tcplAddModel: Draw a tcpl Model onto an existing plot

In tcpl: ToxCast Data Analysis Pipeline

Draw a tcpl Model onto an existing plot

Description

tcplAddModel draws a a line for one of the tcpl Models (see Models for more information) onto an existing plot.

Usage

tcplAddModel(pars, modl = NULL, adj = NULL, ...)

Arguments

pars	List of parameters from level 4 or 5 output
modl	Character of length 1, the model to plot: 'cnst,' 'hill,' or 'gnls'
adj	Numeric of length 1, an adjustment factor, see details for more information
...	Additional arguments passed to curve

Details

tcplAddModel draws the model line assuming the x-axis represents log base 10 concentration.

If modl is NULL, the function checks pars$modl and will return an error if pars$modl is also NULL.

adj is intended to scale the models, so that models with different response units can be visualized on a single plot. The recommended value for adl is 1/(3*bmad) for level 4 data and 1/coff for level 5 data. If adj is NULL the function will check pars$adj and set adj to 1 if pars$adj is also NULL.

See Also

Models, tcplPlotFits

Examples

## Create some dummy data to plot
logc <- 1:10
r1 <- sapply(logc, tcplHillVal, ga = 5, tp = 50, gw = 0.5)
r2 <- log2(sapply(logc, tcplHillVal, ga = 4, tp = 30, gw = 0.5))
p1 <- tcplFit(logc = logc, resp = r1, bmad = 10)
p2 <- tcplFit(logc = logc, resp = r2, bmad = log2(1.5))

## In the dummy data above, the two plots are on very different scales
plot(r1 ~ logc, pch = 16, ylab = "raw response")
tcplAddModel(pars = p1, modl = "hill")
points(r2 ~ logc)
tcplAddModel(pars = p2, modl = "hill", lty = "dashed")

## To visualize the two curves on the same plot for comparison, we can 
## scale the values to the bmad, such that a scaled response of 1 will equal
## the bmad for each curve.
plot(r1/10 ~ logc, pch = 16, ylab = "scaled response")
tcplAddModel(pars = p1, modl = "hill", adj = 1/10)
points(r2/log2(5) ~ logc)
tcplAddModel(pars = p2, modl = "hill", adj = 1/log2(5), lty = "dashed")

tcpl documentation built on Oct. 10, 2024, 1:07 a.m.