I am trying to set a model using the pcrfit() function of the R's qpcR
package. I have defined a dataframe with 45 cycles (rows) and 4
readings (columns) derived from a TaqMan qPCR with relative
quantification. The data is based on the normalized fluorescent
results (main fluorescence minus ROX).
The example reported herein has a couple of readings with abnormal
amplification (which would be the usual kind of results for the PCR I
am carrying out) and two with a proper amplification; the plot section
of the example is meant to describe the data itself.
For each implementation, the error returned is:
> Error in model.frame.default(formula = ~Fluo + Cycles, data = DATA,
weights = WEIGHTS, :
variable lengths differ (found for '(do.optim)')
However each column has 45 rows.
Any tips on how to problem-solve this issue?
Many thanks,
Luigi
PS:
The following example writes a file in the current folder, uses it for
the pcrfit modelling and then removes it.
>>>
# qpcR
if(!is.element('qpcR', installed.packages()[,1]))
{
install.packages('qpcR')
} else {
library(qpcR)
}
# create dataframe
col.0
While removing "do.optim=TRUE" and "robust=FALSE" options everything works fine. Where does these options come from ? They are not in the qpcR library, neither in monpack.lm...
Dear Denis,
thank you for your answer and sorry for the delay in my reply.
The options were derived from the internet, but I can't recall from where I got them. The error here was in the first place to call the cycle column 'col.0' rather than 'Cycles'. Anyway I still obtaining errors for some amplification data, essentially that the FTC function was not found, that I guess is linked to a sort of "rainbow" of the data. When the shape of the data is not sigmoidal the function does not work properly and I need to bypass this in order to proceeed to the computtion of all wells.
Best regards,
Luigi
Dear Denis,
I actually kind of debugged the problem using the modlist() function rather than the pcrfit(), following a suggestion of Andrej-Nikolai Spiess. Anyway, as you said, pcrfit worked good removing the arguments "do.optim=TRUE" and "robust=FALSE".
For the sake of argument, the code I have now is:
# qpcR
if(!is.element('qpcR', installed.packages()[,1]))
{
install.packages('qpcR')
} else {
library(qpcR)
}
(# create dataframe
Cycles
- Badges
- Science method