GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2009-07 > 1248548522


From: Al Aburto <>
Subject: Re: [DNA] Fw: For Age Estimates beware of tossing STRs away
Date: Sat, 25 Jul 2009 12:02:52 -0700
References: <2B73FF420BBD42158EE8B08C1632DEBB@elizabethod> <200907250524.n6P5OTDf007016@mail.rootsweb.com> <DA07C2B8E26742D4BC6BD7D5513B3CB9@elizabethod><D7416507-65EC-4636-878A-21E1E16FEEA7@vizachero.com>
In-Reply-To: <D7416507-65EC-4636-878A-21E1E16FEEA7@vizachero.com>


I think of it differently I suppose (not sure), in terms of resolution
in time. Like a clock with on hour counter, a minute counter, and a
second counter. You need all three to get the best resolution in time.
When the second counter runs threw it's cycle we raise the count on the
minute counter by one and when the minute counter runs through it's
cycle we raise the hour counter up by 1 ... but to work properly you
need to calibrate the cycle lengths somehow ... else there is no way one
will be able to predict the hour from the minute and second ... seems
like a tough problem with Y-STRs ...
Al

> Vincent Vizachero wrote:
> The underestimation of TMRCA resulting from using too many fast
> markers is (in my view) a function of the behavioral property of STRs
> that restricts their ability to outside of a certain range. These
> range constrtaints lead to mutational saturation, so that variance
> starts to asymptote at some limit. If we knew the details of this
> asymptotic curve we could correct for this behavior, but we don't and
> that's what led me to propose to omitting the very fastest markers as
> a stop-gap.
>
> VV
>

This thread: