how to handle missing values

[cbaakman]

Currently reported types of missing values:

• atom types that are not in the forcefield, in which case there's no charge values/vanderwaals parameters
• conservation scores that are NaN, because the protein wasn't aligned at that particular position

Possibilities about how to fix this:
solution 1: skip PDB entries with missing values
solution 2: remove the residues/atoms with missing values from the PDB file, after it's loaded in.

[cbaakman]

If variant is mapped to multiple PDBs, we can choose the one with no NaNs. However, this information isn't known until the script starts trying to preprocess the PDBs. So we need a more flexible script that can cancel preprocessings during the run.