Size of combined BSSeq objects

[kperzel]

Hi,

I'm working with a large number of samples that were processed in different batches so I built two BSSeq objects with the same loci, which turned out to be 3GB and 6GB, and went to combine them using combine() as the documentation says. This led to an object of 46GB, which is difficult to work with. I tried instead to cbind() the two objects and this created an object of 9GB. As far as I can tell, these two objects contain all the same information, what is the reason for the major size difference?

Thank you

[PeteHaitch]

combine() has some internal inefficiencies compared to cbind() (although, of course, cbind() is only appropriate if the two objects contain the exact same loci in the exact same order).
It's on the list of things to fix.

In the interim, you can run realize(combined_BSseq, NULL) to simplify a BSseq object which will reduce its memory footprint.