I am trying to use the PAA package to normalize and analyze my protoarray data for immune response profiling in plasma samples. I am attempting to do rlm normalization with respect to the IgG and anti-IgG levels as is depicted in quite a few publications. The PAA package is very non-descriptive as to how to input the control data and how that control data will be handled. In my data, each "block" of proteins has 4 IgG measurements and 4 anti-IgG measurments and there are in total 48 blocks in each protoarray. I make my elist files with the following code:
tch.elist <- loadGPR("TCH gpr files","targets.tch.txt")
tch.control.elist <- tch.elist[grep("IgG",tch.elist$genes$Name),]
Note that I checked the indexes of the controls I was attempting to input and they were indeed correct using this strategy. In fact, the elist files look pretty well put together and I don't think there is anything wrong with the way the GPR files were being read in. The tch.elist file contains all the proteins (including controls), whereas the tch.control.elist contains only the controls. I have also attempted only using the non-control proteins in the first elist to no avail. Neither the vignette nor the package manual describes exactly how these files need to be specified and how it treats the controls with respect to the data (e.g. does it go block to block and apply the control measurements in that block to the same proteins?). This is important, as I really need to ensure that this data is being handled appropriately.
Anyways, the code for running the normalization is as follows:
tch.norm <- normalizeArrays(tch.elist,method="rlm",controls.elist=tch.control.elist)
and it generates this error message:
Error in `[<-`(`*tmp*`, current.row, i, value = 1) :
subscript out of bounds
Has anyone else encountered this problem, and if so, were able to resolve it? I have not found any chat room discussion around the PAA package and really need to figure this out to move on with data analysis.