No abstract

A methodology is described for analysing the Cambridge Structural Database (CSD) in terms of molecular conformations. Molecular species that have more than a single occurrence across the complete CSD are identified, either as the sole crystal component or co-crystallized with other components. Cluster analysis, based on a root-mean-square fit of coordinates and chemical connectivity, is performed to identify conformational variance for each molecule. Results are analysed in terms of the number of discrete conformations observed versus the number of crystal environments and number of acyclic torsion angles in the molecule. Special subsets of environments are also analysed, namely polymorphs, co-crystals and solvates. In general, conformational diversity increases with an increasing number of different crystal environments and with an increasing number of flexible torsion angles. Overall, molecules with one or more acyclic flexible torsion angle are observed to exist in more than one conformation in ca 40% of cases. There is evidence that solvated molecules exhibit more conformational flexibility on average, compared with polymorphs and co-crystals.

A method is described which enables automated superimposition of molecular conformations by the matching of chemical graphs, allowing for topological symmetry in the molecular structure. This algorithm is implemented in the program Tormat. The implemented method makes allowance for enantiomer inversion, explicit and implicit treatment of H atoms, and partial structure alignment.