doubt regarding normalisation when multiple groups

[Fabio Marchianó]

Hi,

I would like to perform a pairwise comparison having 5 different groups.

I defined my treatment like so c(1,2,4,3,1,3,1,3,4,3,4,3,3,4,0,0,0,0,0,2,4,0,4,0,3,2).

At the end i would like to test in a pairwise fashion each group vs the group 0. 

Before running calculateDiffMeth I'm then using reorganize to create new objs containing 1 and 0, 2 and 0, 3 and 0 and 4  and 0.

My dubt is: should I reorganize the already normalized complete dataset ( filtered.myobjDB.norm <- normalizeCoverage(filtered.myobjDB, method = "median")) or should i  reorganize  from a not normalised Object (filtered.myobjDB in my case) otherwise it will be normalised twice?

calculateDiffMeth 's effect states: method to calculate the mean methylation difference between groups using read coverage as weights (default). When set to "mean", the generic mean is applied and when set to "predicted", predicted means from the logistic regression model is used for calculating the effect.

If I got it right i should start from already normalised Object then reorganize it and then set calculateDiffMeth effect to "mean" to avoid to normalised twice.

Thank you for your suggestions,

Fabio