Fwd: Cell surface marker data analysis

Hello, everyone.

The following is a re-post of a query that I asked back on Jan. 12.  I have 
been asked to post a summary of the replies, which I will do in a few 
days.  However, since I received a lot of "out of office" replies and 
relatively few other responses, I must have caught a lot of you still on 
vacation.  So, I am giving you another chance to put in your 2 cents before 
I post the summary.  There is a figure that makes the story easier to 
follow available 
at: 
http://www.cyto.purdue.edu/MD-parts/b29ed6aa86684747e3d13dccf5c4017d0302771d.JPG 


Thank you.

Rick Meister

>Hello, all,
>
>The attached file is an overlay of three histograms of an FITC-labelled 
>(indirect ) monoclonal antibody to a cell surface protein.  This cell 
>surface protein is expressed in these cells only after an inducer is 
>transfected into the cells.  So, the three plots are:
>
>blue = transfected with empty plasmid ( - )
>purple = transfected with plasmid containing a mutant gene which should 
>have little or no effect
>red = cells transfected with plasmid containing intact gene (WT), which 
>should yield (+) cells
>
>The cells were run on a FACSCalibur and were gated on a more or less 
>homogenious SS vs FS population.  Unstained and isotype controls looked 
>essentially identical to the empty plasmid control (blue), and were set to 
>lie within the first decade of the 4-decade log scale.  Finally, the 
>bright (3rd -4th decade) population of the WT-transfected cells (red) is 
>similar in intensity to that of cells which constitutively express the 
>cell surface protein (bright +).
>
>The PI of the project was not satisfied with the original data analysis 
>which consisted of the usual statistics (%, Mean, CV, Geometric mean, 
>etc.) for regions M1 and M2, defined by the negative controls as 
>"FITC-negative" and "FITC-positive" respectively.  He contends that all of 
>the WT-transfected cells have shifted as a population and, therefore, all 
>are essentially "positive" .  He wants to take the mean fluorescence 
>intensity of the entire population under M3 (ch. 1-1024) [approximately 3, 
>17 and 113 for plasmid, Mut and WT respectively] and normalize the values 
>to the empty plasmid (blue) by dividing through by 3.	Then, he would 
>claim that the Mutant-transfected cells were 6X brighter (and the 
>WT-transfected cells 38X brighter) than the plasmid control and use that 
>for a measure of transfection efficiency.
>
>I have argued that: a) any cell within the first decade of the log scale 
>has the same fluorescence intensity as the empty vector cells and should 
>not be called (+);  b)        that 100% transfection efficiency is not 
>likely (but would be necessary if all stained "positive"); and c)  that 
>dividing through by the MCF of the empty plasmid control is not valid 
>because the fluorescence signals went through a log amplifier before being 
>converted to digital format (ie, MCF of 6 is not necessarily 2X as bright 
>as and MCF of 3 on the log scale).
>
>Any opinions on whether there is a grain of truth to either of our arguments?
>
>Thanks in advance,
>Rick Meister
>
>
>
>Richard K. Meister
>Manager
>University Cell Analysis and Sorting Core
>424 DHLRI
>473 West 12th Avenue
>Columbus, Ohio 43210
>(Phone): 614-292-3569
>
>
>Content-Type: text/plain; name="warning1.txt"
>Content-Disposition: inline; filename="warning1.txt"
>Content-Transfer-Encoding: 7bit
>MIME-Version: 1.0
>X-Mailer: MIME-tools 5.411 (Entity 5.404)
>
>This attachment - 'W6_32 overlay.JPG' -  17.56 KBytes - can be viewed at
>http://www.cyto.purdue.edu/MD-parts/b29ed6aa86684747e3d13dccf5c4017d0302771d.JPG 
>