PCR machine recomendations PLEASE :-)

Leman leman at leman.org
Thu Oct 21 19:10:48 EST 1999

On Thu, 21 Oct 1999 17:13:12 +0100, Robert Hartley
<rh at mblab.gla.ac.uk> wrote:

>I am buying a PCR machine and have about 5-6K to spend.
>I would like comments from people who have used/compared the MJR PTC200
>gradient the PE 9700 and stratagene Robocycler. which one did you prefer
>in side by side comparison and why? 
>Are there any other, simple to regularly use, machines that are worth

Do you really need a cycler with T gradient capabilities?

We bought MJR PTC-100 machines (no gradient) for ~$4,500 each. For
optimization purposes we used an Eppendorf Mastercycler that belongs
to a guy across the hall. Just set the gradient to be +/- 5.5C (from
the middle of the block) and you'll get 1C increment per well. Soon
enough we realized that what seems like optimal conditions on a
Mastercycler is not always optimal on MJR PTC-100. Reluctant to spend
another few thousands on a gradient machine, I started running
numerous "tweak-and-see-what-happens" reactions on our MJR. After
awhile, I found out about the wonders of touch-down protocols. The
protocol I started with looked like this:

12'00" @ 95C 		'to activate PE AmpliTaq Gold
Tdenat=94C 		'all denaturation steps are @ 94C
Telong=72C 		'all elongation steps are @ 72C
Tann=66C 		'initial annealing T is 66C
FOR n=1 TO 16		'16-cycle touch-down
  0'30" @ Tdenat
  0'30" @ Tann
  1'00" @ Telong
  Tann=Tann-1C 	'next cycle the annealing will be 1C lower
			'at this point Tann=66-16*1=50C
FOR n=1 TO 25		'25 more cycles
  0'30" @ Tdenat
  0'30" @ Tann		'annealing @ 50C every cycle
  1'00" @ Telong

24h00'00" @ 4C	'if you run it overnight

The idea behind the touch-down approach is that at high annealing
temperatures mispriming is very unlikely. Therefore, if you synthesize
any product at all at this stage, it will be only highly specific
fragment. As you decrease the annealing temperature, the probability
of primers' binding to the template increases, but so does the
probability of mispriming. However, at any given temperature the
probability of correct priming is always higher than that of
mispriming, especially if this temperature is unusually high. In other
words, for a few cycles you will create almost exclusively the desired
product. By the time the annealing temperature is low enough to start
worrying about mispriming, the template will be those highly specific
amplification products and the original "seed" template DNA. Of
course, due to the exponential nature of PCR, the former will greatly
outnumber the latter. Therefore, even if the annealing temperature at
this stage is well below what it should be, your primers are much more
likely to anneal where you want them to. By keeping the temperature
lower than usual at this stage you increase the efficiency of
annealing, thus increasing the yield.

You start at 66C, because if you get non-specific products at this
high annealing temperature you should throw your primers away. You end
at 50C, because if your primers don't bind well at a temperature this
low they should go into the garbage as well. You don't want to step
down in big T steps, because then you increase the chances of
mispriming too quickly. You don't want to do it in small steps,
because then most of your Taq will burn out before you complete the
touch-down and get to the "real business". (16cycles)*(-1C) should do
the trick.
To my disappointment, I found out that this protocol doesn't always
work well (although it still performs much better than any standard
protocol). Then, I increased the duration of denaturation and
annealing to 0'45" in both touch-down and "regular" parts of the
program. The results improved. The real breakthrough happend when I
started adding DMSO to my reaction mix. Anywhere between 0.1 and 3%
(final) helps, although I found 1% to give best results in most
"difficult" cases. Now, if I don't see good amplification with this
protocol and 1% DMSO in every reaction, I either throw away the
primers or start worrying about integrity of the template. Both really
pay off, especially since we sequence those PCR fragments, and each
sequnecing reaction costs about $6. If you get a garbage sequence in 6
samples, you already wasted as much money as you would spend on a
better primer pair. A good example of when optimization would not be a
smart thing to do.
Just some things to consider...
Good luck,


More information about the Methods mailing list