Since these are a
Curves+-specific questions, you may have a better chance to get them answered from the Curves+ user community.
I understand that each MD package has a unique native format. However, each MD package should (I believe) have a method to convert its native format to the standard NMR-like MODEL/ENDMDL ensemble format, as in the PDB.
As far as 3DNA goes, the
find_pair -curve+ option handles a single structure in PDB format (e.g.,
find_pair -curve+ 355d.pdb). It aims to simplify the preparation step to the Curves+ program.
HTH,
Xiang-Jun