Residue-residue to atom-atom mapping

The result of the DCA analysis is a list of nucleotide contacts. These need to be mapped to atom-atom contacts to be usable as constraints in Rosetta. Two options are implemented.

P-Only mapping: pOnly

A simple variant is to take the P atom of each nucleotide backbone and add these as atom-atom constraints.

Complex atom-atom mapping using a contact database: minAtom

This method relies on pre-built contact lists included in the prediction package. For each type of nucleotide contact (A-A, A-C, A-G, A-U, ...) a list of representatives is stored, containing the PDB code of an entry with known native structure and the numbers of the residues. From these lists a database is built containing all possible atom-to-atom combinations for the two residues, as well as their distances extracted from the crystal structures. Next, to reduce this database to the relevant atoms involved in forming a contact, a second database is created, containing only those atom-to-atom contacts that satisfy a mean and standard deviation condition. The defaults for those are set to lower than 6Å and 3Å respectively. DCA contacts such as “A-U” are then looked up in the mean database, and all atom-to-atom combinations found are added as constraints.