Haplogroup Designation on "Y Results"

[Itzhak Epstein]

The "Y Results" STRs of several members of my surname project, myself included, are close. One of us was deep-tested and was identified as "R1b1b2h". Accordingly, FTDNA marked each of our haplogroup designations under the personal page's "Ydna results" tab as "R1b1b2h".

However Our project's various results tables, including "Y Results", show all of the untested members as "R1b1".

Do we have to wait for FTDNA 2.0 to get it right??

[efgen]

Itzhak,

On the public Y Results pages, FTDNA only displays predicted haplogroups to the depth that it can be confident of, based on its prediction algorithm. FTDNA predictions are based on 12-marker matches only. So if you go to the Haplogroup page of one of your members who is showing as R1b1 on the Y Results page, I would expect that his SNP-confirmed 12-marker matches range from R1b1 to R1b1b2h. R1b1 would be the least common denominator, so that's the prediction that would be shown.

Elise

[Itzhak Epstein]

Itzhak,

On the public Y Results pages, FTDNA only displays predicted haplogroups to the depth that it can be confident of, based on its prediction algorithm. FTDNA predictions are based on 12-marker matches only. So if you go to the Haplogroup page of one of your members who is showing as R1b1 on the Y Results page, I would expect that his SNP-confirmed 12-marker matches range from R1b1 to R1b1b2h. R1b1 would be the least common denominator, so that's the prediction that would be shown.

Elise

Thanks Elise:

1. If the "public" pages are based on 12 markers, how does FTDNA calculate the often different personal pages' results?

2. As can be seen from my project's results -- http://www.familytreedna.com/public/...xed_columns=on -- our (historic Epsteins) 12 markers are very unique, and we look like clones, and one of us was deep tested. What algorithm?

Some projects are more significantly affected by this problem. Charles Kerchner's R1b project -- http://www.familytreedna.com/public/...xed_columns=on -- does not attempt to sub-group by haplogroup, but would benefit from it. Sean Silver's Jewish R1b project-- http://www.familytreedna.com/public/...xed_columns=on -- attempts (with your help?) to identify internal groups by haplogroup but seemingly not with FTDNA's help.

3. My main concern is with the relevance of haplogroup identification. If R1b1b2h separated from R1b1b2 well before the bronze age, then they have very little to do with each other for even ancient genealogy, much less for conventional genealogy. Thus, Sean (for instance) and I are probably pursuing different patrilineal research investigations. I am still adjusting to the significance of the relatively (few years) recent scientific differentiation of R1b1b2's subclades. Explaining to the various Epstein family historians how what is being told about R1b1b2h's origins relates to my family's origin myths will not be easy.

Itzhak

[efgen]

Hi Itzhak,

Both the private and public predictions are based on 12 markers, but apparently FTDNA errs on the side of caution when it shows the prediction publicly, and therefore doesn't show the deeper prediction without a very high confidence level.

I don't know the exact algorithm that FTDNA uses, but I do know that it's based on the confirmed haplogroups of the exact and near 12-marker matches. If there's one SNP-confirmed match who is only tested to the R1b1b2 level, while other matches are SNP-confirmed to the R1b1b2h level, then FTDNA can't say with 100% confidence that the predicted person would also be R1b1b2h.

Although you and I are aware that your 12-marker results are very unique, the prediction algorithm doesn't see that -- it simply sees that the predicted person has SNP-confirmed matches at differing levels of R1b1b2 and makes the best prediction it can based on that information.

R1b1b2 is the ancestor of R1b1b2h. So it's not an incorrect prediction, it's just not as specific as it could be.

Elise

[darroll]

This is where the National Geographic comes to play.
They take your DNA and match it to a person on the ground somewhere
around the world. d

[Itzhak Epstein]

Hi Itzhak,

Both the private and public predictions are based on 12 markers, but apparently FTDNA errs on the side of caution when it shows the prediction publicly, and therefore doesn't show the deeper prediction without a very high confidence level.

I don't know the exact algorithm that FTDNA uses, but I do know that it's based on the confirmed haplogroups of the exact and near 12-marker matches. If there's one SNP-confirmed match who is only tested to the R1b1b2 level, while other matches are SNP-confirmed to the R1b1b2h level, then FTDNA can't say with 100% confidence that the predicted person would also be R1b1b2h.

Although you and I are aware that your 12-marker results are very unique, the prediction algorithm doesn't see that -- it simply sees that the predicted person has SNP-confirmed matches at differing levels of R1b1b2 and makes the best prediction it can based on that information.

R1b1b2 is the ancestor of R1b1b2h. So it's not an incorrect prediction, it's just not as specific as it could be.

Elise

Yes Elise, you make sense, but...

So FTDNA's algorithms seem to be set to two paths -- conservative and somewhat speculative. It would be nice if the default public data would be set on the cautious side and project administrators would be able to reset them to the liberally interpreted option. How does one communicate with Adrian about including such a feature in FTDNA 2.0?

Itzhak

[Itzhak Epstein]

I have just noticed the following note above my project's public y-dna results table:

[B]DNA Test Results (Alleles) for Project Members

* Haplogroups in green have been confirmed by SNP testing. Haplogroups in red have been predicted by Family Tree DNA based on unambiguous results in the individual's personal page. This has been placed on this GAP page for your ease and convenience. Please note that for any predicted results we see no reason for ordering a SNP test to confirm the Haplogroup. if a

[Itzhak Epstein]

My previous attempt at this post was garbled by the system. It is the dash that I enclosed in quotes now that had messed it up,Here is a second try.

I have just noticed the following note above my project's public y-dna results table:

[B]DNA Test Results (Alleles) for Project Members

* Haplogroups in green have been confirmed by SNP testing. Haplogroups in red have been predicted by Family Tree DNA based on unambiguous results in the individual's personal page. This has been placed on this GAP page for your ease and convenience. Please note that for any predicted results we see no reason for ordering a SNP test to confirm the Haplogroup. if a "

[Itzhak Epstein]

My previous attempts at this post were garbled by the system. It is the dash that had messed it up,Here is a third try.

I have just noticed the following note above my project's public y-dna results table:

DNA Test Results (Alleles) for Project Members

* Haplogroups in green have been confirmed by SNP testing. Haplogroups in red have been predicted by Family Tree DNA based on unambiguous results in the individual's personal page. This has been placed on this GAP page for your ease and convenience. Please note that for any predicted results we see no reason for ordering a SNP test to confirm the Haplogroup. if a [dash] is in the HAPLO field then we feel that the comparative results are not clear and unambiguous and if the kit holder wants to know their SNP with 100% confidence they may consider ordering a SNP confirmation test.


This means that, while I am R1b1b2h, my R1b1 is the best as I can get and that nobody should urge me to take a deepSNP test. I, of course, know better, but not everybody does.