review: 96-well autosampler from Cytek

From: Laird Bloom (LBloom@Phylos.com)
Date: Mon Jun 11 2001 - 14:32:35 EST


	This is a follow-up to a query I posted several months ago about the
96-well plate automatic loader for the FACSCalibur available from Cytek
(Cytek Development, contact Eric Chase	510-657-0102.  Website
http://www.cytekdev.com/    ).	Many people were interested in the unit, but
nobody reported any direct experience with it.	We have since had the
opportunity to use the Automated Microsampler (AMS) for several months (via
Ray Lannigan at Tritech Field Engineering, 800-886-7004),  and it seems very
good overall.  Below is a summary of our experience with it.

	The sampler consists of a robotic plate carrier that fits under the
cytometer's sample tube, a control unit that takes about 1 foot of bench
width to the right of the cytometer, and a set of wash and waste bottles
that fit behind the controller.  There are two cables that connect to the
cytometer's internal circuitry and a fluidics/pneumatics bundle joining the
plate carrier with the control unit.
	The unit works by positioning the plate below a probe that is
connected by a short (3-inch) piece of tubing to the cytometer's sample
uptake tube, then lowering the probe into the well and creating a seal
around the edge of the well, stirring the contents of the well with a small
paddle attached to the bottom of the sample probe, and applying air pressure
to the well to drive the sample up the tube.   After the sample is taken,
the plate carrier moves a washing station under the probe, and the probe is
rinsed and backflushed with water from the AMS wash bottle.  The user can
vary the duration of the rinse and backflush to minimize carryover from
sample to sample (minimal in our hands, less than 1 percent with 2-3 seconds
for each step).   The system requires about 3 sec. of boost time to get the
sample to the cytometer and another 3 seconds or so to synchronize the AMS
unit with CellQuest (more below).  Altogether,	it takes about 15-20 seconds
per sample in addition to the actual time of sampling and counting, and a
96-well plate takes about 40 minutes of hands-free operation to process if
the cell density is reasonable (100,000 cells in a volume of 100 ul).
	Once we learned how to set it up properly,  we were very satisfied
with its performance, and we have been able to run through 200+ samples with
a minimum of attention.  Operating the AMS takes a bit of getting used to.
The interface with the computer is somewhat awkward because the software on
the AMS doesn't actually communicate with CellQuest.  Instead, the user
specifies a desired number of events to be saved and a maximum acquisition
time, and the AMS takes over control of the mouse and literally moves the
cursor over the "acquire" and "save" buttons in CellQuest when these values
are met.   The AMS counts threshold events that occur after a preset time,
and it does not begin counting exactly when CellQuest does, so the number of
events saved will vary slightly from file to file and will be close to but
not exactly equal to the number of events set on the AMS. With a little
practice, it's not hard to keep CellQuest and the AMS synchronized, but it's
important to keep CellQuest from terminating the acquisition before the AMS
is ready.  The main limitation of this setup, aside from the unequal event
numbers in each file,  is that there is no way to set the number of events
based on any parameters that come from CellQuest.  In other words, it's
impossible to acquire a set number of events that occur within a specific
gate.  For many applications, this will not be a serious limitation (just
acquire more events, up to 60,000 maximum, and gate during analysis in
CellQuest), but for analyzing cells representing less than about 20% of the
population, this unit will not work well.
	We have noticed a few  quirks.	(1) In a few wells, the unit has
kept sampling cells beyond the specified cutoff number (e.g., 35,000 when we
asked for 5000), and despite consultation with Cytek, we have not been able
to solve this problem.	This results in  strange-looking histograms but does
not cause data loss.  (2)  Occasionally, only 94 or 95 data files are saved
for a 96-well plate, or an empty data file was saved.  This seems to be
related to a failure of the probe to make a good seal around the well (often
at the edge of a plate) and may be specific to the unit we've tested (one
solution:  put standard beads in every row or two to show where the well was
dropped).    (3)  The two units we've tested out have had problems with
overheating, which apparently have been fixed by  several new design
features.
	We have found Cytek very helpful and responsive to our questions,
and willing to redesign parts when problems have arisen.  Although this
product is quite new, it seems to be a good addition to a lab needing to
handle hundreds of samples at a time.  The time saved by automatic loading
is tremendous.	The price, around $15K, is substantially better than the
$35K unit that BD will be selling.


Laird Bloom
Phylos, Inc.
128 Spring Street
Lexington, MA 02421
lbloom@phylos.com



This archive was generated by hypermail 2b29 : Wed Apr 03 2002 - 11:57:43 EST