Hello All, I have a question regarding the proper way of analysing FACS-data. I realise that the question may be off topic and I apologise. If it is too far off, please advise me where else to go. I am relatively new to FACS measurements (though not new to programming). In the laboartory where I just started, they treat cell cultures in triplicates and measure for some parameter on the FACS, e.g. DNA fragmentation or exposure of phosphatidylserine. After FACS measurement, the percentage of cells with for instance fragmented DNA or exposed PS is found, and the mean form the triplicates is taken. Later bar-graphs with standard variations are created. As we collect at least 3 x 10.000 events, I somehow feel that collapsing them all onto basically just three measurements, and then ask if those three measurements are different from three other experiments may not be the best way to go. I would like to know if there is a significant difference after treatment with some compound, and also compare treatment times. Is there a better way than taking the means of triplicates? Any solution should produce bar-graphs with quantitative differences. Thanks in advance. Ulrik

Hi Ulrik, I'm far from being an expert in analysing FACS-data but I can suggest you some tools implemented in rflowcyt (see the vignette): - you should use visualization tools (e.g. ECDF plot, density plot) to assess the quality of your replicates and verify the reproducibility of your experiment - rflowcyt also proposes some statistical methods (WLR.flowcytest, KS.flowcytest...) to test for the difference between two univariate distributions, e.g. treated and untreated with or without a time parameter. You can also try to post your message to Purdue University Cytometry Laboratories' e-mail archive http://www.cyto.purdue.edu/hmarchiv/cytomail.htm It is the main cytometry e-mail archive. Hope it helps, Nolwenn Ulrik Stervbo wrote: >Hello All, > >I have a question regarding the proper way of analysing FACS-data. I >realise that the question may be off topic and I apologise. If it is >too far off, please advise me where else to go. > >I am relatively new to FACS measurements (though not new to >programming). In the laboartory where I just started, they treat cell >cultures in triplicates and measure for some parameter on the FACS, >e.g. DNA fragmentation or exposure of phosphatidylserine. After FACS >measurement, the percentage of cells with for instance fragmented DNA >or exposed PS is found, and the mean form the triplicates is taken. >Later bar-graphs with standard variations are created. > >As we collect at least 3 x 10.000 events, I somehow feel that >collapsing them all onto basically just three measurements, and then >ask if those three measurements are different from three other >experiments may not be the best way to go. > >I would like to know if there is a significant difference after >treatment with some compound, and also compare treatment times. Is >there a better way than taking the means of triplicates? Any solution >should produce bar-graphs with quantitative differences. > >Thanks in advance. >Ulrik > >_______________________________________________ >Bioconductor mailing list >Bioconductor at stat.math.ethz.ch >https://stat.ethz.ch/mailman/listinfo/bioconductor >Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor > >