engineering and spurred activity investigating molecular packing in the solid state.
Packing similarity [120] became an important factor to assess whether different
crystal structures could be polymorphic. The XPac [121] approach went further to
assess the degree to which crystal structures formed by different compounds might
be closely related. Implemented in Mercury around the same time, COMPACK
[122] looks at the similarity over multiple molecules when comparing crystal
packing. The related packing feature search [120] in Mercury will retrieve structures
with a similar crystal packing feature to that selected, such as a hydrogen-bonded
dimer. Other approaches of comparing crystal packing have developed, CrystalCMP
[123] employs differences in molecular rotation and relative positions, and Salbego
et al. proposed an index for ‘supramolecular clusters’ [124].
As computers have developed, the types of analysis scientists would like to do
have become more complex. There has been a trend towards informatics with the
rise of ‘big data’ and using computer programming to directly search and analyse the
database. The CSD Python API (Application Programming Interface) [14] has
enabled this type of query; with much of the CSD functionality now available
through fast computer algorithms, the knowledge has become easier to discover.
This means that it is possible to use most of the database searching functionality in a
machine-to-machine mode, allowing for scripting and automated workflow enactment and providing the potential to power machine learning and big data
approaches.
3.2.3 Web Tools
Technological advances and the proliferation of the Internet led to online access to
the database through WebCSD in 2010 [125], which gave easy access to the CSD via
a website without needing to install software locally on the user’s computer. This has
been superseded by a second version that reflects the changes in web technologies.
The ICSD [126], PDB [127] and COD (Crystallography Open Database [68]) [128]
databases provide a similar level of text and bibliographic searching via a web
interface. The distribution of the CSD and associated tools has also been transformed
by web developments, which have been vital as it has grown in size – in 2009 the
CSD was distributed on a CD, whereas today it is installed by download from the
Internet.
CrystalWorks [129] is a web-based search tool that takes advantage of the fact
that the different chemical crystallography data sources often need to be searched in
the same way. By indexing and aggregating the CSD, ICSD and COD, it enables a
simultaneous metadata search across them all at a high level, e.g. bibliographic,
composition or unit cell queries. It should be noted that the full CrystalWorks is
currently only available to UK academics through the Physical Sciences Datascience Service [130]; however, there is now a developing interface for all that
provides access to open (small molecule) crystallographic data. Conversely, while
there are a lot of similarities in how different crystallographic databases are searched,
there are also disciplinary differences, for instance, in the cases of the ICSD, the
Leading Edge Chemical Crystallography Service Provision and Its Impact on. . .
113
Packing similarity [120] became an important factor to assess whether different
crystal structures could be polymorphic. The XPac [121] approach went further to
assess the degree to which crystal structures formed by different compounds might
be closely related. Implemented in Mercury around the same time, COMPACK
[122] looks at the similarity over multiple molecules when comparing crystal
packing. The related packing feature search [120] in Mercury will retrieve structures
with a similar crystal packing feature to that selected, such as a hydrogen-bonded
dimer. Other approaches of comparing crystal packing have developed, CrystalCMP
[123] employs differences in molecular rotation and relative positions, and Salbego
et al. proposed an index for ‘supramolecular clusters’ [124].
As computers have developed, the types of analysis scientists would like to do
have become more complex. There has been a trend towards informatics with the
rise of ‘big data’ and using computer programming to directly search and analyse the
database. The CSD Python API (Application Programming Interface) [14] has
enabled this type of query; with much of the CSD functionality now available
through fast computer algorithms, the knowledge has become easier to discover.
This means that it is possible to use most of the database searching functionality in a
machine-to-machine mode, allowing for scripting and automated workflow enactment and providing the potential to power machine learning and big data
approaches.
3.2.3 Web Tools
Technological advances and the proliferation of the Internet led to online access to
the database through WebCSD in 2010 [125], which gave easy access to the CSD via
a website without needing to install software locally on the user’s computer. This has
been superseded by a second version that reflects the changes in web technologies.
The ICSD [126], PDB [127] and COD (Crystallography Open Database [68]) [128]
databases provide a similar level of text and bibliographic searching via a web
interface. The distribution of the CSD and associated tools has also been transformed
by web developments, which have been vital as it has grown in size – in 2009 the
CSD was distributed on a CD, whereas today it is installed by download from the
Internet.
CrystalWorks [129] is a web-based search tool that takes advantage of the fact
that the different chemical crystallography data sources often need to be searched in
the same way. By indexing and aggregating the CSD, ICSD and COD, it enables a
simultaneous metadata search across them all at a high level, e.g. bibliographic,
composition or unit cell queries. It should be noted that the full CrystalWorks is
currently only available to UK academics through the Physical Sciences Datascience Service [130]; however, there is now a developing interface for all that
provides access to open (small molecule) crystallographic data. Conversely, while
there are a lot of similarities in how different crystallographic databases are searched,
there are also disciplinary differences, for instance, in the cases of the ICSD, the
Leading Edge Chemical Crystallography Service Provision and Its Impact on. . .
113