R help - Jan 2010 - Estimate Slope from Boltzmann Model (package: DRC)

Dear R Community,

I am using the package DRC ( to fit a boltzman model to my data.  I
can fit the model and extract the lower limit, upper limit, and ED50
(aka V50), but I cannot figure out how to get the slope of the curve
at ED50.  Is there a simple way to do this?  I've searched the mailing
list and looked through the package documentation, but could not find
anything.  I am new to r, and especially the DRC package, so any help
would be appreciated.

This gets me 3/4 of what I need:

DATA:
Dose,Response
0.85,0.0085
0.90,0.0096
0.95,0.0155
1.00,0.2725
1.05,0.3634
1.10,0.3508
1.15,0.5002
1.20,0.8826
1.25,0.7023
1.30,1.7934
1.35,0.7389

CODE:
library( drc )
mydata <- read.csv( file="test.csv", header=T )
drc.preTBS <- drm( response ~ dose, data = mydata, fct= LL.4( names c(
"Slope", "Lower Limit", "Upper Limit",
"ED50") ) )

For the data above, I would expect a slope at ED50 to be about 0.0714
(according to GraphPad).  How can I determine the slope at ED50 for a
Boltzmann fit using DRC or related packages?

Thanks much!
-Bob (Veterans Administration)

Hi Bob,

you fitted a log-logistic model specifying the model function LL.4().

If you want to fit a Boltzmann or logistic model then you can use the model
function L.4()
in place of LL.4():


drc.preTBS.2 <- drm( Response ~ Dose, data = mydata, fct= L.4( names c(
"Slope", "Lower Limit", "Upper Limit",
"ED50") ) )

summary(drc.preTBS.2)


The slope is get from the above fit is -13.7329718. The "slope" in
GraphPad is then
obtained using the following transformation:


-1 / (-13.7329718) = 0.07281745


which is close to the number you report. So, I would say that the two fits
(obtained using
GraphPad and using "drc"/R) are the same.

Depending on the software used, different parameterizations of the logistic
model are
used. So even though the parameters unequivocally have the same names in various
software
programmes they need not necessarily be identical!


Christian