DNA-R1B1C7-L Archives

Archiver > DNA-R1B1C7 > 2012-04 > 1333427451


From: Mike Wdna <>
Subject: Re: [R-M222] FW: Geographical distribution of M222+
Date: Mon, 2 Apr 2012 23:30:51 -0500
References: <000101cd10a2$dbd729b0$93857d10$@com>
In-Reply-To: <000101cd10a2$dbd729b0$93857d10$@com>


Sandy, that's a great a analysis. I'm not a statistician but it seems
like reasonable perspective. I think you send this to the academics,
like Busby, Myres and just about all of them doing age estimates based
on 15 markers, or typically, even less.

Actually, a good checkpoint for you is Dienekes. He goes off on
complaints about STR diversity and seemed to agree with Busby that a
lot of markers aren't needed, just "good' (linear) ones.

Ken Nordtvedt seems to come down more on your side where his
simulations show that the more markers the better. He says we should
consider each STR as its own "experiment" and therefore it is only to
be expected to get better precision with more markers. Ken does throw
out multi-copy markers, though.

Regards,
Mike W


---------- Forwarded message ----------
From: Sandy Paterson <>
Date: Mon, Apr 2, 2012 at 2:33 AM
Subject: [R-M222] FW: Geographical distribution of M222+
To:


Here's an attempt at checking how many markers are required before reaching
the stage where estimated TMRCAs begin to look sensible. .....

Although this is just a crude examination, it is clear that 15 markers is
hopelessly too few. The TMRCA estimates using 15 markers are all over the
show. Even 50 markers is dicey, with the slow mutators producing a 25%
error....

This thread: