth. Almost every haplogroup analyzed has this issue (but some are worse than others).

The FTDNA model for YDNA match lists and 95 % confidence intervals with the Tip tool are just too optimistic about matches and make the vast majority of testers appear to be closer than they really are in reality. These FTDNA models assume that our genetic genealogy data are normal distributions. There is a measurement of normalcy called kurtosis. I have ran several haplogroups for this measurement and they are either very close 2.0 (normalized kurtosis) or just above 2.0. According to statistics, any model based on normal distributions must be rejected if the models regularly produce kustosis of 2.0 (normalized where 2.0 is subtracted). So any statistician would reject the FTDNA models since they routinely fail the Kurtosis test for normalcy.

However, statisticians are purists and the models can be used if you add some common sense via genetic distance filters. These filters can be based on two pieces of information - 1) surnames of testers (Irish and English) exceeding a 50 % NPE rate; 2) confirmed haplogroups that are incompatible that exceed 50 %. This issue is covered in my presentation:

https://www.youtube.com/watch?v=pmv7RyrVTa4


Robert, I don’t necessarily agree with your comments that the FTDNA model and TiP tool are too optimistic about matches and make “the vast majority” of testers appear closer. Nor I believe is your comment correct about kurtosis leading to rejection of all the statistical models in use today (and many of which I might add which have been written up in peer-reviewed publications without that particular point ever being raised).

Also by the way lack of divergence is particularly an issue for R1b, but the original poster isn’t in that haplogroup. It’s especially an issue in the more heavily-tested haplogroups like R1b where there are lots of branches in short amounts of time. How much of an issue it is is really haplogroup-specific depending on branches discovered to date.

However ALL of this is off-topic for the original poster’s question. They just want to know if their Y37 possible match is a confirmed match, and you and I certainly I think agree that there are reasons they should still consider it POSSIBLE rather than confirmed until further testing of both themselves and their match yields more data.

RobertCasey
11-24-2019, 07:50 AM
Robert, I don’t necessarily agree with your comments that the FTDNA model and TiP tool are too optimistic about matches and make “the vast majority” of testers appear closer. Nor I believe is your comment correct about kurtosis leading to rejection of all the statistical models in use today (and many of which I might add which have been written up in peer-reviewed publications without that particular point ever being raised).

Please watch the YouTube video which clearly shows a dozen examples of just how bad the FTDNA match lists are (and the Tip Tool based on same math).

The first three that I analyzed (Y67):

L226 - O'Brien - Surname signature - one mutation - false positives = 91.1 % - with GD = 5, only reduced to 83.6 % false positives - these testers have very low genetic distance from the L226 signature
L226 - Casey - Surname signature - five mutations - false positives = 83.6 % - with GD = 4, reduced to 5.2 % false positives - these testers have very high genetic distance from the L226 signature
A5629 - Gleason (Maurice Gleason's surname cluster) - false positives = 74.4 % - with GD = 4 has 33 % false positives and GD = 3 has 0 % false positives - genetic distance filter could eliminate all false positives

What are the peer reviewed publications ?


These are clearly no where close to 5 % that FTDNA advertises.

Note that Maurice's haplogroup has some convergence which few haplogroups have.

royaljoker
11-24-2019, 11:08 AM
Can I also ask. If I have a common 12th Ancestor with someone. What is the highest level that I should match with this person. At 111 STR? and also what would be the minimum STR level that I would have to match with this person for them to have a MRCA at 12 generation.I'm assuming that at anything including and above 67 STR there won't be convergence right ? and can it happen that I don't match with someone at a 12 or 25 or 37 STR level but do match at 67 and above.

Dave-V
11-24-2019, 02:05 PM
Can I also ask. If I have a common 12th Ancestor with someone. What is the highest level that I should match with this person. At 111 STR?. and also what would be the minimum STR level that I would have to match with this person for them to have a MRCA at 12 generation.

Roughly someone who shares an ancestor 12 generations back with you (as opposed to a 12th cousin which is 14 generations back but a similar answer) could be a genetic distance of anywhere from 3 to about 14 at 111 markers. Although of course the matches don't show on the report higher than a genetic distance of 10 at 111 markers.

At the Y67 level the same match could be a genetic distance of 1 to 9, although matches don't show on the Y67 match report higher than a genetic distance of 7.

There is no minimum STR level to match; what happens as you increase the number of markers is that the range of generations represented by each genetic distance "step" gets a little smaller so in that sense higher number of markers are more "precise" than lower numbers.



I'm assuming that at anything including and above 67 STR there won't be convergence right ?

It can still happen but it's usually rarer. It does depend greatly on what your haplogroup is because some are more prone to convergence than others (and no, there isn't a cross-reference, it's just something that people more familiar (for instance perhaps the admins of the major haplogroup project that covers yours) could tell you).

But anecdotally I would say it is normally less than 5% of matches at Y67. I would bet though people will have examples of worse effects even at Y67 but I would argue they're not the norms.


and can it happen that I don't match with someone at a 12 or 25 or 37 STR level but do match at 67 and above.

Yes it's possible but it usually only happens with distant matches.

Realize that you actually "match" EVERY kit in the FTDNA database since all men are related to each other (currently the common ancestor of all men is placed ABOUT 260,000 years ago). But obviously the reports don't show 750,000 matches (the approximate number of current FTDNA Y-DNA testers).

That's because the reports stop showing matches after a certain genetic distance in order to limit the reports to just your closest ones. At Y37 the match limit is 4, at Y67 it's 7, and at Y111 it's 10. Matches with a higher genetic distance than that are cut off.

What happens is sometimes you have a match that misses the cut-off at one level but makes it at another and they "suddenly" appear as a match. They were ALWAYS a match really, they just missed the lower report. So for example if a match at Y37 is a genetic distance of 5, they won't show on that report. Then you upgrade to Y67 and they're now a genetic distance of 7, and they appear on the Y67 report.

royaljoker
11-24-2019, 02:26 PM
Roughly someone who shares an ancestor 12 generations back with you (as opposed to a 12th cousin which is 14 generations back but a similar answer) could be a genetic distance of anywhere from 3 to about 14 at 111 markers. Although of course the matches don't show on the report higher than a genetic distance of 10 at 111 markers.

At the Y67 level the same match could be a genetic distance of 1 to 9, although matches don't show on the Y67 match report higher than a genetic distance of 7.

There is no minimum STR level to match; what happens as you increase the number of markers is that the range of generations represented by each genetic distance "step" gets a little smaller so in that sense higher number of markers are more "precise" than lower numbers.




It can still happen but it's usually rarer. It does depend greatly on what your haplogroup is because some are more prone to convergence than others (and no, there isn't a cross-reference, it's just something that people more familiar (for instance perhaps the admins of the major haplogroup project that covers yours) could tell you).

But anecdotally I would say it is normally less than 5% of matches at Y67. I would bet though people will have examples of worse effects even at Y67 but I would argue they're not the norms.



Yes it's possible but it usually only happens with distant matches.

Realize that you actually "match" EVERY kit in the FTDNA database since all men are related to each other (currently the common ancestor of all men is placed ABOUT 260,000 years ago). But obviously the reports don't show 750,000 matches (the approximate number of current FTDNA Y-DNA testers).

That's because the reports stop showing matches after a certain genetic distance in order to limit the reports to just your closest ones. At Y37 the match limit is 4, at Y67 it's 7, and at Y111 it's 10. Matches with a higher genetic distance than that are cut off.

What happens is sometimes you have a match that misses the cut-off at one level but makes it at another and they "suddenly" appear as a match. They were ALWAYS a match really, they just missed the lower report. So for example if a match at Y37 is a genetic distance of 5, they won't show on that report. Then you upgrade to Y67 and they're now a genetic distance of 7, and they appear on the Y67 report.

Thanks, and I guess if I match with someone who has missed one of the cut-offs. then that person is more distantly related or maybe less precise and more likely to be a convergence result.

And at the 37 STR level. Should I be a full match with someone who has the same ancestor as me at 12 generations.

There are Indian Quraishis who I have matched at 12 STRs but I've completely ignored them as they either don't have higher testing or I haven't matched at a higher level, especially because for most of these guys I can't find their STR's on a FTDNA research page. Is there a good chance that they are related to me, if they are an exact match at 12 STR but they don't show up for any higher test. Especially if one considers that the Quraishi dynasty that we are meant to belong to is meant to be one of the oldest of India. Maybe even having arrived in about the end of the 10th Century. Could it mean that these guys are related to me but just very far back.

And thanks again for all your help. Its really invaluable !

Dave-V
11-24-2019, 02:40 PM
Please watch the YouTube video which clearly shows a dozen examples of just how bad the FTDNA match lists are (and the Tip Tool based on same math).

The first three that I analyzed (Y67):

L226 - O'Brien - Surname signature - one mutation - false positives = 91.1 % - with GD = 5, only reduced to 83.6 % false positives - these testers have very low genetic distance from the L226 signature
L226 - Casey - Surname signature - five mutations - false positives = 83.6 % - with GD = 4, reduced to 5.2 % false positives - these testers have very high genetic distance from the L226 signature
A5629 - Gleason (Maurice Gleason's surname cluster) - false positives = 74.4 % - with GD = 4 has 33 % false positives and GD = 3 has 0 % false positives - genetic distance filter could eliminate all false positives

What are the peer reviewed publications ?


These are clearly no where close to 5 % that FTDNA advertises.

Note that Maurice's haplogroup has some convergence which few haplogroups have.

Robert I moved this discussion to the thread where you originally posted the video, so that we don't occupy the original poster's thread with the tangent.

See https://anthrogenica.com/showthread.php?18573-TMRCA-estimates-based-on-surname-cluster-milestones-and-variable-years-per-YSNP&p=624441&viewfull=1#post624441

Dave-V
11-24-2019, 02:49 PM
And at the 37 STR level. Should I be a full match with someone who has the same ancestor as me at 12 generations.

It's impossible to say. They might be a close genetic distance match (probably 3 or so at best, less likely 0,1, or 2) or they might be so high a genetic distance that they don't show up on the report.

If you have someone that you think is related 12 generations back, upgrade to the highest levels of markers you can or better yet, take SNP tests like the Big Y, and look for whether the additional markers (STR or SNP) continue to suggest that the two of you are closely related. Y-DNA won't ever PROVE it's 12 generations though.


There are Indian Quraishis who I have matched at 12 STRs but I've completely ignored them as they either don't have higher testing or I haven't matched at a higher level, especially because for most of these guys I can't find their STR's on a FTDNA research page. Is there a good chance that they are related to me, if they are an exact match at 12 STR but they don't show up for any higher test. Especially if one considers that the Quraishi dynasty that we are meant to belong to is meant to be one of the oldest of India. Maybe even having arrived in about the end of the 10th Century. Could it mean that these guys are related to me but just very far back.

I'm sorry I just don't take 12 marker comparisons as any indication. In my view they're only good to show that people are so far apart even on 12 markers that they CAN'T be related to each other. But to me there's just too much "noise" at 12 markers to make any type of prediction as to how related they might be.

But if those people don't show up as matches at the higher number of markers, then they're not related to you (at least not closely. NORMALLY people show up as matches at the higher levels if they have 16th century common ancestors, that's USUALLY too close and results in genetic distances within the report thresholds. So I would say people that don't show up as matches at Y37 or certainly higher numbers of markers do not normally share common ancestors from the 16th century. But it's not IMPOSSIBLE.

RobertCasey
11-24-2019, 03:21 PM
Genetic distance match lists are somewhat optimistic from FTDNA. Lower resolution are especially low quality matches. It is encouraging that you have no Y12 matches and that your Y25 and Y37 matches are quite consistent. This is pretty unusual as Y12 is normally 10X of Y25 and Y25 is normally 5X of Y37. So your matches are hanging in there which means that they have a better chance of being real matches. Y67 is the gold standard and I recommend that upgrade which will surely reduce the number of matches and make them higher quality. Only upgrade to Y111 if several of your Y67 matches have tested to Y111.

I assume that you are predicted to be R-L657 (it would red on the YSTR report). According to YFULL, the TMRCA of R-L657 4200 YBP (years before present). This is way too old to be much of use for genealogical matches. With Y67 in hand, you can use the YSNP predictor tool, NEVGEN which is very reliable if your line pretty prolific in numbers or well tested (it is manual to add one haplogroup at a time so only larger haplogroups get predicted). But this prediction could get you down to a YSNP branch in the 1500 to 2500 YBP time frame which could be used a genetic distance filter if it does not match. Posting Y37 would not be that helpful as Y67 is really needed for any thorough analysis.

Upgrading to Y67 would give you a better chance to separate Arab and Indian matches. If your Y67 matches have several Y111 tests, upgrade to Y111 as well. Sounds like the European R1a does not show up but the Arab and Indian has not been filtered out with Y37. You should go to YFULL, click on R1a, then search for L657. Look at the flags for the branches under L657.

https://www.yfull.com/tree/R1a/

There are some lower branches where Indian and Saudi are separated but there is a lot of geographic diversity shown with the dozen flags posted for L657. It looks like L657 for Arab/Indian testers is the R1b of Europeans - very broad across many countries. Looking at L657 at YFULL really reveals how diverse L657 really is. Big Y700 could get you assigned to one of the lower branches and would probably reveal a new branch (you have to pay $50 to get placed on the YFULL tree). You can look at the public FTDNA haplotree as well - but it is very difficult to navigate (as usual). FTDNA has even more diversity of flags than YFULL. Your chain for FTDNA is (you have to open each branch):

https://www.familytreedna.com/public/y-dna-haplotree/R

M173 > M420 > M459 > M198 > M417 > PH6162 > Z283 > Z93 > Z94 > F2597 > M636 > L657

royaljoker
11-24-2019, 03:59 PM
Genetic distance match lists are somewhat optimistic from FTDNA. Lower resolution are especially low quality matches. It is encouraging that you have no Y12 matches and that your Y25 and Y37 matches are quite consistent. This is pretty unusual as Y12 is normally 10X of Y25 and Y25 is normally 5X of Y37. So your matches are hanging in there which means that they have a better chance of being real matches. Y67 is the gold standard and I recommend that upgrade which will surely reduce the number of matches and make them higher quality. Only upgrade to Y111 if several of your Y67 matches have tested to Y111.

I assume that you are predicted to be R-L657 (it would red on the YSTR report). According to YFULL, the TMRCA of R-L657 4200 YBP (years before present). This is way too old to be much of use for genealogical matches. With Y67 in hand, you can use the YSNP predictor tool, NEVGEN which is very reliable if your line pretty prolific in numbers or well tested (it is manual to add one haplogroup at a time so only larger haplogroups get predicted). But this prediction could get you down to a YSNP branch in the 1500 to 2500 YBP time frame which could be used a genetic distance filter if it does not match. Posting Y37 would not be that helpful as Y67 is really needed for any thorough analysis.

Upgrading to Y67 would give you a better chance to separate Arab and Indian matches. If your Y67 matches have several Y111 tests, upgrade to Y111 as well. Sounds like the European R1a does not show up but the Arab and Indian has not been filtered out with Y37. You should go to YFULL, click on R1a, then search for L657. Look at the flags for the branches under L657.

https://www.yfull.com/tree/R1a/

There are some lower branches where Indian and Saudi are separated but there is a lot of geographic diversity shown with the dozen flags posted for L657. It looks like L657 for Arab/Indian testers is the R1b of Europeans - very broad across many countries. Looking at L657 at YFULL really reveals how diverse L657 really is. Big Y700 could get you assigned to one of the lower branches and would probably reveal a new branch (you have to pay $50 to get placed on the YFULL tree). You can look at the public FTDNA haplotree as well - but it is very difficult to navigate (as usual). FTDNA has even more diversity of flags than YFULL. Your chain for FTDNA is (you have to open each branch):

https://www.familytreedna.com/public/y-dna-haplotree/R

M173 > M420 > M459 > M198 > M417 > PH6162 > Z283 > Z93 > Z94 > F2597 > M636 > L657

I am getting some european matches at the 12 STR level. But I completely ignored them.
I used Nevgen and got L657 ->Y4-> y6->Y5 as the most likely haplogorup. Interestingly the arabs I'm being matched with are also Y5. I'm not sure if Y5 is a particularly indian marker. I read somewhere that Y6 is more arab. But on Yfull, Y5 comes up as 4100 years old. So I guess it would be much more diverse.

geebee
11-27-2019, 03:07 AM
All I can tell you is my own experience. I have no idea how typical -- or a-typical -- it is.

At FTDNA, I have no matches at all above 37 markers. At 37 markers, I have only a single match, with a genetic distance of 2. But the most significant part is that my match has a known variant of the same rare surname that I have.

The original form was "Buchhammer", which is taken from a place name which would now be written as "Buchheim". But that is due to the standardization of German and German place names. In the Austro-Bavarian dialect, "ham" is the form used in place of "heim".

My surname uses "Book" in place of "Buch" but retains the Bavarian ending "-hammer". My match's surname also uses "Book", but the more standardized ending "-heimer". However, this was apparently a fairly recent development on my match's side -- it was evidently his father or grandfather who made the change.

Regardless of the slight variation in the spelling of the surname, we're fairly closely patrilineally related. Our most recent common Y-line ancestor was a 4th great grandfather to both of us -- a son of the man who brought the German form of the surname to America in 1749.

The Tip tool gives only a 77.24% percent probability for the connection to be within only six generations, yet it is. So the tool obviously is not always "optimistic".

Interestingly enough, my first actual Y DNA match was at Ancestry, on a test that included several more markers (46). My match, who has since passed away, had exactly the same surname I do, as well as the distinction of being a former lieutenant governor of Delaware.

Ancestry showed a genetic distance of 1 in my match list online, but I found a reporting error in a download of that list. Online, two copies of a multi-marker STR were displayed, as DYS459a and DYS459b. For me these values were 9 and 10, while for my match they were shown as 10 and 10.

However, the download showed that both my match and I actually had a third copy of this marker, DYS459c. My value for this marker was listed as 10, while my match's value was listed as 9. The trouble is, Ancestry didn't know the actual order of these repeats, and the convention was to report them in numerical order -- from least to greatest. Therefore, the order should have been 9, 10, 10 for both of us, which meant that the online match list should have reported DYS459a = 9 and DYS459b = 10 for both of us.

I reported this to Ancestry and they conceded their error and corrected the online match list -- something I doubt they'd do today. But this correction did have one odd effect. When there was supposedly a genetic distance of 1, Ancestry predicted that our most recent common Y-line ancestor was six generations back. After the GD was corrected to 0, the number of generations changed to just one. So my match and I became brothers!

In actuality, the match was from my grandfather's generation. Our most recent common Y-line ancestor was the immigrant ancestor himself -- my match's 3rd great grandfather and my 5th great grandfather, so we're 4th cousins twice removed.

A point I'd like to make is that STR mutations can occur at any time, regardless of any "estimated mutation rate" for that marker. It just happened that -- apparently -- there were no mutations between my 5th great grandfather and me. But, that means no mutations on any of the specific 46 markers in Ancestry's test.

Obviously, there has been at least one mutation on either my FTDNA match's side or mine -- or both -- since our 4th great grandfather. As I recall -- though I can't now confirm it -- the mutation or mutations did not involve a marker (or markers) among the 46 used by Ancestry. So there's no way to know which side it was on, or whether it was a single mutation involving two repeats, or two different mutations. (I would suspect the former.)

RobertCasey
12-02-2019, 05:03 AM
Ancestry showed a genetic distance of 1 in my match list online, but I found a reporting error in a download of that list. Online, two copies of a multi-marker STR were displayed, as DYS459a and DYS459b. For me these values were 9 and 10, while for my match they were shown as 10 and 10.

However, the download showed that both my match and I actually had a third copy of this marker, DYS459c. My value for this marker was listed as 10, while my match's value was listed as 9. The trouble is, Ancestry didn't know the actual order of these repeats, and the convention was to report them in numerical order -- from least to greatest. Therefore, the order should have been 9, 10, 10 for both of us, which meant that the online match list should have reported DYS459a = 9 and DYS459b = 10 for both of us.

Having extra copies of multi-copy markers can be very useful if you match others that also have extra copies of markers. These extra copies are created via a massive mutational event in a special area that can sometimes recombine. This event is RecLOH and if you have extra copies of 459, you will regularly also have extra copies of 464 and CDY markers as well.

https://en.wikipedia.org/wiki/RecLOH

These markers are located in the area of the Y chromosome called the palindromic region:

https://en.wikipedia.org/wiki/Palindromic_sequence

Once you get the markers of your matches, matching the extra copies can make very good connections. But these extra copies are short lived as after not many generations the tend to revert back to original number of marker values. There are two ways to handle these. If you have access to the YSTRs of your matches, you can look at common extra copies which could be very useful. But most of your matches will lose these extra copies and in order to determine your real genetic distance, there are two approaches: 1) just truncate the higher marker values and only look at the lower values that match the standard length (this is not recommended); 2) Normalize these markers by deleting pairs back down to one value.

For your 459 = 9, 10, 10 set of markers, they could be normalized back to two markers. The only choices are 9,10 and 10,10. These extra copies are usually pairs, so it is more likely to be 9,10 but could be 10,10 as well. Normalizing data allows you to compare apples to apples better. It helps with genetic distance, YSNP prediction and charting. But you are throwing away data that could be very relevant for a short time. My fairly large haplogroup project, R-L226, has nineteen testers out of 711 YSTR testers that show extra copies of markers. Most of these are independent RecLOH mutational events. 464 markers tend to persist the longest and with very high mutation rates, CDY tends to go back to two values very quickly. Also, all three do not necessarily have extra copies during the RecLOH event.

If most of your matches are 9,10, this is probably your original values as well. If most of your matches are 10,10, this is probably your original values.

geebee
12-03-2019, 08:39 PM
Having extra copies of multi-copy markers can be very useful if you match others that also have extra copies of markers. These extra copies are created via a massive mutational event in a special area that can sometimes recombine. This event is RecLOH and if you have extra copies of 459, you will regularly also have extra copies of 464 and CDY markers as well.

https://en.wikipedia.org/wiki/RecLOH

These markers are located in the area of the Y chromosome called the palindromic region:

https://en.wikipedia.org/wiki/Palindromic_sequence

Once you get the markers of your matches, matching the extra copies can make very good connections. But these extra copies are short lived as after not many generations the tend to revert back to original number of marker values. There are two ways to handle these. If you have access to the YSTRs of your matches, you can look at common extra copies which could be very useful. But most of your matches will lose these extra copies and in order to determine your real genetic distance, there are two approaches: 1) just truncate the higher marker values and only look at the lower values that match the standard length (this is not recommended); 2) Normalize these markers by deleting pairs back down to one value.

For your 459 = 9, 10, 10 set of markers, they could be normalized back to two markers. The only choices are 9,10 and 10,10. These extra copies are usually pairs, so it is more likely to be 9,10 but could be 10,10 as well. Normalizing data allows you to compare apples to apples better. It helps with genetic distance, YSNP prediction and charting. But you are throwing away data that could be very relevant for a short time. My fairly large haplogroup project, R-L226, has nineteen testers out of 711 YSTR testers that show extra copies of markers. Most of these are independent RecLOH mutational events. 464 markers tend to persist the longest and with very high mutation rates, CDY tends to go back to two values very quickly. Also, all three do not necessarily have extra copies during the RecLOH event.

If most of your matches are 9,10, this is probably your original values as well. If most of your matches are 10,10, this is probably your original values.

My entire point was that this was a reporting error on the part of AncestryDNA -- they had simply put my cousin's copies of DYS459 in an unconventional order. By convention, we both should have had the three copies of DYS459 listed in this order: 9, 10, 10. However, following your suggestion of "normalizing" these values, we would both still end up with 9, 10 -- regardless of whether we truncated the highest value of "10" or eliminated the apparently duplicated copy, which is also "10".

Incidentally, I do have extra two extra copies of DYS464, as did my match at Ancestry. They were listed as DYS464a 12, DYS464b 12, DYS464c 15, DYS464d 15, DYS464e 15, and DYS464f 16. Currently, FTDNA simply shows my DYS464 as 12, 12, 15, 15, 15, 16. (Since they no longer use the labels "a", "b", "c", etc.)

I wish I could tell the STR values of my matches. Unfortunately, FTDNA no longer shows where any mismatches are located -- they just give genetic distance.

If I select "show matches for the entire database", I only have one match at 37 markers, and none above that. This match has a GD of 2 and actually is related to me, but he hasn't tested more than 37 markers. He and I are 4th cousins once removed. My match's 3rd great grandfather is my 4th great grandfather.

At 25 markers, I have four additional matches, all with a GD of "2". But, again, I have no way of knowing how we compare at individual markers. I also have quite a few additional matches at 12 markers, but since DYS459 isn't among the first 12 markers -- it's a moot point.

EDIT:

I vaguely remember sending FTDNA an email about DYS459, since they only showed DYS459a = 9 and DYS459b = 10, while Ancestry showed three values. I asked whether my results at FTDNA actually had included three values as well, and as I recall their only answer was, "That's possible".

2nd EDIT:

My YDNA results are also somewhat unusual in that when I first submitted my Big Y 500 results to YFull, YP619 was considered to be at the same level as YP444. However, I was positive for YP619 but negative for YP444. At the next update, YFull showed YP619 and YP444 as being at different levels, with YP444 below YP619. FTDNA has since adopted this same change, and my haplogroup is shown there as R-YP619*.

I'm sure I don't "own" the haplogroup -- but so far I seem to be the only one in it. In actuality, of course, I have two brothers and a host of close patrilineal relatives (cousins, uncles) who would likely have the same terminal SNP if they were to test for it.