GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2005-10 > 1128723709


From: "John McEwan" <>
Subject: PCR primer design: was RE: [DNA] S21+ extends from 37 markers to 48 markers.
Date: Sat, 8 Oct 2005 11:21:49 +1300
In-Reply-To: <000f01c5cb58$6efecc80$71509045@Ken1>


Dear Ken

You asked
.... Question: can I expect the different labs to generally have made
their own version of the primer(s) which will differ from the primer(s)
of other labs? Or is the use of different primers the exception rather
than the rule? And is there a similar issue concerning the primers which
isolate a region where SNPs are located? Will one lab usually be using
different primers than another?...

I will take a stab at answering this question in some more detail than
Gareth and David (as usual my slow considered reply meant they have
answered it while I was composing this). Sorry for the detail and length
but this question has appeared in various guises many times previously
and I think a fuller explanation will help many especially regards the
new 464x tests.

Normally, when a STR (often called a simple sequence repeat e.g. has a
sequence like AGTAGTAGTAGT) is found and reported the primers (short
18-25bp oligonucleotides flanking the variation) used for the PCR are
reported as well. Many primers to detect an STR are often possible, but
by the time a whole host of conditions are satisfied very few have all
the properties required so most people stick with what is reported. If
any body is interested you can take the STR sequence below:
http://www.ncbi.nlm.nih.gov/entrez/viewer.fcgi?db=nucleotide&val=1223298

This variant has the primers used to amplify it reported as well. You
can take the sequence in the "FASTA" format shown below.

>BMS332
ATCTGAACTAAAGGACAAACCTATCTTCTGACAAAACCCTTTTAGCACAGGACTCTGTGTGTGTGTGTGT
GTGTGTGTGTGTGTGTGTGTGTGTGTGTGTGTAGGTATGTGTTTTAGGCACTGGTGGGAATAAAGACACT
CTGATTTAGCTGAGAACTTTCCATGCAATTATTTT

And paste it into a primer design program slot

http://frodo.wi.mit.edu/cgi-bin/primer3/primer3_www.cgi

and push the button and it will come up with the "best" primers and note
these differ from the ones provided. Fiddling with the numerous
conditions results in different primers.

Sorry for the bovine example but it reflects my background. Moooving
right along :-)

When multiplexes (combinations of STRs amplified in one PCR reaction)
are used additional complexities are created and often to satisfy
product size and interference between primer constraints different
primer sets are used than those originally reported. However, if labs
use the same "official" multiplexes then they use the same primers. Some
labs of course spend a lot of time designing their own multiplexes (to
reduce cost, improve success rates and accuracy of scoring) and these
are often not reported and yes they will differ between labs.

For SNPs the question is even more moot because there are NUMEROUS
technologies used. These range from sequencing, primer extension,
hybridisation, mismatch repair... In general the "primers" used differ
between all of them and when multiplexed the same issues described above
are present.

Now we get to the core of your question. What happens if there is a
mutation and a base (or two) differs where the primer is meant to bind?
Well the answer is it depends on a number of factors. For PCR reactions
used for STRs normally the last 5 or so bases at the 3' end (this is the
end where extension takes place and by convention is the right hand side
of the sequence) needs to have a perfect match. If it does not have a
perfect match the reaction fails (which often happens for other reasons
as well) but if it is consistent this is often called a "null" allele
and a SNP or indel under where the primer binds is suspected. However,
the PCR reaction often still works to SOME extent if there is a mismatch
elsewhere in the primer sequence. The exact results depend on the exact
PCR conditions used, which almost certainly do differ between labs (in
fact often you get differences between PCR machines from different
manufacturers). As a rule of thumb if there are more than 2 mismatches
in total across both primers the reaction does not work under any
"normal" conditions.

This property can be used by positioning a set of primers so the 3' end
of a primer is on the variant in an attempt to detect it more clearly.

Now I do not know what the various labs do, I suspect there are a lot of
differences, but this does mean that if a SNP is present in an
individual where the "probing" sequence normally binds, the results you
get will depend on the exact methods and "primers" used and this may
lead to lab to lab differences. However in the vast bulk of cases (>99%)
this problem is not present, otherwise it would have been recognised and
described.

Hope this alternative explanation clarifies some aspects.

Cheers

John McEwan




This thread: