Quantitating Westerns

Dima Klenchin klenchin at REMOVE_TO_REPLY.facstaff.wisc.edu
Sun Jun 25 10:54:16 EST 2000


In article <8ivppq$qf$1 at uni00nw.unity.ncsu.edu>, kaghoram at unity.ncsu.edu (Karthik Aghoram) wrote:
>
>Hi,
>
>Is there a reliable way to quantitate protein expression from Western
>blots (am using a chemiluminescence system for band visualization)?  I
>have heard some phosphorimagers can also quantitate luminescent signals,
>but the one here (Mol Dynamics) does not.  Any ideas would be highly
>appreciated. 

No matter how linear your detection system is _supposed_ to be, 
when quantifying anything you must have calibration curve that covers
your range of signal and is not saturated in this range. 

Simple, really. In ideal world the standards are purified proteins 
which are run as serial dilutions. This, however, is not necessary
in many applications. For example, when I assay for some weak 
membrane binding,  I simply run serial dilutions of total cell lysate
as calibration curve, and equivalent amount (by volume, _not_ by
total ptotein) of total cellular membranes treated differently as
samples. Then I either do phosphoimager or film (in fact, to make
it more reliable, I use BOTH - data from phosphoimager + data
from two different exposures on film, averaging results in the end).

Scanning and qualtitation can be done in anything capable of 
counting pixels. From freeware such as NIH Image to overpriced  
specialized programs to Adobe Photoshop. As long as you have 
correctly done calibration curve (run on the same gel!), it really
does not matter.

90% published figures on western "quantitation" is done by 
counting pixels after film scanning without any calibration curve. 
This is a wonderful method to obtain any desirable result regardless
the reality.

        - Dima






More information about the Methods mailing list