LD: Lethal dose

In lethal: Compute lethal doses (LD) with confidence intervals

Description

Compute generalized additive model for lethal dose finding and corresponding confidence intervals using parametric bootstrap approaches

Usage

## S3 method for class 'formula'
LD(formula, groups = NULL, experiment = NULL,
   lethal.dose = c(50, 10),
   dose_trafo = c("sqrt", "log", "none"),
   data, family = negbin(theta = c(0.01, 1000)), ...)

## S3 method for class 'LD'
LD(object, lethal.dose = NULL, group = NULL, ...)

## S3 method for class 'LD'
confint(object, level = 0.95, lethal.dose = NULL,
       B1 = 20, B2 = 100, newdata = NULL,
       myapply = mclapply, ...)

### just wrappers to LD functions as defined above
### (i.e., with same arguments)
## S3 method for class 'formula'
lethal.dose(...)
## S3 method for class 'LD'
lethal.dose(...)

Arguments

formula	Specify the main dose-response model in the form outcome ~ dose. Only used to extract outcome and dose varibles. Internally a more complex model is fitted. See details.
groups	(optional) define the name of the grouping factor (if available in the data set). This could be for example the species, the strain, or any other factor (with two or more groups).
experiment	(optional) define the name of the experiment variable if multiple replicates (biological or technical) are used. This variable is included as randondom effect in the lethal dose model.
lethal.dose	vector of LDXX values. Per default, c(50, 10) for LD50 and LD10. If no value is specified in functions for class "LD", the LD values of the fitted "LD" object are used.
dose_trafo	define a character string for the transformation that is used to stabilize the estimation of the dose effect. Note that this does not influence the form of the dose-response relation ship as we use a smooth function of dose (or a smooth function of the transformation of dose) in the model. This is really more a technical option.
data	data set with an outcome variable (i.e., number of surviving cells/bacteria/...), a dose variable and optionally a grouping variable (with two or more groups) that is used to comparte the groups. Further more, the data set can optionally contain an indicator for the experiment, which is used to model possible experiment effects.
family	distributional assumption. Currently only negbin is
object	Lethal dose model fitted using function LD.
group	(optional) specify a factor level of groups to extract only LDxx values for one of the groups.
level	coverage probability of confidence interval
newdata	(optional) list of data frames, one per strain, with a fine grid for dose. Only for expert users. Usually this data set is built internally.
B1	number of new sampled data sets.
B2	number of draws from the distribution of the fitted coefficients.
myapply	either lapply or mclapply (for parallel computing on Linux machines). In the latter case, use the additional argument mc.cores to specifiy the number of processes to use.
...	further arguments.

Details

Explain the model that is used to derive the lethal doses here...

Author(s)

Benjamin Hofner <[e-mail]>

See Also

methods for fitted LD models: predict.LD, summary.LD, etc.

Examples

## model survival fractions for UV radiation experiment
data("geoderm.uv")
mod.uv <- LD(value ~ time, groups = "strain", experiment = "replicate",
             dose_trafo = "sqrt", data = geoderm.uv,
             family = negbin(theta = c(0.5, 10)))

## for more examples see ?summary.LD

lethal documentation built on May 2, 2019, 4:19 p.m.