A number of cents from my end (see purdue CD2 for further information)
Key problem in non-ratiometric counting methods are system dead times and flow
inconsistency. Make sure your reference particles are not out of scale in any
signal as that can fool your electronics.
Absolute counts for beads can be calculated accurately from % solids and
diameter and polymer density
(0.5u latex at 2.5% solids come at 3.64x10^11/ml)
Absolute counts with the Coulter XL work very well as long as the acquisition
rate is below 1300 events per second. When counting 100ul (five 20ul cavities) A
display of time versus count tells you about clogging and excess flow rate (>300
on 300 second timescale). Unfortunately versionII software cuts of too early, so
only the data for 80ul can be used.
Counting methods
- Fixed volume counting or volume integration :
This method is used in most haematology analysers. The volume measurement is
achieved by two electrodes acting as level sensors in a known geometric set-up
(Partec CAII) or by loading a cavity in a ceramic valve (Coulter XL) or a loop
made of tubing as done in high pressure liquid chromatography instruments. In
all cases all events within the volume are measured.
- Time integration :
This approach is based on the assumption of a constant volume flow over time. It
is best achieved by (syringe) pumps delivering the sample. Such systems are
implemented in the Ortho Cytron absolute and the former Skatron Argus flow
cytometer now Bio-Rad Bryte HS.
- Spiking with reference particles (ratiometric counting) :
This method can be used with all flow cytometers. It corrects for system dead
times and the variations in flow rates that can occur in instruments that use
differential pressure to deliver their samples. An example of bacterial counting
using a bead standard is given in Figure 5. The tight cluster of beads also
serves as an on line alignment control, particularly important when measuring
environmental samples that are more likely to block the flow path.
Cell disaggregation
Single cell suspensions are essential for any form of accurate counting.
Aggregates only give rise to one single colony or event. If for example one cell
in a triplet is positive for a dead cell marker the whole aggregate is
registered dead but will grow when sorted onto agar plates. Skin sample are far
worth as hundreds of bacteria can be attached to a single squame.
Cells can be dissaggregated by either chemical or mechanical methods.
Mechanical methods have a broader application spectrum but can lead to problems
with filamentous organisms. Shearing by needles leads to problems with clogging
and is very tedious. Shearing with homogenisers is difficult with small volumes
and causes problems with foaming and sample carryover.
Ultrasonic treatment is the most convenient method, but it is important
to apply reproducible energy levels. The geometry of the set-up and the material
of the sample container has to be taken into consideration. When using a probe,
energy loss can occur by coupling to ice cold water surrounding the sample
container or by airbubbles trapped at the bottom of pointed vessels.
Transmittable energy in an ultrasonic waterbath is sensitive to the level of
water, its temperature and dissolved gas as well as its cleanliness. Soft
container materials like polypropylene do absorb the energy in both systems.
some points of discussion
The high number of counts achieved by flow cytometry make the technique
superior to others with the regard of counting accuracy. By careful pipetting
technique and using 0.05% tween 20 to avoid cell sticking we could achieve
counting variations within 1% of the expected numerical counting error.
The detection sensitivity of optical systems is limited by the
statistical abundance of an event and the signal intensity separating the event
from background noise. To identify a cell cluster it is desirable to have at
least 100 cells in it. Thus if there is one organism per µl in the final sample
volume, 100µl have to be measured. Relative frequency is another limit to the
measurement. To detect a log 3 reduction equivalent to an event frequency of
0.1% 100.000 events need to be screened to see 100 wanted cells. Good signal to
noise ratio is therefore important, as, at lower relative frequency, the
labelled cells are end up within the standard deviation of the unlabelled
events.
Signal to noise ratio is also limiting the speed of measurement, as with
increased sample / volume throughput the variations broaden but in particular
the background fluorescene increases due to free fluorochrome. We currently run
flow rates below 10µl/min which leaves us with a practical sensitivity of
approximately 10^3 within 10 minutes. Lower concentrations require patience or
pre-enrichment by physical or biological means.
Key problem in other counting methods are system dead times and flow
inconsisteny. Make sure your reference particles are not out of scale in any
signal as that can fool your electronics.
Gerhard Nebe-v.Caron
Unilever Research, Colworth,
Sharnbrook, Bedfordshire
GB - MK44 1LQ
Tel: +44(0)1234-222066
FAX: +44(0)1234-222344
gerhard.nebe-von-caron@unilever.com
______________________________ Reply Separator _________________________________
Subject: absolute counts
Author: novod@muss.cis.mcmaster.ca at INTERNET
Date: 26/02/97 22:37
Hello Everyone,
I am trying to do absolute counts of bacteria. I have been using a
kit but its about $100 for 100 samples. My problem is that the kit keeps
dissapearing faster than the bottle of vodka I keep on the lab shelf for
days that experiments don't work. I am wondering if anyone has a
cheaper/longer lasting suggestion for absolute counting of cells on the
FACScan..
Thanks for your help.
I will compile the suggestions and post them for future reference.
Dave
***************************************************************************
I have nothing to put here that anyone would care about. But since
everyone has one of these, I figured I should too . . . . .
***************************************************************************
This archive was generated by hypermail 2b29 : Wed Apr 03 2002 - 11:49:28 EST