Hello everyone,
I've downloaded two MeDIP-Seq datasets from the same patient (Normal tissue, Cancer tissue) and I want to identify differentially methylated regions between those datasets.
I pre-processed my datasets, I made the alignment step using BWA and now I'm in the MEDIPS part. I am following this tutorial I made every single step until chapter 6. I'm know having trouble finding the way to proceed to the DMR analysis.
Let's assume that my bam file names after BWA alignment are normal_tissue.bam and cancer_tissue.bam!
If you look at the above link, chapter 6, you'll see that they are using more replicates so I can't really understand what am I supposed to do with my two BAM files. Also I can't really understand what they mean with things like: MEDIPS SET, INPUT SETS and Cset.
Can anyone make any R coding suggestions based on my BAM file names and the above tutorial (chapter 6)?
I'm totally new in the field of bioinformatics so any comment would be very helpful!
Thank you very much.
Consider copy number variations, which are quite common in cancer. I you compare MeDIP enrichment between two samples when CNVs are likely to be present then how do you distinguish an apparent difference in enrichment that's actually due to methylation differences from one due simply to a CNV? That's why you end up needing at least 3 types of samples for this sort of analysis.
Thank you for your answer, but could you be more specific? what do you mean by 3 types of samples? My datasets are MeDIP-Seq from both Normal and Cancer tissue and RNA-Seq from the same parts.
I found some examples online of people using MEDIPS with only two samples so I thought I could do that too. Do you think I can use another tool to make my DMR?
Thank you!
As long as you're OK with possible confounders due to CNVs then you can go ahead and proceed. I haven't used MEDIPs in years, so I don't recall whether it strictly requires replicates or not. Of course, without replicates any method will be highly questionable.