need a curve fitting package

Duke Groebe drg at prophet.pharm.pitt.edu
Wed Feb 16 15:16:18 EST 1994


I> In article <1994Feb15.030258.13398 at cs.yale.edu>,
johnk at spasm.niddk.nih.gov
> (John Kuszewski) wrote:
> 
> > Hi,
> > 
> > I'm looking for a curve-fitting package to do a specific
> > problem in chemical kinetics.  (Sorry for the long 
> > description)
> > 
> > The function to be fitted is 
> > 
> > I = exp(-kf1 * tf) * (kex/(kex + kf2)) *
> >     (1 - exp( -(kf2 + kex)* tex))
> > 
> > kex is a known constant.
> > I, tf, and tex are known variables.
> > 
> > I need to optimize values of kf1 and kf2 against
> > sets of several (I, tf, tex) values.
> > 
> > Does anyone know of a program that'll do this sort 
> > of fit? Nothing I have (facsimilie, Kalidegraph, 
> > Mathematica) can do it.  I'd prefer something for 
> > a Sun or Mac that I can get for free.
> > 

Have you tried PROPHET?  It is a statistics/molecular modelling/sequence
analysis package licensed by the NIH and owned by Bolt Beranek and Newman,
Inc. (BBN).  We have it running on a Sun Sparcstation 2 (thus my e-mail
address). Contact BBN for more info:

Telephone: 617-873-2669
E-mail (info): prophet-info at bbn.com
E-mail (help): prophet-help at bbn.com

We're fitting differential equations that describe

L + R <> L.R > L-R
    +
    B <> R.B

in order to get the reversible on and off rates for L + R.  We measure R.B
formation and provide values for the other rates.  Works pretty well.

Good luck,

Duke

******************************************************************************
The opinions expressed here are my own.  MY very own and NOBODY else's!  I
THOUGHT OF THEM FIRST, HA HA, AND YOU DIDN'T!!  SEE!?  There's my name up
there!  And the date - Ooohh, the DATE!  That's important!  HEE HEE, I'M
GONNA BE RICH!  DO YOU HEAR ME!?  RICH!!  HA HA!!  They're mine...heh, heh,
all mine....  GOD!  I'm so smart!




More information about the Bio-soft mailing list