Hello all,

I need to ask two questions. I am new to this area. I have my epic array data from 4 different cells with 2 replicates. I have to use Minfi package. I have loaded my idat data into the environment. I checked tens of workflows and all of them contained sample sheet. There is no sample sheet for my data. Should I create it somehow or should I start analyzing without the sample sheet? If I should create, how can I do it? And the second question is how will I handle with the replicates? I will continue with bump hunter and Limma for differential methylation. At what point the replicates should be merged, if they should be merged? Thanks in advance

> RGset
class: RGChannelSet 
dim: 1105209 8 
metadata(0):
assays(2): Green Red
rownames(1105209): 1600157 1600179 ... 99810982 99810990
rowData names(0):
colnames(8): 207716530101_R01C01 207716530101_R02C01 ... 207716530101_R07C01
  207716530101_R08C01
colData names(14): X Date ... Basename filenames
Annotation
  array: IlluminaHumanMethylationEPIC
  annotation: ilm10b4.hg19

You don't need a sample sheet. It is just a simple way to identify the Idat files and possibly include sample information. But you appear to have read the data in already, so you are good to go.

It is not clear what you mean by replicates. Are these technical replicates or something more like a biological replicate? Ideally you would have biological replicates so your results would be more likely to generalize to the underlying population, but perhaps you have replicates that are simply different plates of cells? In that case your replicates are more like technical replicates and you could either collapse to a single value by taking the mean of the M-values for each replicate (using e.g., avereps from limma), or perhaps by fitting a GLS model by specifying the within-replicate correlation (using duplicateCorrelation).