In <46ovj1$n5e at ixnews3.ix.netcom.com>, jsalb at ix.netcom.com (Jesse Salb) writes:
>Anyone know of enzyme kinetics software for Windows which will do
>Lineweaver Burk plots and is very flexible with respect to plot
>appearance? I'm using Grafit now but it's highly user unfriendly,
>though admittedly very flexible if you want to spend a few years
First, for anyone looking, you DON'T do Lineweaver-Burk plots to get kinetic parameters. (I'm sure Jesse is using GrafIt to do least squares non-linear curvefitting.) SECOND, GrafIt is a Win version of Enzfitter, an old DOS program available from BioSoft (Missouri, USA) and via the Sigma catalogue. They both originated from Robin Leatherbarrow, an old-line kineticist. Enzfitter defines the category of enzyme kinetics software. Clunky, DOS-based, doesn't do publication-quality output, BUT is easier to use f
or basic curvefitting than any other program on the market. I'm not aware of any program that has a better command of non-linear curvefitting. For example, Enzfitter allows you to choose all possible weighting methods, not just one or two, including weighting by 1/y; 1/(y^2) for cpm data, and weighting by direct use of std. dev. of each data point. Enzfitter makes it easy to plot resaidualls, etc. Best of all, Enzfitter is unique (except for GrafIt) in that if the equation you're trying to fit can be linea
rized, Enzfitter automatically does so, uses least-squares regression (which doesn't require any initial estimates of parameters) to calculate the initial estimates, and then feeds those estimates into the Marquardt-Levenburg non-linear regression procedure (all non-linear methods require initial estimates of the final values).
SO, why try anything else? Well, Enzfitter still costs $395 (certainly was worth it to me, though); doesn't do publicaiton graphs, and can't save the graphs. It's a calculation tool.
The best values in curve-fitting programs that will do publication graphics are:
(1) PSI-Plot (DOS/Windows/32-bit Windows; $50 or 100; Polysoft International, Salt Lake City, Utah; 801-485-0466/ FAX: 801-485-0480). Easiest graphing program I've seen. Totally object-oriented, just click on any plot element to alter it. Has a vast array of math & statistical analyses built-in (Fourier transforms, ODE solver, ANOVA, etc.). Does not have built-in non-linear equations, but it's easy to write and save your own. My favorite for plotting, 2nd or 3rd (after Enzfitter & Prism) for raw curve-fi
(2) Graph-Pad Prism (Graph-Pad, student price $99 from Softshell [Grand Junction CO]: 303-242-7502/ Fax 800-240-6469). Prism has the wonderful feature that you can organize raw data, graphs, ands text notes into separate "folders: within the same data file. All calculations and curve-fitting are automatically redone each time any of the orginal data is changed, or any computation parameters are altered. Good selection of built-in equations. Makes an educated guess at initial estimates for non-linear fitti
ng. (These estimates aren't derived from linearizations, and - although they are user-selectable - they don't work well on equations I entered myself.) Good graphics and a good text editor with greek & math symbols. Can automatically calculate & plot mean & std. dev from several columns representing replicate data runs.
There are many more out there, several with more features and more control over the curve-fitting procedure, but these are the best I know of for the price. Frankly, even though I use OS/2 Warp, I almost always go the Enzfitter first to get the kinetic values, and then use another program to make a nice plot.
| Dr. Peter Gegenheimer | voice: 913-864-3939 FAX:913-864-5321 |
| Departments of Biochemistry | pgegen at kuhub.cc.ukans.edu |
| and of Botany | URL: gopher://rnaworld.bio.ukans.edu:70 |
| | |
| University of Kansas | "When you have excluded the impossible, |
| 2045 Haworth Hall | whatever remains, however improbable, |
| Lawrence KS 66045-2106 | must be the truth." Sherlock Holmes |