Re: relative fluorescence

From: maciej simm (simm@cd4cd8.com)
Date: Mon Jul 22 2002 - 11:52:17 EST


I found the below in the archive, hope it helps
browse the archives here -
http://flowcyt.cyto.purdue.edu/hmarchiv/cytomail.htm

Maciej


Hi Sandeep,

There's probably not too much difference between a properly calculated
geometric or arithmetic mean as long as you use the same in all runs (there
will be a difference between the two, but the calibration should be ok).

What do I mean by "properly calculated":? As I remember (I haven't used it
for six years or so), FACScan research software returns the results of its
calculations in "channel numbers", which need to be transformed into linear
units before they're used with any meaning:

linear value =10^ (channel value/256)

ie divide the numbers given as mean by 256, and then raise ten to the power
of this number. Note divide by 256 if you acquire ten bit data (channel
number 0 to 1023) on a 4 decade scale, if you're acquiring 8 bit (channel
numbers 0 to 255) data on a four decade scale you should divide by 64 (8 bit
three decade data needs a divisor of 256/3 = 85.333)
The outcome of doing the transformation after you've found the mean is that
you end up with the geometric mean - so you may be in luck anyway.

You can tell if you're getting channel numbers or ready-transformed results
quite easily - if you've got a negative peak that is spread over the first
decade of your log scale, it's mean would be around 100-150 if the software
is reporting channel numbers (so you'd have to apply the formula, or your
results will be wacky), and around 3 or 4 if the data are ready-transformed.

Cheers,

Ray

> From: sandeep ramalingam <sandeeprk@yahoo.com>
> Date: Wed, 2 May 2001 01:53:05 +0100 (BST)
> To: Cytometry Mailing List <cytometry@flowcyt.cyto.purdue.edu>
> Subject: Mean Vs. Geometric Mean
>
>
> Hello,
>
> As you probably read earlier, we have a HP based
> FACScan.  We are interested in quantifying antigens
> expressed per lymphocyte and are thinking of using the
> Quantibrite beads.  In the "white paper" on
> Quantibrite beads that we downloaded from the BD
> website, we noticed that they have used Cell Quest
> that calculates the "Geometric mean", while we are
> planning to use FACS research software that calculates
> the "mean".
>
> Does it matter?
>
> With warm regards,
>
> Sandeep.
>
> =====
> Dr. Sandeep Ramalingam,
> Department of Clinical Virology,
> Christian Medical College Hospital,
> Vellore-632004, India.
> Phone(off): 91-416-222102, ext 2070
> Fax: 91-416-232035
> Email: sandeeprk@yahoo.com
>
> ____________________________________________________________
> Do You Yahoo!?
> Get your free @yahoo.co.uk address at http://mail.yahoo.co.uk
> or your free @yahoo.ie address at http://mail.yahoo.ie
>


on 7/18/02 2:43 PM, Stewart, Judith at Judith_Stewart@URMC.Rochester.edu
wrote:

>
> I wonder if somebody out there could explain the benefit of using the
> geometric over the arithmetic mean when converting to equivalent
> fluorescence using Quantium beads.
>
> Also I have a shared user who would like to analyze their data on their own
> system.  They have a PC.  Can anyone recommend a good PC program that will
> read the CellQuest files.
>
> I'm relatively new to this, but I've already got several useful tips from
> reading the comments, thanks to all of you.
>



This archive was generated by hypermail 2b29 : Sun Jan 05 2003 - 19:26:16 EST