Dima Klenchin klenchin at facstaff.REMOVE_TO_REPLY.wisc.edu
Thu Aug 27 18:46:40 EST 1998

In article <6s16s5$o92 at panix3.panix.com>, iayork at panix.com (Ian A. York) wrote:
:In article <35e40dba.2700387 at news.univie.ac.at>,
:Martin Offterdinger <a8803349.nospam at unet.univie.ac.at> wrote:
:>I would like to know if you would consider it as science to quantitate
:>bands obtained by enzymatic Western Blotting with a densitometer. I
:I have seen this done, but only as "semi-quantitative", not
:"quantitative".  As with any other method, you need to take great care
:that your antibody isn't limiting and so forth.  One way of making the
:data a little more convincing is to do serial dilutions of your lysate,
:say 3-fold.  Your densitometry should accurately detect the known dilution
:of each sample, and the ratios between the different samples should remain
:constant at the different dilutions.  If either is not true, your blots
:are not useful for quantitation.
:My own limited experience in trying to get accurate quantitation from
:westerns is that it's not very reliable, but that if you emphasize the
:"semi" part of the semi-quantitative you can get some information from it.

Totally agree! No matter what, there always must be a calibration curve
from which it is seen that the numbers are derived from a meaningful
part of the curve. Calibration should be run on the same blot, of 

Myself, the best I can do so far, is to say that one can be confident that 
the amount of the protein in the experimental point is <25%
but >12.5% of control. Good enough for my purposes.

But literature is full of the most pathetic examples of "quantitative" 
Westerns where people just scan films and compare pixels in two lanes. 
Never mind that there are 4 non-linear processes underlying 
detection. Ten times less pixels => ten times less protein!

        - Dima

More information about the Methods mailing list