STRs vs SNPs for determing subclades

[ElaineB]

I am a member of the very active J-1 project. My 2C1R FTDNA #400208 (REST) is BY68 FGC9941+. The project administrator is awesome, and/but I'd like another perspective.

There has been some discussion on the group page on how SNPS are more accurate than STRS for determining TMRCA.It was also explained how genetic distance does not necessarily indicate closeness of the MRCA and that someone with a GD of 6 may have a closer MRCA than someone with a GD of 1, due to the addition of SNPS info.

I would like your take on the following situation:

The project administrator has placed KNIGHT and REST and CARLEN (GDs of 2 and 3)in different 3 different subclades FGC9941+ and states that there is 400+/-48 years TMRCA. However, I have what I thought was very strong evidence (both paper and anecodotal) to indicate that these three share a common ancestor perhaps 250-300 years ago.

A STRS chart showing TMRCA does back up my genealogy work, but with the SNPS added into the mix, it changes radically, no longer corroborating the evidence.

Comments?

Thanks!
Elaine

[Armando]

STRs mutate randomly. High quality SNPs that can be shown to be negative in most people except for people that share a series of other SNPs can be used to prove branching. If the three people have STRs with a GD of 3 but have high quality SNPs that aren't shared by each other but each person shares those with others then each person belongs to a different branch.

However, the dating of the SNPs is based on estimates and until there have been more NGS testing of well documented family groups there is too little information available.

[ElaineB]

Thanks for your reply, Armando. I think I understand what you are saying. A little more info:

Only one of these three actually tested for FGC9941: REST, who was positive. There is one BigY person downstream, also FGC9941+.

These 2 did *not* test for FGC9941:
KNIGHT was placed in a FGC9941+ subclade all by himself, and CARLEN was placed in a different FGC9941+ subclade with two others.

REST was placed in a yet another FGC9941+ subclade with one other-- (who is a 67/6 STR match).

What I'm trying to understand is how the PA determines through the "computer models" he uses, that these three could not be in the same subclade. Can you tell me what he is looking at to make that kind of determination?

Also (probably a really beginner question), is my assumption that these people in the same FGC9941+ subclade are more closely related than they are to those in a different FGC9941+ subclade? Thanks.

[Adam]

Dear Elaine B,
I have stumbled upon this thread and would like to share some thoughts. I am currently running a new FTDNA project and one of the persons you mention, Knight, is in the project. Would you please consider joining the project so that I may perhaps shed some new light on your research? Here is the site: https://www.familytreedna.com/groups...-project/about. To join simply push the 'JOIN' button at the top and log into your FTDNA account. I'll follow-up once I have done a preliminary assessment. Thanks, Adam

[cjm]

STR's are usually only reliable for distance if you can also confirm the subclade. That being said, you can use STRs to narrow down possible candidates for who you are close to, but the final arbiter is going to be your SNPs. On YFull, I have people who are 7 distances from me in STRs, but not in my subclade at all. The closest guy in my subclade who shares my SNPS is over 50 GD away on his Big-Y.

In your case, unless the science gets updated, if you don't share the SNPs, then the common ancestor is the estimated time frame when their subclades intersect.