GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2011-12 > 1323195406


From: Vince Tilroe <>
Subject: Re: [DNA] Correct TMRCA analysis (Europe)
Date: Tue, 6 Dec 2011 11:16:46 -0700
References: <137601ccb3b3$48d55480$da7ffd80$@org><BFECJOAEEPCFBFFLLBGPGEJFIKAA.dna@irishtype3dna.org><CAKWx04Sxhd4FAmtared6esCbrMTapEBXS+eaSpcHCKryCwe8FA@mail.gmail.com>
In-Reply-To: <CAKWx04Sxhd4FAmtared6esCbrMTapEBXS+eaSpcHCKryCwe8FA@mail.gmail.com>


Mike,

I've been wondering about that. Wouldn't your approach tend to
artificially skew genetic distances for haplotypes with nulls towards other
haplotypes that have legitimate (mode minus 1) values? Would it be more
accurate to presume the mode (or implied ancestral value) for nulls, but
add additional Boolean logic columns (i.e. true=1or false=0 for purposes of
variance, GD, ASD, etc.) thus treating them (null425, null439, etc.) as
distinct mutational events?

Just throwing it out there...

Vince T.

On 6 December 2011 05:56, Mike W <> wrote:

> Dennis,
>
> I handle nulls by creating extra columns way to the right for markers like
> 425 and 439 (and 534) and doing if tests to replace nulls with the mode for
> the column minus 1. I then have a macro that copies (& paste special values
> ony) those columns' calculated results back on the the original columns. I
> do this in the modified Generations6-1 from Ken... which is posted at the
> various subclade groups/forums.
>
>
> Mike W
>

This thread: