DNA-R1B1C7-L Archives

Archiver > DNA-R1B1C7 > 2009-12 > 1260643510


From: "Sandy Paterson" <>
Subject: Re: [R-M222] Coyne and Golden project families
Date: Sat, 12 Dec 2009 18:45:10 -0000
References: <mailman.25.1260432006.9391.dna-r1b1c7@rootsweb.com>, , ,<002a01ca7a39$3e11a9f0$aa00a8c0@Sisko>, , ,<000601ca7a65$27e4c4b0$77ae4e10$@com>, ,<BLU116-W13631884A35B92122FE703A98C0@phx.gbl>,<BLU116-W4A0F6353AA142FC91AF48A98B0@phx.gbl>,<000601ca7b41$6a566ce0$3f0346a0$@com><BLU116-W138D07E167C76F715AE3C4A98B0@phx.gbl>
In-Reply-To: <BLU116-W138D07E167C76F715AE3C4A98B0@phx.gbl>


I should be able to produce a table of sorts fairly soon, but from
simulation rather than a TMRCA model. I've changed my simulation software
code so that each simulated haplotype has a 'name'.

The fathers in each generation are numbered from 1 to Ng. The final
haplotype names contain the father number to whom they are linked in each
generation. That then permits a search to determine the most recent common
ancestor for each combination of two people. I should eventually be able to
produce a table showing mean TMRCA for each pair of two haplotypes
stratified by

Gd from modal of nearest to modal of the pair and
Gd between the pair.

That the plan, anyway.


Sandy





-----Original Message-----
From:
[mailto:] On Behalf Of Iain Kennedy
Sent: 12 December 2009 16:01
To:
Subject: Re: [R-M222] Coyne and Golden project families


It appears to propose a continuous spectrum of adjustments as it also says
far off-modal matches adjust the other way ie are closer than the standard
formula predicts. Its a shame he couldn't give us a nice handy table at the
end instead of leaving it as an exercise for the user. At the least it
implies that a close match near the modal is several (>2) times more distant
than a close match of the same GD at the far off-modal.

Iain



> From:
> To:
> Date: Sat, 12 Dec 2009 15:40:21 +0000
> Subject: Re: [R-M222] Coyne and Golden project families
>
> Yes, this is still a good bit short of what I believe to be the case.
>
> However, it does recognise that there is a problem with TMRCA estimates if
> one of the haplotypes is exactly equal to the modal. It also makes it
clear
> that the founder haplotype and the modal is considered to be one and the
> same thing. I suspect that this is because the observed modal can be shown
> to be the best estimate of the founder haplotype.
>
> Since TMRCA is a measure or proxy of 'relatedness', the implication is
that
> 'relatedness' estimates need to be adjusted if the modal haplotype is
> involved.
>
> There's something I don't quite follow, but this may be clarified in the
> full text of the paper. Note that the Abstract uses the wording "...if one
> of the favoured MRCA haplotype choices is the modal (founding)
haplotype..."
>
> This seems to imply that the estimation process involves taking two
possible
> haplotypes for the MRCA (presumably the two that are computed in some
> logical way to be the two most likely ones) and then estimating the time
it
> would take to mutate to the two haplotypes under consideration.
>
> Finally, note the bit that talks about the assumption that haplotype
> frequencies are constant through time.
>
> I think this comes from the concept of a 'stationary population', one in
> which births match deaths is such a way that the average age of the
> population remains constant. This implies that it is a 'mature'
population,
> which M222+ very certainly isn't.
>
>
> Sandy
>

>
> -----Original Message-----
> From:
> [mailto:] On Behalf Of Iain Kennedy
> Sent: 12 December 2009 12:33
> To:
> Subject: Re: [R-M222] Coyne and Golden project families
>
>
> The mathematical explanation of why you have to double the TMRCA if one of
> the samples is on the modal is explained in a JOGG paper by Ken Norvedt:
>
> http://www.jogg.info/42/files/Nordtvedt.htm
>
> Abstract
> The traditional
> estimates of age back to the most recent common ancestor (MRCA) for a pair
> of
> present-day Y-STR haplotypes implicitly assume that the haplotype
> possibilities for that
> MRCA have the same chance of being found in any past generation's
> population,
> i.e., that population's haplotype frequencies are constant through time.
In
> reality, haplotypes are found to cluster
> near clade modal haplotype with high frequency, and this requires those
> frequencies to be changing in time as the clusters tend to widen and their
> central peak frequencies fall. Using the
> full Bayesian rule of probabilistic inference, these time-changing
haplotype
> frequencies are taken into account, and as a result major changes to the
age
> estimates for MRCAs are derived.
>
> **In one extreme, if one of the favored MRCA
> haplotype choices is the modal (founding) haplotype of the clade to which
> the
> pair of present haplotypes belong, the time estimate back to the MRCA is
> doubled from the traditional estimate. **
>
> If, on the other hand, the pair of haplotypes have all their favored
> choices for MRCA haplotype many steps of mutation from the clade modal
> haplotype, the MRCA time estimate can be pushed toward the present and
> become
> younger than traditional estimates. This
> modified method for estimating a TMRCA can be viewed as triangulation in
> which
> a clade's founding haplotype joins with two
> present-day haplotypes in the process of best-estimating the location of
an
> intermediate node (the MRCA being sought) in a portion of the clade's
> phylogenetic tree.
>
>
>
> Of course, JOGG is not quite the same as an academic science journal.
>
> This stills seems to be some distance from what Sandy Paterson was
claiming?
>
> Iain Kennedy
>
> http://www.kennedydna.com
>
>
>
>
>
>
>


_________________________________________________________________
Add your Gmail and Yahoo! Mail email accounts into Hotmail - it's easy
http://clk.atdmt.com/UKM/go/186394592/direct/01/
R1b1c7 Research and Links:

http://clanmaclochlainn.com/R1b1c7/
-------------------------------
To unsubscribe from the list, please send an email to
with the word 'unsubscribe' without the
quotes in the subject and the body of the message

This thread: