GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2007-06 > 1181580864


From: "Joe Knapp" <>
Subject: Re: [DNA] Identifying R1b clusters
Date: Mon, 11 Jun 2007 12:54:24 -0400
References: <a81622ac0706101447u50f0d562kedfbcca0476dd8ca@mail.gmail.com><BFECJOAEEPCFBFFLLBGPKEBPDDAA.scorpion@netconnect.com.au><a81622ac0706101758g74790840v7c7e2e0e2d1eea7e@mail.gmail.com><03ea01c7ac27$8aec1c20$0200a8c0@c452380a><a81622ac0706110626l3be22a0fsd72f08834605f2e1@mail.gmail.com><001d01c7ac30$f79ba990$6400a8c0@Ken1><a81622ac0706110739r6fcc508v1c972013ff801df@mail.gmail.com><004a01c7ac39$0b7f8b90$6400a8c0@Ken1><a81622ac0706110812s35abb744x447d63b72ac34459@mail.gmail.com><008301c7ac3d$7e1d4b20$6400a8c0@Ken1>
In-Reply-To: <008301c7ac3d$7e1d4b20$6400a8c0@Ken1>


On 6/11/07, Ken Nordtvedt <> wrote:
> In an earlier message I expressed my willingness to take the drink. I asked
> for the Ysearch codes for members of one of the unidentified "clusters" so I
> could look myself at the haplotypes to see what, if anything, they had to
> bring them together.

Well, given that that is impossible, it's in the "bring me the
broomstick of the Wicked Witch of the West" category. My impression is
that for privacy reasons there is no public linkage between the two
databases.

Seems to me that any method of identifying a cluster, as long as it is
reproducible using standard tools, is useful, whether or not some
simple cluster definition can be found. I mean, there are some crude
approximations, for example, taking the "modal" of the cluster and
figuring a haplotype's probability of being in that cluster as being
related to the distance from the modal, but that's not perfect by any
means. Same goes for looking for a specific collection of marker
values.

For example, take unidentified group A in the diagram. The modal for
that cluster is:

393,390,19,391,385a,385b,426,388,439,389-1,
392,389-2,458,459a,459b,455,454,447,437,448,
449,464a,464b,464c,464d,460,GATAH,YCA2a,YCA2b,456,
607,576,570,CDYa,CDYb,442,438,531,578,395S1a,
395S1b,590,537,641,472,406S1,511,425,413a,413b,
557,594,436,490,534,450,444,481,520,446,
617,568,487,572,640,492,565

13,24,14,11,11,14,12,12,12,13,
13,29,17,9,10,11,11,25,15,19,
28,15,15,15,17,11,11,19,23,16,
15,17,16,35,36,12,12,11,9,15,
16,8,10,10,8,10,10,12,23,23,
16,10,12,12,15,8,12,22,20,13,
12,11,13,11,11,12,12

Considering the "rare" values above relative to the typical values for
R1b, it can be seen that CDYb=36 is a pretty good differentiator, as
83% of the cluster has that value while only 7% of R1b generally does.
But no combination of such marker-by-marker comparisons exactly
reproduces the cluster.

Joe

This thread: