You do have to be a little careful with intensity comparisons, and unless you
are going the whole hog and using a calibration curve, I would recommend
using general, qualitative terms to describe qualitative data (several times
brighter, an order of magnitude dimmer etc.).  In addition to the bleaching
problem you mentioned, here are some other pitfalls:-

Black level settings are often kept high for making attractive images, but
should be lowered for quantitation such that there is a non-zero value for
each pixel.  Imagine a blob with 100 arbritrary units of fluorescent protein,
and another with only 20.  If the bottom of the 8 bits is set at the 10 unit
mark, the apparent difference in signal will be 90:10 instead of 100:20 if
set at zero; nine to one instead of five to one.  Set the black level to
minus 20 and you get 120:40 or three to one, unless you account for the
background contribution.

Depth within the specimen will affect the delivery of light to and from the
target voxel.  As you scan deeper through biological material, the image gets
dimmer due to losses from refraction, reflection, and absorbtion.  Unless you
are scanning right against the cover slip, you cannot easily accout for this
by sticking to a constant depth, since biological tissue is inherantly
variable.  Refractive, reflective and absorbant regions can be highly
localized, and can cast deep shadows into underlying material.

Quenching is not likely if you are frugal with reagents, but if you are
making quantitative claims, you may need to corroborate your results at
different loading densities, just to be sure that the localized packing of
dye isn't so high that it is quenching itself.

There is plenty that can go wrong, so it may be safer to let the images do
the talking.  A side by side comparison shows there is more stuff in one lot
of cells than there is in the other, and leave it at that.

David Carter
Meridian Instruments
oQQQQQ@

<<I'm confused, why would an '...intensity of 50 be 2,5, (or even 20
fold less) (in terms of amount of antigen) than an intensity of
150...'? 50 is 3 fold less than 150 (assuming no background). The
other problem is that any bleaching will make such comparisons
impossible. Other than that the A/D converter should be completely
linear within it's 8 bit range..

Regards Mark Cannell

> From: Wayne Arnold <[log in to unmask]>
> Date: Mon, 18 Aug 1997 13:57:50 -0400
> Subject: intensity vs amount of protein
> To: [log in to unmask]
>
> I am using a Biorad MRC 600.  Is there a way to correlate
> differen't intensities with amount of antigen present in an
> image. ie I don't want to quantify the antigen, I just need
> to know at a given setting for the photomultiplier can I
> determine if say an intensity of 50 is 2,5, or 20 fold less
> (in terms of amount of antigen) than an intensity of 150.>>