Analyzing methylation in geo-data

Question

[Katrien]

Firstly, I'm relatively new to R and methylation data.

I'd like to analyze this geo dataset (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE21232). It is performed with Human Methylation27 DNA BeadChip arrays. The data only provide beta-values (no A- and B-values). I know how to perform the analysis with lumi once I have a MethyLumiM object in R. But I'm not able to get there.

I followed this manual about GEOquery (https://bioconductor.org/packages/release/bioc/vignettes/GEOquery/inst/doc/GEOquery.html) and was able to create an ExpressionSet (without the annotation library). However, I do not know how to proceed.

Any help will be greatly appreciated. Thank you!

[ddiez]

I can't check with an example at the moment but from the methylumi package documentation for MethyLumiM:

Objects can be created by calls of the form new("MethyLumiM", exprs, methylated, unmethylated, detection, methylated.N, unmethylated.N, ..., assayData). The "exprs" is a matrix of M-values, which is the log2 ratio of methylated and unmethylated probe intensities; "methylated" and "unmethylated" are intensity matrix measured by methylated and unmethylalted probes of Illumina Infinium methylation microarray; "detection" is the detection p-value outputted by Illumina GenomeStudio software; "methylated.N" and "unmethylated.N" are bead numbers for methylated and unmethylalted probes. "exprs", "methylated" and "unmethylated" information are required for MethyLumiM class. When creating a new MethyLumiM object, the information of "exprs", "methylated", "unmethylated" and "detection" can also be provided directly through "assayData".

This suggests various options to obtain the desired object. I would first try coercing with as, which probably won't work. Another is to create a new object with new as in the quoted text. You can give the objects in your ExpressionSet that match (like the matrix, phenoData) and try to provided the ones not included.

Answer 1

[ddiez]

Following my comment I tested this code that seems to work:

library(GEOquery)

x <- getGEO("GSE21232")[[1]]
x
ExpressionSet (storageMode: lockedEnvironment)
assayData: 27578 features, 16 samples 
  element names: exprs 
protocolData: none
phenoData
  sampleNames: GSM530883 GSM530884 ... GSM530898 (16 total)
  varLabels: title geo_accession ... data_row_count (35 total)
  varMetadata: labelDescription
featureData
  featureNames: cg00000292 cg00002426 ... cg27665659 (27578 total)
  fvarLabels: ID Name ... ORF (38 total)
  fvarMetadata: Column Description labelDescription
experimentData: use 'experimentData(object)'
Annotation: GPL8490 

library(methylumi)
y <- as(x, "MethyLumiM")
y
MethyLumiM (storageMode: lockedEnvironment)
assayData: 27578 features, 16 samples 
  element names: exprs 
protocolData: none
phenoData
  sampleNames: GSM530883 GSM530884 ... GSM530898 (16 total)
  varLabels: title geo_accession ... data_row_count (35 total)
  varMetadata: labelDescription
featureData
  featureNames: cg00000292 cg00002426 ... cg27665659 (27578 total)
  fvarLabels: ID Name ... ORF (38 total)
  fvarMetadata: Column Description labelDescription
experimentData: use 'experimentData(object)'
Annotation: GPL8490

I am not familiar with this package and not sure whether this is all you need to work with it.