Manual creation of differential methylation data tables, RnBeads

Question

[adrianstankiewicz85]

Dear all,

I would like to manually extract differential methylation data tables from RnBDiffMeth object. By the "data table" I mean a table in which I have both genomic region annotations and corresponding methylation statistics (such tables are normally attached in report generated by rnb.run.differential function).

I have created RnBDiffMeth object using command:

regions <- rnb.execute.computeDiffMeth(
x = Inference_result$rnb.set,
pheno.cols = c("i_vs_c", "b_vs_rest", "b_vs_c", "i_vs_rest", "e_vs_c"),
region.types = c("tiling", "cpgislands", "genes", "promoters", "tiling200bp")
columns.adj = rnb.getOption("covariate.adjustment.columns"),
adjust.sva = rnb.getOption("differential.adjustment.sva"),
pheno.cols.adjust.sva = rnb.getOption("inference.targets.sva"))

To extract data tables from this object, I have tried functions exportDMRs2regionFile (which does not generate statistical data) and get.table (which, as far as I see, has only "orphaned" statistics, with no key which would allow me to match these statistics to corresponding genomic region). Is there a way to annotate get.table with genomic data?

I am doing this as a way to learn the RnBeads package to find a solution or workaround for another issue Im having.

Best, Adrian

Answer 1

[fmueller]

Hi Adrian, the get.table() method returns the differential statistics you are looking for. If you want to add the region annotation data, you can extract this from your RnBSet object using

regionAnnot <- annotation(rnb.set)

This will return a table with exactly the same row ordering as get.table() if you use the corresponding RnBSet object, that you used for differential methylation computation.

Hope that helps.

Best, Fabian