GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2005-05 > 1115058179


From: (John Chandler)
Subject: Re: [DNA] Editing help
Date: Mon, 2 May 2005 14:22:59 -0400 (EDT)
References: <IGEOKAGLHNEKPCKPADIGMEJHJNAA.bbailey.lowedna@baileyconnection.com>
In-Reply-To: <IGEOKAGLHNEKPCKPADIGMEJHJNAA.bbailey.lowedna@baileyconnection.com>


Bill wrote:
> STR's are great for possibly 10-20 generations....

The word "great" may be a stretch, so let's focus on the concept of
"useful". A marker is useful whenever the probability is low for
having more than one mutation on that marker within the stated time.

At the nominal mutation rate of 0.002 per generation, the "typical"
STR is therefore useful for 50 generations. The likelihood of more
than one mutation in that time on two lines of descent is only about
1.7%. For a group study, where the ancestral haplotype can be
determined by comparing many samples, each person is compared against
the ancestor, instead of against his putative cousins -- it is
therefore only his one line of descent that counts, and so the useful
time span doubles to 100 generations.

Some markers mutate much faster than average and some much slower,
but the usefulness of a collection of markers follows the average.

> And because of this we keep sliding out onto thin ice using STRS to
> guesstimate haplogroup and sub-clades.... Many of the guesses may be
> close but many times these guesses are dead wrong.

You have to decide what's important. If your interest is purely in
genealogy, haplogroups have very little value. It may be fun to guess
which Son of Adam or which Daughter of Eve founded your line, but
you'll certainly never fill in the connection. Admittedly, there are
times when the haplgroup can be very helpful. For example, the RICE
surname is always described as being of Welsh origin, but we have some
RICE families with non-R1b haplogroups, even though Wales is
overwhelmingly R1b. From those facts we can deduce that a lot of
Rices are not genetically Welsh in origin after all -- but we *can't*
actually deduce that their surname isn't Welsh.

> (2.) To request FTDNA and other testing companies to test for deep-level
> SNPS. If this is not economically at the moment, then to search for
> technologies that will bring these SNP subclade test costs down.

If by "deep level" you mean "I1a" instead of just "I", that's within
reach now, but it still doesn't come anywhere near to genealogy. If
you mean "deep enough to approach the realm of genealogy," then the
problem isn't just cost. Current technology isn't capable of even
discovering enough SNPs in our lifetime to do the job, let alone
testing economically.

John Chandler

This thread: