On Mar 18, 9:19am, "Kevin O'Donnell" wrote:
> Subject: Re:
> <<Your message>>
>> . We've tried laborious/exhaustive DNA extractions, carefully
> > monitoring DNA concentration, ensuring the integrity etc. of our reagen=
> t
> >stocks, but nothing seems to get over the reproducibility problem. =
> If there is
> >anyone out there who has experienced similar problems
> > we would like to hear from you, believe me!
>> We have just been through exactly this problem! A promising assay just =
> wasn't
> giving reproducible results - it was driving us mad. Eventually, we trac=
> ed it to
> the *water*. The water used was filtered and autoclaved but when run =
> side by
> side with reactions using water from a different source it was clear that=
> it was
> indeed the culprit. Why? We don't know. It might be fluctuations in =
> the pH or
> ionic concentration of the water source.
>
What type of PCR machine are you using? There have been reports of
reproducibility difficulties between various wells of the machine. I know
this was documented in the original Perkin-Elmer machine (there was an
article in Biotechniques about this). This was demonstrated by preparing a
master mix (4.8mls) and running 48 identical reactions in all the wells of
the machine. The variation from well to well was quite impressive.
With regards to the water, I would a little skeptical of pH or ionic solutes
if you are using water prepared from a source that bases purity on electrical
resistance of the water. We use a Millipore system that measures the ohms
(pure water is something like >18Mohm). All we do after that is purify the
water through a 0.22 micron filter and store in aliquots. One point to
consider is that these systems don't measure organics (unless they carry a
charge). The final filtration is simply to ensure elimination of bacterial
contamination.
> We've decided to splash out (sorry) on some bottled water in future.
>> This may well not be what's causing your problems of course. For example=
> :
>> >Our latest ploy is to hang rosary beeds on the
> > PCR machine and play Elvis Costello very loud, but this has met with =
> only
> > limited success.
>> Your problem may well be the variation in quality of Elvis Costello album=
> s.
> Have you been playing 'Goodbye Cruel World' on the days it hasn't worked =
>> and 'Imperial Bedroom' on the days that it has?
>
I've been told that a horseshoe over the machine works even you don't believe
that it brings good luck.
Mike Kurilla
