Defining FM intervals with unsampled regions

[Franz Goerlich]

Hello all, 

I am looking at the liquid hexane example. I want to test FM with a mapping where the four central carbon atoms are retained. Now I'd like to define the min/max in which the splines for the dihedral are fit, but the problem is that there are large unsampled regions (see figure from Voth et al attached). How would I circumvent this? Is it possible to only fit the splines in certain intervals?

I believe @Cecília Álvares asked a similar question about this a while ago.

Best,

Franz