Differential methylation analysis and methylation clustering with RnBeads

[colaneri]

Hi, I want to use RnBeads to run a differential methylation analysis to compare cases and controls. Basically I have 12 cases and 11 controls. I have two potential confounding variables that I will like to include in the analysis (Age and Sex). My sample_annotation table looks like this:


Sample_Name
Sentrix_ID
Sentrix_Position
Sample_Plate
Sample_Well
Sample_Group
Age
Sex

The data type is infinium methylation data (450)
The goals for this analysis are:
1) Find differentially methylated sites among cases and control
2) Make clustering analysis to see if there are signatures of methylation that separe cases of controls

I noticed that before running:

rnb.run.analysis(dir.reports=report.dir, sample.sheet=sample.annotation, data.dir=idat.dir, data.type="infinium.idat.dir")


You may set up global options with rnb.options()

Will be me options look like this?:


rnb.options(filtering.sex.chromosomes.removal=TRUE, identifiers.column="Sample_ID", differential.comparison.columns = "Sample_Group", covariate.adjustment.columns = c("Age”, “Sex”))



1) But what about clustering?

Does the rnb.run.analysis function first create and RnBSet object and then I have to run clustering with this object? Or the clustering is part of the whole analysis performed by the rnb.run.analysis() function?


2) What about gender prediction?
Does gender prediction run when I use the rnb.run.analysis function with the argument explicated above?


I will appreciate if someone can answer some of this question.
ALe

[daanum]

Clustering samples of RRBS data

Hi,

I would like to see if the biological replicates of my two groups fall together. Is there a simple tool to check the clustering of samples. i am working with methylation data using RRBS technique.

Please reply asap if you know the answer.

Thanks in advance

Cheers

[dpryan]

Hi,

I would like to see if the biological replicates of my two groups fall together. Is there a simple tool to check the clustering of samples. i am working with methylation data using RRBS technique.

Please reply asap if you know the answer.

Thanks in advance

Cheers

If you make bigWig files out of the methylation tracks then you could use deepTools (either plotCorrelation or plotPCA) to look at sample clustering.