Slots for MultiAssayExperiment

[eleonore-schneeg]

Hi,

I wanted to know if you have any means of adding slots for each experiments?
I would want for instance have slots for raw, normalised, batch corrected data for each of my experiments.
I tried to generate a list of lists object to build the MultiAssayExperiment

objlist <- list("RNAseq" =list(raw=raw, normalised=normalised) , "TMT proteomics" = list(raw=raw, normalised=normalised) )
myMultiAssay <- MultiAssayExperiment(objlist, patient.data, dfmap)

But this returns an error:

Error in validObject(.Object) :
invalid class "ExperimentList" object: 1:
'list' class is not supported, use a rectangular class
invalid class "ExperimentList" object: 2:
'list' class is not supported, use a rectangular class

Do you support slots in the MultiAssayExperiment package or how could I implement this?

Thanks!

Eléonore

[LiNk-NY]

Hi Eléonore, @eleonore-schneeg

You could use one SummarizedExperiment for each experiment to represent the raw and normalized data.

library(SummarizedExperiment)
ExpA <- SummarizedExperiment(assays = SimpleList(raw = matrix(1:4, nrow = 2, dimnames = list(NULL, c("A", "B"))), norm = matrix(5:8, nrow = 2)))
ExpB <- SummarizedExperiment(assays = SimpleList(raw = matrix(9:12, nrow = 2, dimnames = list(NULL, c("A", "B"))), norm = matrix(13:16, nrow = 2)))
MultiAssayExperiment(list(ExpA = ExpA, ExpB = ExpB))

Though your mileage will vary depending on what you want to do.
You may also want to consider having two MultiAssayExperiments, one for each data type, or one MultiAssayExperiment with four 'assays'.

Best regards,
Marcel

[eleonore-schneeg]

Hi Marcel, @LiNk-NY

Thanks for your reply.

The summarisedexperiment solution is what I'm looking for, although it requires elements in the list to have the same dimensions.
Is there any way to bypass this? I want to have different slots with filtered data, so this makes me having matrices of different sizes..

Thanks!

[LiNk-NY]

Hi Eléonore @eleonore-schneeg

There is no way to bypass this as the class was designed to hold transformations of the same data.
You could take the last option I mentioned and create separate SummarizedExperiments for all matrices, i.e., four 'assays'. That's provided that you want to coordinate all these datasets or you could do a pair of MultiAssayExperiments, one for each data type.

Best,
Marcel