GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2007-05 > 1178895606


From: James Heald <>
Subject: Re: [DNA] Different Subject
Date: Fri, 11 May 2007 16:00:06 +0100
References: <449241.34619.qm@web81115.mail.mud.yahoo.com>
In-Reply-To: <449241.34619.qm@web81115.mail.mud.yahoo.com>


Gary Felix wrote:

> Dupuy's paper's not only found that mutation rates increases with allele length. This paper also found that:
>
> "there was a significant relative excess of losses in long alleles and gains in short alleles (p=0.043)."
>
> This means there is an added mechanism pushing mutations in the ancestral direction. Therefore if a mutation is going to occur at DYS 385b, which has a value of 16 then if this is from R1b it would tend to mutate to 15, if Q it would tend to mutate to 17. This occurs because the mutation would expect to be in the direction of the ancestral value. Without this second mechanism YSTR's in general would have the net effect of increasing for a sloped distribution at any dys.
>
> If only allele lengths were considered this effect would not be noticed. Ancestral values must be considered as well.
>
> Incidentally there are only Two Father Son pairs in the Mexico Project. One pair is one step at 389-2 at 12 markers and the other is one step at 464d at 37 markers (KSHRJ and NTUMF in ysearch). Humm.....both multi copy.


Um, I very strongly suspect -- and I would certainly urge as an
appropriately simple initial hypothesis -- that there is no additional
tendency in any "ancestral" direction, beyond any tendency caused by the
absolute length of the allele.

If it is correct that DYS 385b in R1b is slightly more likely to mutate
16 -> 15 than 16 -> 17, I would expect it to be similarly more likely to
mutate 16 -> 15 in haplogroup Q.

The modal haplotype of haplogroup Q is the result of a random lottery of
which lines happen to die out or daughter out. That's not indicative of
any unusual stability of the modal haplotype, just that this founding
line happened to be the one that survived, against a background of
almost all lines dying out. The ancestors of this particular line
happened not to have all daughters.

The survival is unusual. In a population not undergoing net growth, the
tendency is for almost all lines to tend to die out or daughter out,
with probability 1.

See the graph at
http://en.wikipedia.org/wiki/Galton-Watson_process

The haplogroups that exist today are the very few examples of
patrilineal lines which have not (yet) become extinct.

All best,

James.



This thread: