Predicted Y-haplos: who/what to SNP

[Sydni]

With implementation of the new Y-tree my surname project now has 4 subgroups in which no members have tested SNPs so they have only predicted haplogroups - and those differ for subgroup members from the same documented genealogic families.
As administrator how do I determine which man (or men) from each subgroup would be the best representative to SNP test and then, from the extensive pre-loaded list of SNPs - how do I determine which would be most informative to test (since my group fund will be paying).
Suggestions?

[lgmayka]

Anyone considering the purchase of SNP tests should join a haplogroup project and seek its administrator's assistance.

[Earl Davis]

Hi Syndi. What is the surname project you administer?

Earl.

[Sydni]

Earl,
I'm the administrator of the Thurman & Variations Project and in the past I have done as lgmayka suggested and have consulted with people like Steve Gilbert because I have an R-U152 subgroup. However, I'd like to be able to educate myself to better advise the Thurman participants since many are not actively engaged in seeking out new tests.

[Earl Davis]

I am not sure my reply will be particular useful and is probably something you have already done.

In a situation like this where no one within a sub group in a surname project has SNP tested I would log in as GAP and look at the matches one by one that each member has at Y67 or Y111. Some of them will probably have a number of matches outside your project's surname that fall in the period just before when surnames started to form. I would then try and find those matches in their relevant surname project and invariably some of them will be sitting in a group where someone has tested.

For example in the WYNN project (of which I am not a member) there is a large group that have not SNP tested. If the GAP went through all 12 or so members in that sub group they would find that two of them just about match me at Y67. They could then find me in the DAVIS project. They would see that I tested DF27+ and based on their close genetic distance to me they could take an educated punt on tested for DF27. They would also see a large number of SNP's that I am negative for that they might initially avoid. They would also find that I have BIGY tested and have a named 'recent' SNP less than 2,500 years they may be interested in exploring.

Earl.

[Sydni]

Thanks for the input Earl,
I follow my group's matches closely but I'm not sure that looking at mixed surname STR matches is the way to determine who/what SNPs to test.
In one of my subgroups which has several close mixed surname matches at Y67 and Y111 we're using SNP testing to rule out convergence as a reason for the matches.
I'm surprised this predicted terminal SNP question isn't an issue in other projects.
Are most participants with predicted haplogroups testing the entire pre-loaded list?

[Sydni]

For any other surname project administrators wondering how to advise/learn about efficient SNP testing it appears that the FTDNA helpdesk is referring us to the Haplogroup administrators as Igmayka suggested above.
In a related question I asked the helpdesk why a haplogroup is shown as "predicted" for a man who had tested downstream SNPs (Z36) as a "stand-in" for a deceased known cousin (depleted DNA sample) who is R-U152 positive. I was told that the cousin would need to test R-U152 and contact administrator Steve Gilbert for further testing advice. Prior to the Z36 test the participant and I were already told by Steve that re-testing R-U152 for this sample would be redundant since the two men are known relatives.
It looks like even FTDNA is learning about how to advise on SNP testing...

[Sydni]

FTDNA now has a webinar discussing this issue. Essentially an administrator must look at the SNPs tested on a participant's "Haplogroup Origins" matches and look at the Haplotree to see which seem most likely. This is difficult if your participant has no tested close matches and in that case it's best to contact a haplogroup administrator who can often look at the STR pattern to suggest SNPs to test. Deep Clade will possibly be offered again. Hope this helps anyone else with this question.