Scatchard Analysis Software
peterg at rnaworld.bio.ukans.edu
Thu Jan 5 19:18:43 EST 1995
In <3ehenh$fbr at lastactionhero.rs.itd.umich.edu>, Chris Beecher <cab at umich.edu> writes:
>I am looking for PC or Mac software that can do a Scatchard analysis for
>at least one-site ligand binding experiments.
>I am already familair with Ligand and it barely gets by, can anyone
>recommend something better without spending an arm and a leg?
>Univ of Mich
Ligand binding is not determined from a Scatchard plot but from least-squares
non-linear regression fitting of the hyperbolic binding curve to the raw data
(bound ligand vs total ligand). For a simple example, see Chen et al, FEBS
Lett. 298, 69-73 (1992). Any sort of linear transformation of the data
introduces statistical errors which make the results less reliable than direct
curvefitting to the unmanipulated data. This is especially true for a
Scatchard plot, in which calculation of [free ligand] is often unreliable at
very low [total ligand]. The *easiest* program for non-linear curvefitting is
the old PC program Enzfitter (BioSoft; avail from Sigma and other scientific
software distributors). Enzfitter is expensive ($250-350) but lets you enter
your own equations; it uses linear transforms, if possible, to obtain initial
estimates for the non-linear curvefitting. It also lets you adjust the type
of robust weighting. It has excellent graphing capabilities; they require
work and are not publication-quality, but you can transfer everything to s
simple plotting program for final figures if necessary. A slick (and $$)
versions of this program are Graft-It for Windows, and Ultra-Fit for the Mac.
Probably mush easier to use, but I haven't tried them.
The *cheapest* way is to use the freeware program Hyper (for Windows). It's
available from ftp sites; I don't remember which one (maybe IUBio?) Hyper will
fit to the Michaelis-Menten equation, which is a plot of [bound ligand] vs
[total ligand] for any ligand which is not a substrate. Hyper is excellent in
that it also plot all the common linear transforms (Lineweaver-Burk;
Eadie-Hofstee-Scatchard; Haines-Woolfe) to let you see how their answers
differ from the true answer.
A compromise is to use any one of the many scientific graphing programs which
have built-in non-linear curvefitting capabilities. SigmaPlot is one; PSI
Plot (only $50 academic) is my favorite. These curvefitters require a little
work to set up the equations, especially if you want to use robust weighting.
If you can get the $, get Enzfitter or a similar program. You'll use it
| Peter Gegenheimer | pgegen at kuhub.cc.ukans.edu |
| Departments of Biochemistry | voice: 913-864-3939 |
| and of Botany | |
| University of Kansas | FAX : 913-864-5321 |
| 2045 Haworth Hall | "The sleep of reason produces |
| Lawrence KS 66045-2106 | monsters." Goya |
More information about the Bio-soft