Wolfgang Schechinger hubahopp at
Fri Jul 30 08:11:28 EST 2004

Dear Josué,

the (not-)existence of this problem depends on what you want to amplify. As
long as you "only" do PCR on plasmids or isolated pieces of DNA with them,
you don't have to care much, as there usually is no wrong sequence to bind
to and you have lots of template present. 
For genomic DNA however, it might be a good advice to screen the primers
against your target organism's database for mispriming possibilities. If
you have some idea about your mRNA's / ssDNA's tertiary structure, it might
be a good advice to avoid primers binding to stable (=GC rich) or long
stems, at least this point is worth consideration in troublesooting not
working reactions.

Overlaps of 3, 4 or 5 bases usually are not critical at all, because they
easily melt, as long as they won't be extended by the polymerase, so try to
avoid them at the end of primers if possible. At least, try to make them
being rich in A/T for lower binding energy.

In my experince, it is important that primers don't tend to form extendible
oligomers as this will eat up your primers and nucleotides and thus kill
your PCR with thermodynamics (and likely produce wrong positives). Then, if
possible, design your primers in a way that they have similar theoretical
melting temperatures (means they behave identical in annealing) and
finally, run a gradient PCR to find the optimal annealing temperature
(highly recommended, saves LOTS of time and money in the end).

Don't rely blindly on the suggestions of a computer program. Usually, it's
enough to do some "eyeballing" and then check the primers for crude
incompatibilies (hp, dimers) with a primer analysis program.



At 13:40 29.07.2004 -0000, jmfopouk at wrote:
>Hi, am getting confused about avoiding primer-dimer formation during
primer design. The 
>software (Primer Express) displays primers' secondary structures  and the
primer-dimer that 
>might be able to form. I wonder wether the best primers combination is the
one showing short or 
>long overhangs!!!
>Thanks for helping.
Wolfgang Schechinger


More information about the Methods mailing list