GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2006-11 > 1163176241


From: "Alfred A. Aburto Jr." <>
Subject: Re: [DNA] Conference [recLOH]
Date: Fri, 10 Nov 2006 08:30:41 -0800
References: <200611091650.kA9GonYe029348@mail.rootsweb.com>
In-Reply-To: <200611091650.kA9GonYe029348@mail.rootsweb.com>


> Thomas Krahn wrote:

[snip,snip]

>
>12th Query
>How many haplotypes have at least two identical values at DYS464 and at
>the same time identical values at CDY and at DYS459?
>COUNT * FROM HAPLOTYPES WHERE ((DYS464a LIKE DYS464b OR DYS464a LIKE
>DYS464c OR DYS464a LIKE DYS464d OR DYS464b LIKE DYS464c OR DYS464b LIKE
>DYS464d OR DYS464c LIKE DYS464d) AND (CDYa like CDYb) AND (DYS459a LIKE
>DYS459b)) 995 5.29%
>
>So more than 5% of all haplotypes have a triple doublett or a deletion
>on P1/P2.
>This seems to me an impressive number. If we had a bare random
>distribution, those haplotypes should occur in less than 1% of the
>population.
>
>

From the J database I have collected, assuming the same criterion as
your 12'th Query, I get 19 cases out of 339 giving 5.6%. I'm happy to
see that the results are very close to the same!

What you say about the probability of your 12'th Query assuming a random
distribution is good. I tested it a little bit starting from the modal
haplotypes for J2, using random steps of -1,0, or 1, and only got 1 case
out of 800 (0.13%) trials where DYS459, DYS464, and CDY each had 2
indentical markers. There is no way I think to get anything above that
percentage (0.2% say) without some coherent process (recLOH) occuring.
But I didn't test cases where the steps would be more than 1 with
different probability.
Al

>Thomas
>
>

This thread: