In the spirit of sharing all the perspicuity offered by the 2160 denizens of the Purdue cytometry mailing list, I am relaying the answers I received to what I thought would be the "elementary" question of whether the sample concentration ( hence the count rate) affects the data one obtains during a FACScan run. I had expected to get several definitive answers from my more experience brethren. What I found out is I'm less of a tenderfoot than I thought. Therefore, I feel confident enough to state that I personally find Calmain Prussin's and Paul Fallon's answers to be the most cogent and Hank Pletcher's last comment about "three different flow cytometrists" the most apposite. My surmise of the situation is that there is indeed something in the circuitry that detects and eliminates cell signals that are too close to each other. However, if the cells are really, really, really close to each other, the circuitry will think two cells are one and the signals will register as one just as two cells stuck together register as one and we all know we have to turn on our doublet discriminators when we do cell cycle analysis. If true, this would imply that when we compare antigen levels between samples, like before and after treatment, we should be careful to suspend and hence run the cells at the same concentration, lest we get artifactually high MFI's for the sample that is the most concentrated. Argh! something else to be careful about. So here they are: ____ Hi Dan, I have a FACScan also, and my understanding is that if there are too many cells for the machine to gather all the information on, a percentage of cells will be ignored and their data will not be included in the events you've collected (eg, more than 1 cell passes through the flow cell in the path of the laser simultaneously). You can slow the flow rate to low on the control switch, and/or dilute the sample a little. I try not to have the cells more concentrated than a couple million per ml. I run a machine for research, not a clinical one -- I do not know if the clinical protocols allow you to do this. Good luck! Beverly Barton, Ph.D. Assistant Professor Department of Surgery UMDNJ-New Jersey Medical School PS--Call BDIS and ask tech support--they will have the definitive answers! Best, Beverly ------- Increasing the flow rate increases the frequency of doublets (2 cells) being detected as single events and thus yielding false double positive cells. A couple of years back we tried running our FACSCalibur at rates over 3,000 events per second in order acquire sufficient number of events to detect Ag specific cytokine responses. in a pilot, we stained separate samples for CD3 FITC and CD3 PE and then mixed the cells together after staining. Gating on lymphocytes by scatter reduced the number of double positive cells, since most of the doublets had a greater forward scatter. Even though we gated on the lymphocyte cluster, as the rate increased, so did frequency of double positive events. For the events we were studying, we found that 3,000 events was the maximum event rate we could use and still be assured that our double positive cells were actually not 2 signal positive cells triggering at the same time. Calman Prussin ----- My vote is on ignore. FACS user since 1983. Deb ----- I think the electronics detect coincident events and drop them from analysis. The net effect is lower flow rate of events than reality. The fluorescent measurements should not be effected. The was a posting in the archives that measured the flow rate accuracies of the FACScan, but not the fluorescence. The fluid rate effects the sample core, so you will expect lower CV's on the Lo flow rate. ------- Hi Dan, My understanding is that if another pulse (i.e. cell) arrives before the instrument is finished processing the first one, it aborts both events and waits for the next one to come through. You don't ever see these abort events, but the cell sorters have counters that can keep track of them. Because cell sorters go much faster, and because we worry about recovery, we keep track of these things. I seem to remember that the deadtime, or processing time, of a FACScan is around 250usec. That's why they recommend no more than 1500 events/sec as a trigger rate. Newer FACSCaliburs have significantly shorter deadtimes (perhaps 50usec), and cell sorters are even faster (3-8usec). As a result, going too fast just makes you less efficient, but doesn't alter the data. It just means you don't see cells that go through too close to each other. Of course, cells like to associate with each other, so I suppose you could argue that there is a selective cell loss. BTW, you should know that if you ask three different flow cytometrists a question, you'll get three different answers. I've been asking folks at BD about system aborts for many years, and I still don't know the whole story. I do have an old timing chart for abort events on the old cell sorter, if you're interested. Hank Pletcher ----- I use a facsCalibur - even for events within the range of tolerance for my machine I will notice that at 20% max I will have lower CV than if I concentrated the sample and ran it at 99% max thruput. on all parameters. -maciej ----- test this on your machine by making a mixture of stained and unstained beads at high concentration. I ran them up to 3000/sec and maintained low CV and good S/N separation but you can easily determine that the electronics are ignoring more and more particles as you go up in event rate. ----- Hey Dan, I would think that there would be an optimal concentration for efficiently running a sample on any given cytometer. Too low of a cell concentration and the event rate will be fast enough to have a unusable signal to noise ratio. Too high of a cell concentration and the cytometer will not only not miss events but will be unable to recognize individual cells from doublets/groups of cells that will pass within the same time window. Addition of a piggyback cells which are singly positive in a given channel while the first cell is positive in another will be recognized as double positive. Good luck, Paul Fallon -------- In addition to these replies, 9 people on the mailing list have their E-mail auto replies on and hence informed me that they are out of town.
This archive was generated by hypermail 2b29 : Sun Jan 05 2003 - 19:01:23 EST