A little off topic, but the lab recommend methylpy -- https://bitbucket.org/schultzmattd/methylpy/wiki/Vignettes. I used to use Bismark but switch over to this program. One perk of methylpy is it allows you to incorporate a non-conversion control -- like lambda DNA -- which is used downstream for a per site binomial test to call sites as methylated or un-methylated. This is built into the program. The output is very similar to Bismark, so any parsing scripts you have would be easily adaptable. There is also a differentially methylated finder (DMR) tool too. I hope this helps.