[ sedfit help home ]

[ Data ] [ Copy ] [ Display ] [ Model ] [ Parameters ] [ Run ] [ Fit ] [ Generate ] [ Statistics ] [ Notes ] [ Options ] [ Help ]

This command executes a boundary model, i.e. it performs a simulation of the sedimentation process, using the parameters that were entered in the parameter box.  

The difference between the Run and the Fit command:

The Run function does not make a nonlinear regression of the data.  As always in nonlinear regression, one needs to provide starting guesses that are reasonably close to the optimal parameters.  This can be trivial in simple models, but it can be quite difficult for complex models.  Therefore, it is necessary to test a given set of parameter manually, and compare the data with the model functions for these parameters.  This is precisely the function of the Run command.  

However, the Run command is slightly more advanced:  Several parameters of a boundary model are always trivial to optimize mathematically, but are difficult to guess manually.  These are linear parameters -- for those there is a unambiguous best-fit value.  (This best-fit value is calculated by an algebraic equation.)  In order to make the estimation of the nonlinear parameters easier, and the assessment of the quality of starting guesses more transparent, Sedfit optimizes the linear parameters already in the Run command.  

These linear parameters are:  

    baseline, systematic noise parameters, loading concentrations (except for the self-association models), and size-distributions

The fit command is used only to optimize the nonlinear parameters, which requires non-linear regression.

Examples for nonlinear parameters are:

    sedimentation coefficients, diffusion coefficients, molar masses, frictional ratio, meniscus, bottom, binding constants, loading concentration in self-association models. 

This means that in general, after setting up the parameters in the parameter box, one will first execute a Run command, possibly readjust manually some of the nonlinear parameters, and then use the Fit command to let the parameters converge. 

After the simulation is done:

The calculated concentration distributions at the times of the loaded scans are displayed as lines in the data graphics, and the new residuals are displayed in the residuals plot.  For the distribution models, a distribution plot appears. 

All the information of the calculated concentration, residuals, size-distributions can be retrieved in different forms using the copy commands. 

Additional information will be given in the left hand side of the Sedfit window, including the rms error (rmsd), the number of data points (n) and the sum of squared residuals (SSR). This information can be restored with the menu function show last fit info again.

Please Note: This function requires data to be loaded.  For generating simulated data, use the function generate.