Volume 1 Chapter 7 Atomic Overlap in Multiple Fragment Situations

Back to Table of Contents

7.13 Handling Structural Inversion

The molecular structures stored in the CSD represent one enantiomorph (one 'hand') of the molecule in question. However, some 80% of these structures crystallise in space groups in which molecules of both `hands' are present. For the other 20% only one enantiomer may exist in the crystal structure.

Many of the geometrical parameters described in earlier sections will, of course, have identical values for both enantiomorphs. This includes all distances between two points, and all intervector angles. However, parameters such as torsion angles and ring puckering parameters will change sign when recalculated for the mirror-image molecule in moving from one fragment to its inverse.

Thus, for 3D searches which test enantiomorph-sensitive parameters, it is quite possible for the first-located fragment to fail the search criteria, but for these criteria to be satisfied by the inverse fragment (the enantiomorph) in which the sign of each atomic coordinate has been changed.

QUEST3D contains facilities which permit the examination of both enantiomorphs in these situations. Three settings are provided:

Now, it might appear that 'normal' should be the appropriate setting for use in 3D searches involving enantiomorph-sensitive parameters such as torsion angles. However, this makes the assumption that the enantiomorph represented by the crystallographic x,y,z-coordinates actually corresponds to the actual chemical enantiomorph under study! There is an assumption that the absolute configuration depicted by the crystal structure analysis is always correct.

The crystallographer can only ensure that this is true by either:

In general, most crystallographers will try to ensure the accuracy of their published absolute stereochemistries in organic compounds. Indeed, it is very very rare to find that the experiment at (a) above has been carried out incorrectly.

However, for some structure types (e.g. steroids, carbohydrates, etc.) where the absolute stereochemistry is well known and is not in doubt, it is quite possible for the incorrect enantiomorph to be assigned and refined perfectly well in the determined space group, especially where the axial system describing the crystal structure has inadvertently been chosen to be left- rather than right-handed.

This issue is complex, and the CCDC is now engaged in a project to identify (as far as possible) those cases of erroneous enantiomorph assignments. For a test set of over 200 pyranose carbohydrates, about a dozen potential errors have been flagged so far.

For these reasons, the 'normal' command should, perhaps, be used only when the dataset has been thoroughly examined.

Back to Table of Contents

Volume 1 Chapter 7 Searching for Non-Bonded Contacts.