doNls {mixlow}R Documentation

Conduct nonlinear least-squares analysis of sigmoidal concentration-response data

Description

Organizes input and calls the nls function to estimate the parameters of sigmoidal concentration-response curves. Users can supply starting values for parameters or can use default values that are based on the data.

Usage

doNls(mixlowData, parameterDefaults, lambdaThreshold=0.05, numRandomTries=10, 
verbose=FALSE) 

Arguments

mixlowData A list obtained from prepareData.
parameterDefaults A list obtained from getNlsParameterDefaults. The default values can be edited as desired before sending to the doNls function.
lambdaThreshold A scalar designating any lower boundary for lambda. If a model estimates a lambda value less than the threshold, that model will be discarded and a model that does not use a lambda parameter will be used instead.
numRandomTries A scalar designating the number of attempts that should be made to reach convergence for each model. For each attempt, starting values are randomly selected from a normal distribution with mean equal to the given parameter starting value and a standard deviation equal to the given standard deviation.
verbose An optional logical value. If TRUE, information from intermediate steps in the doNls analysis is printed. Default is FALSE.

Details

See doNlme for a description of the parameters estimated. See getNlsParameterDefaults for a description of default parameter values. Parameter values of NaN indicate the following default values will be used: param.g=0, param.g.std=1, param.p = the closest tested log concentration corresponding to 1/2 the mean response of control wells, and param.p.std = the absolute value of param.p/2. The default for param.lambda is NaN, which indicates that a lambda term will not be used. If a lambda term is used, its value should be between zero and 0.5.

Value

A list of type nlsData of length equal to the number of trays with the following components:

nlsEstimates Data frame containing parameter estimates
nlsGraphing List used by plotNlsData for making graphs:
tray
String tray name
drug
String drug name
xx
Concentration values for graphing predictions
pred
Predicted values
y
Adjusted observed responses
x2
Adjusted observed concentrations


Adjusted responses take into account responses in optical control wells (see prepareData). In the adjusted concentrations, values of zero are replaced by concentrations of 1/1000th of the lowest nonzero concentration (for log-plot graphing purposes only).

Author(s)

John Boik jcboik@stanford.edu

References

Boik J.C., Newman R.A., Boik R.J. (2008) Quantifying synergism/antagonism using nonlinear mixed-effects modeling: a simulation study. Statistics in Medicine 27(7), 1040-61

See Also

prepareData, plotNlsData

Examples


# mixlowData data object is obtained using the prepareData function
data(mixlowData)
trays <- getTrays(mixlowData)
parameterDefaults <- getNlsParameterDefaults(trays=trays[1:9])
nlsData <- doNls(mixlowData=mixlowData, parameterDefaults) 

[Package mixlow version 0.01.1 Index]