Archiver > GENEALOGY-DNA > 2003-11 > 1068478666

From: "Alastair Greenshields" <>
Subject: Re: [DNA] Measurement of fragment lengths
Date: Mon, 10 Nov 2003 15:37:46 -0000

Malcolm wrote:
"the group administrator would be unaware of this information as it would
probably be reported as
Male A 11.46 = 11.5 = 12
Male B 11.44 = 11.4 = 11
a one-step difference."

Nancy wrote
"I think it is rare to have a non-integer NOVEL MULTIPLEX FOR
instances out of 1036 markers from 74 individuals....It could be the result
of a variant allele."

The paper that Nancy mentioned compared the results of the 20-plex with a
commercially available 6-plex. It was the 6-plex that picked up the
variants which were missed by the 20-plex. This could be because many of
the primers were slightly different, but also because the 6-plex was
compared against allelic ladders that are supplied with the kit, whereas the
20-plex alleles were numbered using software that places the results by
comparing them against internal size standards.

One cannot always pick up every non-integer value every time, especially if
you rely on analysing each marker just once. However, the occurrence of
these non-integer values is fairly low.

Malcolm picked up a point about 'calling' an allele (i.e. deciding which
repeat number to give a particular marker). If a value is in doubt when it
comes out of the machine, then it is simple re-run and double-checked.
Being able to compare two cousins and double-check the results that way is
not an option, as the lab runs blind on each sample (i.e. only a code number
is seen by the lab which doesn't have the surname on it).

A good strategy is to use multiplexes that overlap each other. This
effectively means that markers are tested more than once AND using different
primers. Additionally, sequenced allelic ladders for each of the markers
should be constructed also and run as quality controls in conjunction with
an internal sizing standard. Neither methods are cheap for the lab to do,
but all are adopted by us at DNA Heritage.

Hope this helps.

Kind Regards,

Alastair Greenshields
DNA Heritage

This thread: