We've had an Alf express for over a year now. Over all a good, robust
machine that with some (perhaps extensive) optimization will give 800 bp +
average, >99% accuracy) reads. Its easy to pour gels, easy to load (the
wells are backlit) -- clean up is not too bad. I would warn you to use
thermosequenase, not the pharmacia kits --
but as Pharmacia and Amersham have merged, your service rep. may give you
the same advice.
Throughput is not that high -- 10 samples/gel. But if you really push
it, you can get 2 gels run a day (one per 12 hrs) and still get reads over
900 bases. We've only been able to do this since we switched from Long
Ranger to AMERESCO's PAGE-PLUS acrylamide. [By the way, remember when you
are asking about read length to ask for read length at a certain accuracy.
Alfx and LiCor's can potentially read >1000 bases at >99% accuracy
(ambiguities don't count). On a decent Alfx gel I get one read over 1000
and average in the high 800's.] You could get higher throughput if you
just need short reads -- there are short plates that run very fast -- but
I've never used them.
The software is quite good -- it runs in Windows 95. There are a few
irritating things about it, but nothing worth mentioning.
Okay, now the down side. You can't load the wells with a 8 or 12x
hamilton shringe loader -- at least not one that will pull up from a
standard 96 well plate. This is pretty minor because it doesn't take that
long to load 40 samples (4x10) -- maybe 1/2 hour with a little practice --
maybe faster. The plates are pretty expensive, $900 for the thermoplate,
$150 for the front plate. Don't break them. I worry about the durability
of the theromplate as it is a pretty complex thing. But the Alfx's worst
problem is its "clone 10" problem. The samples are numbered 1acgt,2acgt,
etc.; from the left (close to the laser source) to the right. (I should
add that this is a problem you won't even notice unless you get everything
optimized so you reads are averaging over 800 bases and are pretty even.)
Each of your samples (acgt) is called a clone. The first 8 clones are
capaple of delivering reads over 900 at >99% accuracy routinely. I don't
remember seeing clone 9 break 900 bases and clone 10 never (in any of the
120+ gels I've run) breaks 800 bases. Again, unless your willing to put a
lot of effort into optimising your sequencing you may never even notice
this effect.
Finally, I highly recommend Pharmacia for their service and sales
representatives -- at least the ones we deal with, Marc Goldstein and John
Trumbull.
Phillip SanMiguel
Bennetzen lab
Purdue University
Paul G. Rothberg wrote:
> We are considering Visible Genetics vs. Pharmacia automated sequencers.
> Input would be greatly appreciated from users. Send to
>JBRADLEY at CCTR.UMKC.EDU
