Re: Events below threshold
Leary, James (jleary@utmb.edu)
Fri, 05 Sep 1997 17:12:00 -0500
I agree with Dave Parks that new designs of commercial flow
cytometers do not always represent progress, e.g. on the
multi-thresholding issue. I'm pretty busy and don't comment often on this
usenet, but this an area that has been at the heart of much of my
technology development work in flow cytometry over the past 16 years.
It's obvious from comments on the usenet that many people are unaware of
work that has been done in this area. Multi-thresholding is a very
important limitation of most present day instruments that I have been
pointing out periodically over the past 16 years without much interest in
the general flow community, particularly among the commercial instrument
manufacturers.
We have been doing multiparameter thresholding for over 10 years in
our home-built high-speed flow cytometer/cell sorter system which among
other things serves as a multiparameter thresholding of events which may
be below the normal data acquisition thresholds. This is the basis of
how we can accurately count all cells at rates in excess of 100,000
cells/sec while determining with the main data acquisition system
threshold which events we wish to store as listmode data. The system
works well and has an advantage over trying to trigger on a single rare
fluorescent signal to cut down on the number of events stored. We use a
boolean logic combining thresholds on three (typically fluorescent)
parameters. We use a double-thresholding logic on a light scatter signal
(since that is the only thing that all cells must have) to determine
whether something is likely to be a cell at all followed by a secondary
threshold on the same parameter. Then we look for how the signal crosses
these two thresholds to detect doublets or very closely coincident
events. Signals which satisfy all of these multi-threshold requirements
then serve as the system trigger for the main data acquisition system.
Using this system we have processed 100 million cells in about 15
minutes with a results of only storing listmode data on the rare cells of
interest. We are in the process of raising this speed limit to more than
200,000 cells/sec to perform some applications in molecular toxicology
and combinatorial library screening for which we are also developing new
single cell PCR and single-cell DNA sequencing (using TA cloning
methods).
For an easier reading version of the above and the general problems
in multi-thresholding you might find useful my chapter on "Strategies for
Rare Cell Detection and Isolation" in Methods in Cell Biology vol. 42,
1994. For a more detailed and technical description you might want to
read several manuscripts we have published in the SPIE journals (Vol.
2386; pp. 150-163; Vol. 2678: 240-253; Vol. 2982: pp. 342-353). For the
braver at heart you might want to explore our basic patent on
multithresholding methods (U.S. Patent 5,204,884 (1993)) which sets the
groundwork for our extremely sophisticated cell sorting algorithm patents
(U.S. Patents 5,199,576 (1993); and 5,550,058 (1996)). If you have
difficulty obtaining any of these publications I would be happy to end
you a copy.
James F. Leary, Ph.D.
Chief, Molecular Cytometry Unit
Division of Infectious Diseases
Professor of Internal Medicine; Pathology; Genetics;
Biophysics; Immunology/Microbiology; and Biomedical Engineering
University of Texas Medical Branch
Galveston, TX 77555-0835
Tel: 409-747-1927; fax: 409-772-6527