Generating D-amino acids and other NCAA alpha-amino acids through covalent modification for RoseTTAFold-All-Atom

[tsamdin]

If I want to predict the structure of a protein that contains both L and D amino acids, could I submit a sequence where the D-residues are given as glycine, with each glycine bearing the appropriate covalent modification and change in stereochemistry to give the D side chain? I think I would just have to provide each sidechain as the appropriate sdf file? Would this approach using glycine as base work for other NCAA alpha-amino acids with L or D stereochemistry?