Real-time PCR quantitation

Nicolas von Ahsen nvahsen at hotmail.com
Sun Mar 21 06:20:36 EST 1999


Dear David,

I have to comment on Januarys opinion on the LightCycler (LC, Roche)
and add some of our experience with this machine.
I haven't tried out one of the ABI machines. The LC is great for
(point-)mutation detection, which really gives him a position in
routine diagnostics. What is annoying is the price of the ready-made
LC-PCR mixes, you better make your own. This holds true for both, the
SYBR-green and the hybridisation probe format. For the latter the
labelled oligos are also expensive. But you can do thousands of
genotypes with one synthesis and need no restriction enzymes. The
original LC (IdahoTech) was a very open system, their web page still
is a great source of hints towards rapid cycling. The Roche policy is
probably another way...
But you asked for quantitative PCR: Well designed PCRs are easily
ported into the LC format. Rapid cycling is not very sensitive towards
annealing temperature, so PCRs usually run after one or two tries. One
should use short amplicons (100-300 bp) so that only 5 sec. at 72C are
needed. This way a new quantitative PCR can be started every 30 min or
so, which greatly improved our throughput. Runs are very reproducible,
it is amazing how exact PCR is. The CV from run to run is 10-20%. As
data are collected online every cycle it is always possible to detect
the log linear phase and use it's slope for quantification.
The software is still not perfect, but I don't consider it to be
"beta". No chrashes at all, runs very stable. There are still many
ends where the software needs minor improvements (especially post-run
modifications are not possible inside the software), a new release was
already scheduled for last year and should be out now by April - let's
wait and see. The software puts all data into ASCII files, so editing
these files is easy and they can be imported into e.g. MS-Excel.
And I don't know about other quantification systems except for the
Cobas Amplicor, and that's not what you want. I don't think there are
some.

Best regards
Nicolas




More information about the Methods mailing list