Analyze methylation microarray

Question

[youseeisme]

Hi! I have genome wide DNA methylation assays GSE145254 which was conducted using the Illumina Infinium MethylationEPIC BeadArray technology (Methyl850K chip). There are 11 psychiatric cases and 12 healthy controls samples balanced by age decade and sex. I wanna compare these samples: 11 psychiatric vs 12 control.

What is the difference between them (show by graph: clustering as example)?

Which CpG islands are more outlier (p-value 0.05)?

[Hugo]

Hi, maybe you can find bicyle (https://www.sing-group.org/bicycle/) useful, it is a bioinformatic pipeline to analyze whole genome bisulfite sequencing data. It can process data from directional (Lister) and non-directional (Cokus) bisulfite sequencing protocols, and from single-end and paired-end sequencing, and performs methylation calls for cytosines in CG and non-CG contexts (CHG and CHH).

[ATpoint]

https://bioconductor.org/packages/release/workflows/html/methylationArrayAnalysis.html

If you require more elaborate responses and please consider to show some effort towards what you tried so far and where you got stuck. Open-ended questions are generally not well-appreciated.

[JC]

Are you asking someone to do your analysis? At least offer some payment ...

[youseeisme]

I don't get any payment for this...

[JC]

Then you will need to learn how to do this. The analysis is too broad and requires expertise.

[youseeisme]

Therefore I asked. I don't know in which direction I should search. My coworkers also don't know answer on this question. One of them said: "Microarray is biased. Better if you will be using nanopore for researching methylation." If you can give me any advice. I will be regard.

Thanks

[Kevin Blighe]

To your co-worker: next generation sequencing technologies have bias, too. Everything and everyone has bias.

You posted the question here, too: [removed on July 2, 2020 as malware detected at URL]

Look, mmmmm, go HERE, click on the R script tab, and then you will be presented with code that allows you to retrieve the normalised values.

The data that is available is normalised beta values:

library(Biobase)
library(GEOquery)

# load series and platform data from GEO
gset <- getGEO("GSE145254", GSEMatrix =TRUE, getGPL=FALSE)
if (length(gset) > 1) idx <- grep("GPL23976", attr(gset, "names")) else idx <- 1
gset <- gset[[idx]]

boxplot(exprs(gset))