I've come across this odd threshold behavior on the BD FACSCalibur which I thought I would mention since I haven't seen any notice of it on this list. It can be summarized as follows: The measurement values of first laser fluorescent signals decreases if a forward scatter threshold is set "too near" the forward scatter signal of the particle. Since it is not uncommon to threshold high to eliminate debris, I looked at this a little closer (data below). Although I don't think this will affect the majority of experiments, there could be cases where it does. BD is aware of this behavior (PIR submitted through Dave Houck). The exact amount of measurement change varies from machine to machine, so the data below is more qualitative than quantitative. In the "bead compare" document, the left column of plots are FSc x SSc with different FSc thresholds indicated. The right column are the single beads with the median listed. For this set, the bead measurement dropped from 619 units to 284 units. The "Bright cell stain" document shows this affect on the a bright positive population (CD45) of PBLs. The left column shows FSc histograms at various threshold levels. The right column is the CD45 positive lymphs, showing a decrease in median from 1537 to 1426. If one were measuring percent positive populations, this should not affect the results. The "Dim cell stain" shows this artifact on dimly staining cells (IF-gamma). The columns are as above. There is about a 5% decrease in the percent positive - in this case not a significant change. However, for other instruments, this may be greater (or less). Caveat emptor!
This archive was generated by hypermail 2b29 : Wed Apr 03 2002 - 11:57:48 EST