U106*

[1798]

I entered the modal for U106* in Cullen's Haplotest. The results were
S28=72%, R1b=21%, N/S1=5%, R1b-ub=1%. I didn't know that these two subclades were so close.
I entered my own markers and the results were R1b=69%, S28=15%,
S26=10%, N/S 2= 5%, Leinster=2%. I'm neither S28 or S26.

[MMaddi]

Most subclades of R1b1b2 only differ on 3 or 4 markers out of 67. That's why, if you want to know your subclade, you won't be able to tell (with one exception, see below) with any degree of probability by looking at STR markers. You need a deep clade test to know.

The one exception is DYS492. If you have DYS492=13, there's a 95% probability that you're U106+ (aka S21 ). If you have DYS492=12, that generally means you're U106- and probably P312+. If you have DYS492=14, there's a good chance that you're U152+ (aka S28), although many U152+ men have DYS492=12.

Even given the difficulty of predicting R1b1b2 subclade based on STR markers, I'm not impressed at all by the Cullen predictor. It may do a good job with haplogroup I, but I wouldn't take its predictions for R1b1b2 subclades very seriously. I've seen too many times when it's been way off.

[1798]

Most subclades of R1b1b2 only differ on 3 or 4 markers out of 67. That's why, if you want to know your subclade, you won't be able to tell (with one exception, see below) with any degree of probability by looking at STR markers. You need a deep clade test to know.

The one exception is DYS492. If you have DYS492=13, there's a 95% probability that you're U106+ (aka S21 ). If you have DYS492=12, that generally means you're U106- and probably P312+. If you have DYS492=14, there's a good chance that you're U152+ (aka S28), although many U152+ men have DYS492=12.

Even given the difficulty of predicting R1b1b2 subclade based on STR markers, I'm not impressed at all by the Cullen predictor. It may do a good job with haplogroup I, but I wouldn't take its predictions for R1b1b2 subclades very seriously. I've seen too many times when it's been way off.

It would be great though if some smart man could come up with a Haplo-"R" subclade predictor. As far as deep clade tests go a man could be out money trying out all the new SNPs downstream of U106.
Your point about 492 = 13 for U106 seems right enough.The thing is there are people in the "Q" haplogroup who have 13 at 492 as well.It could be the original value for that STR. Didn't R and Q have the same father? It could mean that U106 is older than P312!!

[MMaddi]

It would be great though if some smart man could come up with a Haplo-"R" subclade predictor. As far as deep clade tests go a man could be out money trying out all the new SNPs downstream of U106.

Until some new marker or markers come along with the same predictive value of DYS492 for R1b1b2 subclades, I think R1b1b2 subclade predictors based on STR values will be a futile effort. There are plenty of STR markers outside the standard 67 available on FTDNA's Advanced Orders menu. I think you can probably get up to 100 markers tested between the standard 67 and the Advanced Orders STR markers.

Perhaps when many more customers order those markers for testing we'll find some useful markers for subclade predicting. But that still means spending more money, without the assurance of finding one or two markers that will do what you want them to do.

My attititude is, if you're interested in knowing your subclade, you might as well spend the money on the deep clade test and new SNPs as they come along. At least that way you're certain and not just relying on probabilities associated with an STR marker.

Your point about 492 = 13 for U106 seems right enough.The thing is there are people in the "Q" haplogroup who have 13 at 492 as well.It could be the original value for that STR. Didn't R and Q have the same father? It could mean that U106 is older than P312!!

Of course when I brought up DYS492 for predicting R1b1b2 subclade, I was referring to a situation where you already had a good idea that you're dealing with an R1b1b2 haplotype.

However, I think most of the more knowledgeable people who analyze R1b1b2 haplotypes believe that the original value for DYS492 in R1 was 12. Take a look at the values for DYS492 in the R1a Project at http://www.familytreedna.com/public/...ction=yresults, the R1b1* Project at http://www.familytreedna.com/public/...ction=yresults and the R1b1b1 Project at http://www.familytreedna.com/public/...ction=yresults. In all three projects, involving a "distant cousin" (R1a), the "grandfather" (R1b1*) and the "brother" (R1b1b1) of R1b1b2, the strong modal for DYS492 is 12.

This implies to me that the modal for R1, the distant ancestor of R1b1b2 and the others, also had DYS492=12 as its modal. Since R1 lived about 18,000 years ago, DYS492 is a very stable marker. This also implies that the MRCA of R1b-U106 men had a rare mutation from 12 to 13 at DYS492 and, since the marker is very stable, that has remained the strong modal in R1b-U106 men.