How to apply preprocessing methods from minfi package on an output file from Genome studio

Question

[halldorbjarni]

Hi guys,

I have an output file from Genome studio that has not been normalized. I want to apply the functional normalization method found in the minfi package to my methylation data but I don't have access to the IDAT files only the Genome studio output file. I found an internal method in the minfi package that reads the data from Genome studio in (read.GenomeStudio) but that doesn't result in a rgSet which the functional normalization method requires to work. Is there some critical information in the IDAT files that makes this impossible or can I somehow cast my GenomeStudio data to an rgSet?

TL;DR is there any way to apply functional normalization from the minfi package to a Genome studio output file from Illumina 450k?

[igor]

What type of object does read.GenomeStudio return?

I think preprocessIllumina produces an output that's equivalent of Genome Studio output (downstream of RGChannelSet). For example, RGChannelSet contains red and green channels, which are lost in all downstream objects.

[halldorbjarni]

Looking at the code it only returns a list of the three matrices: signalA, signalB and beta. And yeah exactly I'm not sure if I can use the functional normalization since I'm pretty sure the control probe information is lost after the GenomeStudio output file is created.

[fortin946]

You are correct, the control probes are lost with the Genome Studio preprocessing and are needed for the functional normalization algorithm; I believe there is a way to get those probes from the Genome Studio software, but that might be hard and I'm not sure you want to adapt the code to that output. Probably easier to get the original IDAT files form your core. Then you could fully use all of the minfi functionalities.

Hope this helps,

JP