GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2007-02 > 1172015501


From: Doug McDonald <>
Subject: Re: [DNA] Average and extreme genetic distances in R1b
Date: Tue, 20 Feb 2007 17:51:59 -0600
References: <847340.89922.qm@web31508.mail.mud.yahoo.com>
In-Reply-To: <847340.89922.qm@web31508.mail.mud.yahoo.com>


Jim T wrote:
> From: "Lawrence Mayka" <>
>>> The two people with the greatest GD from each other were
>>> Malugani-QG3AJ and Hart-7FZUQ, with a GD of 76 on 67 markers
>>> (71 on 65 markers, excluding CDY a & b).
>> ..
>>> Caveat: I used a simple stepwise GD calculation that
>>> sometimes gives different GDs than Ysearch's GD calculation.
>> Ysearch considers this a distance of only 56. 7FZUQ seems to
>> have a Null value at DYS406S1, where QG3AJ has a 10. Are you
>> treating that as a genetic difference of 10? Ysearch ignores
>> it entirely, which of course is not correct either.
>
> When comparing a null marker with a marker that has a value, I
> count it as a genetic distance of 1, which I think is better
> than what Ysearch does.
>

I think the correct method is to consider it a difference of one,
BUT to increase the total marker count by one and use
mu0*n/(n+1) as the mutation rate.

For example, two 67 marker haplotypes which differ only in having
one with a null 425 would be counted as 67/68. If the average
mutation rate for 67 markers was .002, I would use for the
67/68 result a rate of .002 * 67/68.

The reason is that a new process was added to the mix,
essentially an SNP, and the rate of it was small enough to
be negligible, more or less.

Doug McDonald

This thread: