preprocessSWAN fails to normalize RGChannelSet

[alexandernassar24]

Hello,

I am having issues normalizing an RGChannelSet. I initialize the object with two table that have probes as rownames and methylated/unmethylated signals for each sample as column names:

#create minfi object
RGChannelSet <- RGChannelSet(Green = green_methylated, Red = red_unmethylated)

#specify annotation
RGChannelSet@annotation <- c(array = "IlluminaHumanMethylation450k", annotation = "ilmn12.hg19")

RGChannelSet

class: RGChannelSet dim: 485512 74 metadata(0): assays(2): Green Red rownames(485512): cg00050873 cg00212031 ... ch.22.47579720R ch.22.48274842R rowData names(0): colnames(74): 6929689068_R01C01 6929689068_R04C02 ... 9296930148_R02C01 9296930138_R06C01 colData names(0): Annotation array: IlluminaHumanMethylation450k annotation: ilmn12.hg19

Running the preprocessingRAW function yields an empty matrix. Running preprocessSWAN on the dataset yields the following errors:

Warning: no non-missing arguments to min; returning Inf Error in sample.int(length(x), size, replace, prob) : vector size cannot be infinite

I'm not sure why the preprocessing is failing. Is it because of a mistake on my behalf in initializing the object? Guidance would be appreciated.

Thank you, Alexander

P.S. As a starting point, I am using the non-normalized table (merged for methylated/unmethylated signals for all samples) in the supplementary files of this GEO accession page: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE52556