Experience and experiments with Matlab...

regressContourPlot
You can fit a linear regression line on 2D histogram data, i.e. if you already don't have the original data set containing two measurements for each data point.
Syntax:[r,slope,intercept]=regressContourPlot(twoDHist,xScale,yScale)
Help is available by typing "help regressContourPlot" at the Matlab command prompt.
createTrend
A trend line can be created from a data set containing two measurements for each data point. In order to create a trend line the X range is divided into the given number of bins and the mean of the Y variable in each bin is calculated.
Syntax: createTrend(XYData) or [trendline,binsX. freq]= createTrend(XYData,XMin,XMax,NumBins)
Help is available by typing "help createTrend" at the Matlab command prompt.
createContour
You can generate a 2D histogram (contour plot) interactively from a set of data containing two measurements for each data point.
Syntax: createContour(XYData)
Help is available by typing "help createContour" at the Matlab command prompt.
Fitting the Hill equation to data points: fithill
The program fits the Hill equation to measurement data. The Hill equation has the following forms depending on whether IC50 or EC50 is fitted on concentrations on a linear or logarithmic scale:
The program can be run in GUI mode or command-prompt mode.
- GUI mode: In order to run the program in GUI mode start it without input arguments, i.e. type "fithill" at the Matlab command prompt.
- command prompt mode:[fittedParameters,rsquared]=fithill(xdata,ydata,PropName,PropVal1,PropVal2,...) OR [fittedParameters,rsquared,fittedN]=fithill(xdata,ydata,PropName,PropVal1,ProbVal2,...)
Output argument 'fittedN' is required if propName 'logconctoplot' is given among the arguments.
xdata = LOG10 of the drug concentrations (LOG10!!!!!!!)
ydata = cell numbers or absorbances
fittedParameters = structure with the following fields:
fittedParameters.min
fittedParameters.max
fittedParameters.kd
fittedParameters.n
fittedParameters.minCI
fittedParameters.maxCI
fittedParameters.kdCI
fittedParameters.nCI
rsquared = goodness of fit
fittedN = calculated curve using the best fit parameters at the concentrations given in the argument list after 'logconctoplot'
The following parameters are followed by two numbers:
If 'ec50' and 'ic50' is not followed by a numeric parameter, it only specifies whether a decreasing or increasing curve is to be fitted.
The following parameters are followed by one number or variable:
confint- confidence bounds will be calculated at the % level specified by the number following 'confint'
logconctoplot - LOG10concentrations (!!!!) at which fittedN will be calculated using the fit parameters
The following parameters aren't followed by anything.
ver - version of the program
version - same as above
help - displays this help
Download: fithill.p