Calculated percentages can be wrong

[SPPearce]

When the levels in the group factor don't appear in the sample table in the same order, then the calculated percentages are wrong.

This can be fixed by replacing line 393 in analysis.R:

meth.per <- t(t(meth.freq)/as.vector(table(samp$group)))

with:

meth.per <- t(t(meth.freq)/as.vector(table(samp$group)[colnames(meth.freq)]))

I don't know if there is anywhere else with the same issue.
I've created a pull request for this point #10