PCR primer design (Rules Of Thumb)

Vladimir Svetlov vvsvetlov at utmem1.utmem.edu
Wed Dec 18 00:31:09 EST 1996


In article <596gfi$8ps7 at ncisun1-nf0.ncifcrf.gov>, pnh at ncifcrf.gov (Paul N
Hengen) wrote:

 > I'm thinking of doing a review about the different methods people use for
 > designing PCR primers, and compare this to the selection of primers done by
 > computer programs to see if in real life these actually work. Some people I
 > know use one set of rules of thumb, while others prefer a completely  
different
> set, or even the computer programs.
 > 
 > What is the BEST set of rules?
 > 
 > Does this work every time?
 > 
 > What is the failure rate for those people who
 > design by hand versus those who use the programs?

I guess there some obvious and widely accepted rules: don't have strong
hairpins, that are stable at the intended annealing temp., don't have
extendable primer-dimers etc. The only non-trivial "rule" implemented in
Natl. Bioscience's "Oligo" program is that lowered internal stability
profile to the 3' end of the primer forces annealing of the entire primer
length thus increasing the specificity of the reaction, especially on the
complex templates. Before bying the program we ran a reversed blind test -
in a set of "intuitively" designed PCR primers already used in the labs
around we set out to detect the ones that would work badly based on the
Oligo 4.0 criteria. One pair that was predicted to performed badly
actually worked out OK, but seven pairs that did not work at all were
picked up by the program. Anything that was designed using this program
since worked, and only in my hands it's 100+ primer pairs. It's good also
for picking sequencing primers regardless of the sequencing method
(Sequenase or cycle sequencing). Moral, I guess, is that optimal designs
according to this program so far always worked, although some pairs that
would be rejected also work, but for me too stringent criteria in this
case is not a problem. 
Regards,
V.



More information about the Methods mailing list