GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2006-10 > 1159724460


From: Doug McDonald <>
Subject: Re: [DNA] significance of partial str repeats
Date: Sun, 01 Oct 2006 12:41:00 -0500
References: <20061001150620.7967.qmail@web37702.mail.mud.yahoo.com>
In-Reply-To: <20061001150620.7967.qmail@web37702.mail.mud.yahoo.com>


Rick Arnold wrote:
> Playing around with the smgf y database this morning I
> noted about half of the STR fields provide value entry
> options for partial repeat values (e.g., dys 388.....11,
> 12, 12.2, 13, ....). I recall reading somewhere -
> probably on this list - that some (most) labs round
> partial repeats off to the closest whole number of
> repeats. The question is, wouldn't partial repeats be
> just as heritable from generation-to-generation as are
> full sequences of repeats? Thus wouldn't a difference
> between values of 12 and 12.2 on a marker be just as
> significant as a difference between 12 and 13 on the
> marker? If so, the reporting of both 12 and 12.2 as 12
> strikes me as ill advised, to say the least. I hope I'm
> wrong.!!!
>


You are correct, sort of.

The difference between 12 and 13.1 is much, much greater
than between 12 and 13, and in fact the difference between
12 and 12.1 is almost as great as between 12 and 13.1.

This is because partial repeats are in fact insertions or
deletions. They can occur anywhere between the two primers,
and so are much more probable than a single point SNP. If
the amplified length is say 200, and the single base
insertion rate is one in one hundred million, the rate of
the partial repeat occurring would be 200/100,000,000 =
2/1,000,000 = 1/500,000 which is 1000 times rarer than
a typical plain SNP.

Thus these partial repeat numbers are effectively family
SNPs. Of course, they are not truly "SNP" but insertions or
deletions. They are inherited.

I suspect the reason some companies don't report them is
that their technology is not sufficiently precise to
measure them reliably. If it is not, it's better to
not muddy the waters with unreliable numbers.

Doug McDonald

This thread: