Distance geometry cannot distinguish between enantiomers. The test for the correct enantiomer can be carried out either at the substructure level or after the coordinates of all atoms have been generated. The safest procedure is to generate all atomic coordinates for both enantiomers and then to refine both enantiomers. The one with the lower refined energy will almost certainly be the correct enantiomer (see file ``dgsa.inp" below). This test is somewhat time-consuming, because one has to carry out a large number of refinement steps for the incorrect enantiomer.
Two other alternatives will work for well defined systems and appropriately defined substructures but may fail for less well defined systems. The first involves testing both mirror images of a particular substructure according to their improper energy. The enantiomer with the lower improper energy is likely to be the correct one. This test will work only when chiral centers are part of the substructure selection; the central atom as well as all of its ligand have to be included. The test also assumes that the improper angles in the parameter and topology file have been well defined. This enantiomer testing method is included in the file ``dg_sub_embed.inp" as a comment; to activate it, the user has to uncomment the appropriate lines.
The second alternative involves testing both mirror images of a particular substructure according to its least-squares fit to a known reference coordinate set. The enantiomer with the lower rms difference is very likely to be the correct one. Again, this enantiomer testing method is included in the file ``dg_sub_embed.inp" as a comment; to activate it, the user has to uncomment the appropriate lines. This feature will not work in the case of extremely underdetermined systems that exhibit very large rms differences between possible conformations.