Plaque PCR mystery

Wen-Yen Kao wykao at
Sat Mar 11 13:26:36 EST 1995

In article <3jk25l$bqt at>, bckraev at wawona wrote:

> Dear netters,
> We use PCR on genomic library plaques to confirm hybridisation positive
> clones at all stages of plaque purification. Although the technique works,
> there is a mysterious phenomenon that we encounter, which is the presence
> of the expected band in all tracks, including various negative controls.
> In the negative control track the band intensity is much much lower, than
> in the true positive tracks and in genomic DNA track. We tried all possible
> remedies, including purchase of a new pipette and using aerosol-resistant
> tips. Those measures reduce the phenomenon, but do not eliminate it completely
> Can anybody explain this? Had anybody a similar experience? ( It does not
> look like a carryover during gel loading, we have checked that )
> Any response welcome
> *************************************************************************
> Alexander Kraev, Ph.D.                 Internet: bckraev at
> Lab. of Biochemistry III               Phone: 0041-1-632-31-47
> Swiss Federal Institute of Technology  FAX:   0041-1-632-12-13
> Universitaetstr.16
> CH-8092 Zurich
> "Some ideas are obscure not because they are complex, but because they 
>  are excluded from our circle of comprehension" - Kozma Prutkov

                              Dear Dr. Kraev,

Though my message may not be very satisfactory for you, I basically can
very much confirm your problem concerning negative control products in
genomic plaque PCR's. All those plaques had already been selected by a 1st
screen of a genomic lambda-ZAP library, cutted with EcoRI. 'Negative
control' for me means: no template DNA provided (water instead). The
single band I see in those neg. controls runs at exactly the size to be
expexted and got weaker, though never disappeared, after replacing dH2O,
10x buffer, dNTP's and primer preparations against new solutions.
Moreover, when I sequenced the corresponding cloned PCR band out of
positive controls, they turned out to be real (i.e. that stretch
contained, based on the previously determined cDNA, the sequence I was
looking for). All those lamda DNA's with apparently the correct inserts
(based on: same size of the PCR product relative to the sequenced ones)
had been converted into bluescript DNA and several been sequenced. So far,
all were false positive (vector sequences instead of the genomic inserts I
am looking for). A colleague of mine kindly carried out further PCR's with
his set of solutions and my specific primers and either my lambda or his
completely non-related template DNA's. Result: Especially with one set of
primers (out of two) ALL templates give one band at the correct size!!
Most probably these primers are contaminated with some sort of that
genomic DNA I am dealing with. I have no clue how it happened but that's
the best answer Ican come up with by now.


                                Thomas Gorr
                                thomas at

|   Wen-Yen Kao, Ph.D.           FAX: (512)471-9651                |
|   Postdoctoral Fellow          Lab: (512)471-7435                |
|   Department of Zoology        Home:(512)478-7147                |
|   University of Texas          Internet: wykao at |
|   Austin, TX 78712                                               |
   Complex problems have simple, easy-to-understand wrong answers.

More information about the Methods mailing list