fit_data

Data Fitting (fit_data)

Data fitting is performed with the ‘fit_data’ function. One needs the structure with detection vectors (r) output from the ‘detector_opt’ program. We refer to this as variable ‘r’. Then, one also must compile the experimental data. This is in a structure, ‘data’, with the following fields:

data.resi   :   Residue number. Optional field (used for plotting)
data.R1     :   R1 rate constants
data.R1std  :   Standard deviation of R1 rate constants
data.NOE    :   NOE rate constants (H–X transfer)
data.NOEstd :   Standard deviation of NOE rate constants
data.R1p    :   R1p rate constants
data.R1pstd :   Standard deviation of R1p rate constants
data.R2     :   R2 rate constants
data.R2std  :   Standard deviation of R2 rate constants
data.S2     :   S^2 measurement (not S, not 1-S^2).
data.S2std  :   Standard deviation of S^2 (~2*Sstd)

Note that the data fields (‘R1’,’R1p’,’NOE’,’R2’,’S2’) and their standard deviations are all matrices. Each row gives a different residue (or different data set, as desired). Each column gives a different experiment. The columns must be in the same order as was used when the experiments were initially entered into the ‘info’ variable before calculation of the rate constants. Then, ‘fit_data’ is called as

fit=fit_data(data,r) %r is the output of ‘detector_opt’

One may also specify a few settings for fitting, given in a structure ‘par’. These include settings for Monte-Carlo error analysis, and optional plotting.

par.n_rep   :   Number of Monte-Carlo analysis repetitions (default=200)
par.conf    :   Confidence interval for error analysis (default=.95)
par.fig     :   Figure numbers for optional plotting 
(give one integer for detector response only, two to also include data fit)

Then, one calls ‘fit_data’ as

fit=fit_data(data,r,par);  %par is optional