Except for the software error I mentioned earlier [1]
I've had no substantive problems with version 1.1. I've
only been running v. 1.1 for 2 weeks, but I highly recommend
it.
Under v 1.0 my machine errored out about every 4 runs on
average (usually FM Sheath pump fatal error) but the last 5
or 6 batches I've run have completed without error. All
these were 8 run batches except the most recent which was a
10 run batch. POP5 is giving me longer reads on average and
no more failures than I was getting with POP6 once I
followed Paul Shinn's lead and did a new spatial (I did a
new spectral as well.) I've been using the recommend 3M foil
tape to seal the plates. This has drastically reduced (if
not eliminated) the sequence quality degredation I was
seeing before using fang piercing. Further, the 3700's
loading speed is greatly increased. This combined with
quicker run speeds results in my being able to do about 8
runs/day. Also v 1.1 uses less buffer per run. Much less,
like half. I think I could do a 16 run batch without running
out of buffer--I would run out of water though.
Why the dire reports of v 1.1 from others [2]? This is
speculation, but it may be that the strange bug I document
(to some extent) in [1] is big problem if you don't notice
it and rely on being able to stop runs. (I haven't heard
anything back from PEBio about this bug, so maybe they don't
believe me.) Also, different labs, different conditions.
Here is my advice for anyone wanting to switch to v 1.1:
1) Follow the installation instructions carefully but in
addition do a spatial and spectral afterward. (The spectral
may not be important.) I think just jiggling the caps a
little while replacing the dummy block may be enough to move
some of the plumes off by a pixel or two. (By the way, the
first time I did the spatial it worked--never had that
happen under v 1.0.)
2)Use the thin 3M tape recommended by PEBio. Bite the bullet
and buy a few rolls (whatever the minimum order is from your
3M rep--their phone number is given in the new 3700 manual).
It worked out to about $0.60/ plate for us. Rub the foil on
the plate thoroughly with a paper towel--so you can see the
impression of the wells below.
3)Use the PEBio 384 well plates. They are cheaper than other
plates anyway. (It costs us about $3 per plate--that is less
than $0.01/reaction.) I use fresh plates, not the ones I
thermal cycled in. If you do a plate transfer anyway (Ie, if
you don't just precipitate to clean up dye terms) why not go
into a clean plate that isn't warped? (I don't denature, by
the way) If you can't bear the idea of spending money for an
extra plate, just peel the foil tape off the plate after a
run, wash it (maybe UV irradiate it) and re-use it. My
plates are very flat when clipped in and sitting on the
deck.
If you load the plates with a Hydra or other 96 well
pipetter you can use a method I've presented in this forum
before to set up plate records [3]. This would allow you to
use a 384-well plate to load just 96 samples, for instance.
4)Shut down the machine and computer after each batch and
then restart. While shut down take the sample probe sheath
(the thing that has the fangs) off and clean the fangs of
any residual adhesive.
Good luck.
Phillip San Miguel
Purdue Genomics Core Facility
Notes (these should be links to both bio.nets archives and
dejanews separated by "or"):
[1]
http://bionet.hgmp.mrc.ac.uk/hypermail/autoseq/current/0013.html
or
http://www.deja.com/%5bST_rn=ps%5d/getdoc.xp?AN=582351672&fmt=text
[2] For example
http://bionet.hgmp.mrc.ac.uk/hypermail/autoseq/autoseq.200001/0003.html
or
http://www.deja.com/%5bST_rn=ps%5d/getdoc.xp?AN=570404458&fmt=text
[3]
http://bionet.hgmp.mrc.ac.uk/hypermail/autoseq/autoseq.199912/0021.html
or
http://www.deja.com/%5bST_rn=ps%5d/getdoc.xp?AN=562599774&fmt=text
---
