calculateDiffMeth-related problem

[Mengjiao Li]

Hi!

Thanks for the wonderful package of methylation analysis, I have a question about the logistic model used in differential methylation analysis.

As you have described in the manual,  'In any given cell, any given base are either methylated or not. Therefore, looking at many cells should yield a similar pattern where we see lots of locations with high methylation and lots of locations with low methylation.'

This is true but if my sample is a mixture of different cell types, like a tissue sample, I just find the %methylation per base would not have an obvious bimodal distribution as your example data showed in manual. It would have much less bases with low methylation, like the picture I attached.

So is this data still suitable to analyze DiffMeth by the logistic model? Is the model only used for the bimodal distribution?

Looking forward to your reply!

Best,

Li.

[Altuna Akalin]

You can use it for this data as well. bimodality is more a like QC check. 

Best,

Altuna

--
You received this message because you are subscribed to the Google Groups "methylkit_discussion" group.
To unsubscribe from this group and stop receiving emails from it, send an email to methylkit_discus...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/methylkit_discussion/594844fd-0071-4d2f-af47-a31e65f065ccn%40googlegroups.com.