Subject: How fast can a FACS run? Dear Calman, A question we all come to at some stage; I'm glad that you raised it. A question that I don't remember seeing discussed: Probably not raised because it is a contentious issue. How fast (events per second) can a FACScan or FACSCalibur run? When I have tried it on a FACScan, it peaked at 10,000 events/s. This implies that its dead time is around 100 microseconds. By the time you get to that rate though, the Scan will be missing more than half the total events, and you will be getting many measurement errors due to undetected coincidences. These occur when two cells overlap in the flow path as they are illuminated, and the machine registers a signal which is a reduced sum of the signals each would separately generate. To minimise these errors, you need to run at a rate where the incidence of these is lower, and/or you need to set gates in s/w that will accept only singlet events. In the MoFlo, the pulse width enables you to do this exclusion/acceptance. You would need to talk to BD about that in their machines. It is harder to do in conventional signal space, since one effect of the undetected coincidences is to smear the normal lymphocyte scatter gate areas in both directions; you are placed on the edge of uncertainty!! We are looking at rare events (0.01-0.03% of total cells) and want to speed up the acquisition. One to three events per ten thousand is probably beyond the Scan when you run it at a rate which will yield a valid statistical sample, ie ca 10,000 events. This will take one-three hours; the time relationship is linear, so 1000 will take 6-18 minutes, and the error then 3%, not including the coincidence errors. Your solution is to reduce the machine's dead-time and run in peak mode (refer to BD for a retro-fit, of whatever their performance is claimed to be). There is a better alternative: rent or buy a MoFlo, either desktop or MLS system; it comes as an analyser or a sorter. The dead-times are respectively 15 and 5 microseconds, so you can run at 20,000 and 60,000 cells/s without much trouble from the sources I described. Take it up with me if you wish. Cheers, Bob _______________________ Calman Prussin Laboratory of Allergic Diseases NIAID/ National Institutes of Health
This archive was generated by hypermail 2b29 : Wed Apr 03 2002 - 11:49:58 EST