Re: st: one/two-site competitive binding ado

[David Airey <david.airey@Vanderbilt.Edu>]

Joseph Coveney kindly replied:

> David Airey asked, "Has anyone ado files to perform nl regression 
> models for one and
> two site competitive binding assays? There are several other 
> pharmacological assays in
> GraphPad Prism 3.0 that I wondered if anyone has duplicated in Stata?"
>
> In the first volume of Stata Technical Bulletin Reprints, pp. 68-73 
> and continuing pp. 73-
> 76, Paul Geiger provided descriptions of using Stata for 
> radioimmunoassay ligand-
> binding analysis.  This would be one-site binding. The program he used 
> was an early
> version of a nonlinear least-squares regression ado-file in Stata, but 
> an analogous setup
> using the current -nl- ought to be straightforward.
>
> If you have the nonlinear function written out for two-site binding, 
> then it ought to be
> fairly easy to implement in Stata using -nl- or even -ml-.  The latter 
> has advantages in a
> -cluster()- option; see the FAQ by Weihua Guan at
> www.stata.com/support/faqs/stat/nl_ml.html .
>
> I'm not sure what GraphPad Prism 3.0 has for the other pharmacological 
> assays.  In
> general, if the functions are available, then setting up the proper 
> ado-file to be called by
> -nl- will be feasible.
Thanks for this information. I'll likely give one of these equations a 
whirl to compare results. Others in the lab uses Motulsky's GraphPad 
prism because it's easy to use and the documentation is well done if 
full of typos. Be nice to have extended capabilities present in Stata 
that are not present in Prism.

From what I can tell, Prism 3.0 has several built in (and modifiable) 
classic equations, such as:

1. one site binding (hyperbola)

Y = Bmax * X
    --------
    Kd + X

two site binding

Y = Bmax1 * X   Bmax2 * X
    --------- + ---------
    Kd1 + X     Kd2 + X

2. sigmoidal dose-response

             (top - bottom)
Y = bottom + --------------------
             1 + 10^(logEC50 - X)

3. sigmoidal dose-response (variable slope)

             (top - bottom)
Y = bottom + ------------------------------
             1 + 10^(logEC50 - X)*hillslope


4. one site competition

             (top - bottom)
Y = bottom + --------------------
             1 + 10^(X - logEC50)

5. two site competition
                               fraction_1              1 - fraction_1
Y = bottom + (top - bottom) *( --------------------- + 
-------------------- )
                               1 + 10^(X-logEC50_1)    1 + 
10^(X-logEC50_2)

and a few others:

6. boltzmann sigmoid

7. one phase exponential decay

8. two phase exponential decay

9. one phase exponential association

10. two phase exponential association

11. exponential growth

12. power series

13. polynomial equations

Cheers,

Dave

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/