methSeg output has some missing segment groups

[Ezel Ergüden]

Hi!

I am trying to segment a genome based on the methylation profile, and I am using methSeg() for this purpose. While it works fine on most samples I did notice that in a few of the samples it is skipping some segment groups, so while it identifies 5 segment groups, actually in the data only 4 are present. I did try running the analysis again and got the same result. There are no segments labeled as belonging to the 4th segment group in the bed file either. 

Appreciate any help or ideas regarding what might be causing this, thanks!

[alex....@gmail.com]

Hi Ezel, 

This is hard to diagnose without seeing the data. 

I would check with `diagnostic.plot` enabled and see if you can get any additional information like the expected segment width or methylation level for the missing segment, maybe they are on the edge of the distributions and very rare.
For confirming the missingness, did you confirm via table(res$seg.group) where res = methSeg(obj)? It is possible that there is only a single chromosome that features the missing segments (usually the mitochondrial chromosome). You could check via table(res$seg.group, res$seqnames) if the missing regions are indeed in one specific chromosome.

I hope this helps.

Best,

Alex