Apologies if 2 of these come through - a few problems with email my end, I
think.
Andrew - I stay away from using anything but de-ionised formamide or water
to resuspend your samples, I had plenty of problems with this when we first
had the instrument. I've found no difference between formamide and water,
except we can leave the plates open with formamide (no need for foil).
Could it be the capillary array itself is degenerating? Do you ever get
good runs, or are they always fairly poor now?
Have you tried any of the G50 clean up plate's for your samples? Maybe some
salt is coming through with your precipitations and is interfering with the
injection. I think the 'rubbish in, rubbish out' saying is even more true
on the 3700 than 377/373 systems (no slur on your samples intended, but we
have seen split samples fail on the 3700 and work on a 377 when using an
ethanol pptn). We can get away with a lot more on our gels than the 3700.
As for the loading/running conditions, we use
run temp: 50C
cuv temp: 50C
sample vol 25 x0.1ul
inj v. 1000
inj time 50sec
In essence only changing the cuvette temp by 10C. This is running POP6 by
the way (we run both genescan and sequencing services - I'd like to use POP5
for both, but no joy yet). It made a bigger difference to our genescan
samples. We now get a very even signal strength across the entire capillary
array whereas previously it fell off a lot on the LHS. I'm happy to say
this change has improved both sequencing and genescan results.
As for the template amounts & sample overloading, this is what we are
working on, we routinely use 0.5 BigDye (v2.0) reactions, even for the
cosmids. Even when we quantitate as accurately as we can sometimes the
samples overload sometimes they do not, the problem we face is if we only
use half of the cleaned up product (instead of diluting after 1 run) we
occasionally fall into a low signal trap...which is much more expensive to
repeat the entire reaction rather than dilute and run again....
Oh, before I go, what is the current state of sequencing with the 3700? Are
users happy? I am getting very good reads at the moment (750+ bases for
pgem standards at less than 1% error), but I have spoken to people seeing
much less than this. Am I likely to see a difference as the array gets much
older, or when I eventually swap arrays can I still expect consistent
results with what I get now (my array is about 100 runs old)? I'm happy to
recommend the product we use to clean up our reactions, but to save any
advertising contact me in person and I'll let you know who the supplier is
(it's a G50 spin plate).
Cheers,
Stuart.
PS What are feelings about Snapshot kits on the 3700?
Stuart Bayliss,
Manager,
Genetics Core Facility,
MRC Clinical Science Centre,
Hammersmith Hospital,
Du Cane Road,
London.
W12 0NN
Tel (lab): 020 8383 3181
Tel (office): 020 8383 8305
Fax: 020 8383 8338
email: stuart.bayliss@csc.mrc.ac.uk
web: http://146.179.68.156
-----Original Message-----
From: Cairo, Al [mailto:Al.Cairo@pfizer.com]
Sent: Thursday, August 24, 2000 10:33 PM
To: Recipients of ABRF List
Subject: RE: Injection on the 3700
Hello Stuart and Andrew,
I agree with Stuart, the problem may lie with Andrew's templates. How does
the control template run? Have you tried looking on agarose gel to get a
feel for the quality and quantity?
Stuart mentioned optimizing front end cleanup and cuvette temp, resulting in
1-2% failure rate. For plasmid sequencing we see about the same using
Sephadex to cleanup our BD CS reactions. I am curious to know the run module
conditions you find optimal. Some of our current conditions using POP6 are:
run temp: 50C
cuv temp: 40C
sample vol 25 x0.1ul
inj v. 1000
inj time 50sec
To avoid overloading (which I assume means too much CS reaction product)why
not tinker with the CS reaction. You might quantitate your templates and
dilute them appropriately prior to CS reaction, or try running a 0.5x or
0.25x CS reaction.
Good sequencing,
Al
This archive was generated by hypermail 2b29 : Fri Sep 01 2000 - 16:10:38 EDT