Merging HETATM and ATOM entries into one DataFrame

Following up on the comment by @wojdyr in #52

> by the way, having atoms in two separate frames is rather not a good idea.

> At first glance it may look like the protein chains are all ATOM, but wwPDB uses different criterium:

> only natural amino-acids (and nucleic acids) are marked as ATOM, and the modified ones are > > > marked as HETATM.
> So MET is ATOM but MSE is HETATM.

> If you keep them both separately such an example:

ppdb.df['ATOM']['b_factor'].plot(kind='hist')

> won't work as expected - it may skip some residues

That's a good point and I haven't thought of that! The reason why I kept these separate is that I am mostly working on cases where HETATMs refer to non-protein residues. The HETATM--MSE issue should definitely be addressed somehow and I would have to think about it more ... Suggestions would be welcome.

Review the existing handling of separate ATOM and HETATM DataFrames, using the MSE example and the non-protein HETATM cases described here to define the desired behavior. The issue names no files or tests; done should ensure modified residues such as MSE are not omitted when atom data is used together.

Written by the indexing model from the issue text.