3.3.2 Molecular Geometry
As structural data was aggregated, tables of derived data could be formed. In the late
1980s, there was an enormous effort by Orpen et al. that produced tables of precise
bond lengths for organic [70] and organometallic [71] compounds. These were
updated in 2006 in the International Tables for Crystallography [140, 141]. The
use of these tables to check geometry in newly determined structures was a laborious
process and still involved manually looking up the data for individual bonds.
Leading on from this in 2004, a derived knowledgebase of geometric parameters
was developed by the CCDC called Mogul [142], which facilitated the automatic
retrieval of bond lengths, angles and torsions and was further enhanced by the
addition of data on ring conformations [143]. Mogul is now widely used in checking
new small molecule structures for unusual geometry by comparison with distributions of known geometry for similar fragments derived from data in the CSD. A
similar process is also embedded in the PDB deposition pipeline to check the
geometry of ligands in protein structures. Utilising this data can be taken further in
the application of a knowledge-based conformer generator [144, 145] which allows
the minimisation of molecular conformations and the generation of conformer subsets based on CSD data.
3.3.3 Non-bonded Intermolecular Interactions
The first computerised library of intermolecular non-bonded interactions was collated in Isostar [146]. This reveals the geometric preferences of hydrogen bonds and
other classical directional non-bonded interactions. Since 2009 further interactions
have been added, such as C-I groups, due to a significant increase in interest in
non-classical intermolecular interactions, such as sigma-hole interactions [147–
151]. It is important that a knowledgebase can potentially grow and the value of a
greatly increased quantity of crystal structure data is demonstrated in Fig. 9, where it
is clear that more data points provide a considerably more conclusive understanding
of preferred directionality.
The most significant evolution from Isostar was into Full Interaction Maps [152],
which allow the concurrent visualisation of multiple different non-covalent interactions. Full Interaction Maps use Isostar data to construct contour density maps
around a whole molecule, as seen in Fig. 10. These maps can be used to evaluate
the stability of polymorphic structures, assess multiple types of non-covalent interactions simultaneously and provide a platform for understanding crystal
morphology. A comparison of how well the hydrogen bonds of a given crystal
form coincide with the likely interactions highlighted by Full Interaction Maps can
indicate the stability of the form. This approach is now used by pharmaceutical
companies and the CCDC to help determine the stability of a new drug form and the
risk of a more stable polymorph being found with different physical properties
[153, 154].
Leading Edge Chemical Crystallography Service Provision and Its Impact on. . .
115
As structural data was aggregated, tables of derived data could be formed. In the late
1980s, there was an enormous effort by Orpen et al. that produced tables of precise
bond lengths for organic [70] and organometallic [71] compounds. These were
updated in 2006 in the International Tables for Crystallography [140, 141]. The
use of these tables to check geometry in newly determined structures was a laborious
process and still involved manually looking up the data for individual bonds.
Leading on from this in 2004, a derived knowledgebase of geometric parameters
was developed by the CCDC called Mogul [142], which facilitated the automatic
retrieval of bond lengths, angles and torsions and was further enhanced by the
addition of data on ring conformations [143]. Mogul is now widely used in checking
new small molecule structures for unusual geometry by comparison with distributions of known geometry for similar fragments derived from data in the CSD. A
similar process is also embedded in the PDB deposition pipeline to check the
geometry of ligands in protein structures. Utilising this data can be taken further in
the application of a knowledge-based conformer generator [144, 145] which allows
the minimisation of molecular conformations and the generation of conformer subsets based on CSD data.
3.3.3 Non-bonded Intermolecular Interactions
The first computerised library of intermolecular non-bonded interactions was collated in Isostar [146]. This reveals the geometric preferences of hydrogen bonds and
other classical directional non-bonded interactions. Since 2009 further interactions
have been added, such as C-I groups, due to a significant increase in interest in
non-classical intermolecular interactions, such as sigma-hole interactions [147–
151]. It is important that a knowledgebase can potentially grow and the value of a
greatly increased quantity of crystal structure data is demonstrated in Fig. 9, where it
is clear that more data points provide a considerably more conclusive understanding
of preferred directionality.
The most significant evolution from Isostar was into Full Interaction Maps [152],
which allow the concurrent visualisation of multiple different non-covalent interactions. Full Interaction Maps use Isostar data to construct contour density maps
around a whole molecule, as seen in Fig. 10. These maps can be used to evaluate
the stability of polymorphic structures, assess multiple types of non-covalent interactions simultaneously and provide a platform for understanding crystal
morphology. A comparison of how well the hydrogen bonds of a given crystal
form coincide with the likely interactions highlighted by Full Interaction Maps can
indicate the stability of the form. This approach is now used by pharmaceutical
companies and the CCDC to help determine the stability of a new drug form and the
risk of a more stable polymorph being found with different physical properties
[153, 154].
Leading Edge Chemical Crystallography Service Provision and Its Impact on. . .
115