Dima Klenchin klenchin at
Tue May 23 20:31:27 EST 2000

In article <OF60D10268.AA29E2CB-ON032568E8.005A6B50 at>, a.krimer at wrote:
>The original Scatchard paper is about the interaction between Albumin and
>cupric ions. Scatchard developed a mathematical model fitting this
>phenomen. Today it is used for calculating any interactions between a
>receptor and a ligand.
>Therefore the protocol for receptor analysis is quite different and the
>value of Scatchard's work is the mathemathical analysis that could be
>performed by simple calculations.
>Dear Chris, may be Jeff is lazy, may be not; but your knowledge about
>receptor analysis seems to be not your best.

Jeff is most definitely lazy. Scatchard plot is covered in _every_
biochemistry book without exception. It is nothing more than a 
simple transformation of mass action law (covered in every 
chemistry book). The analysis is exactly the same for any given 
pair of interactors, be it ions (for example, Ca2+ chelation by BAPTA)
or ligrand-receptor interaction (with the exception that one must always
do controls and subtract non-specific signal generally significant
enough in biological assays). 

The Scatchard plot has (had) two values: first, it showed how a complex
parameter can be easily and elegantly obtained from a simple experiment, 
second, at allowed reasonably precise measurements of Kd without the need
for complex non-linear interpolation, which is not easy to do without

Nowadays, using Scatchard for purposes any other then illustration 
(say, when there are two binding sites) is a crime agaist common sense.
Even Excel, not to mention a huge number of specialized programs, many of
them freeware, can perform non-linear regression analysis in a matter of 
seconds and give a more reliable result than the one obtained from 
Scatchard (which is, of course, is going to be as good as the data set). 

There is a multitide of papers discussing in great detail just what erroneous
results straight Scatchard plot can produce. Among them, my favorite is
the one that analysed a great number of published Scatchards that showed
SD for data points and concluded that majority of them showed sighs of 
extensive data massaging (if not an outright fabrication) on the basis of 
the fact that the data points showed on Scatchard plot had generally 
the same value of SD, whcih should not and cannot be the case with
real life data. 

        - Dima

More information about the Methods mailing list