>It may be that this is a trivial question: >What is the effect of combining log and lin amplification, e.g. for >FITC and 7-AAD, on hardware compensation on instruments >using analog log amplifiers? My guess is that the setup probably >only has been designed and validated for log to log compensation. >Probably someone can comment on where the compensation is >actually performed on such instruments (before or after log >amplification)? > Hardware compensation pretty much has to be done on the linear signals from the preamps before they go into the log amps, since the universal compensation model is a linear one. The original paper on compensation (Loken, Parks, and Herzenberg, 1977) antedates the introduction of log amps to the Herzenberg lab; compensation was done on linear data, and was made necessary because only rectangular count and sort gates were available. Software compensation is also done on linear data; some programs transform log data to linear, compensate, and go back to log. The procedure Joe Trotter described was necessitated by the lack of red-green and green-red compensation on the FACScan and a lot of other systems. In general, there is not a lot of crosstalk between red (FL3, PE-Cy5 or PerCP) and green (FL1, FITC) channels when one stains surface antigens, and compensation between each of these and yellow-orange (FL2, PE) is all that is needed. Problems come up when brighter stains, e.g., 7-AAD, are used, and there can be a great deal of crosstalk between FITC on nuclear antigens and 7-AAD, probably due at least in part to energy transfer. This may represent some departure from the typical model of crosstalk on which compensation is based, and, in general, the settings needed for proper compensation will be different from those expected on the basis of analysis of singly stained samples. -Howard
This archive was generated by hypermail 2b29 : Wed Apr 03 2002 - 11:51:44 EST