In Derwent Chemistry Research chemical structures are searched against a database of manually curated chemicals from patents (Derwent Chemistry Resource) and a database of chemicals indexed in automated fashion (SureChEMBL).
All chemicals inside Derwent Chemistry Research link to either a patent record or a WoS literature record.
When clicking on a chemical structure within the patent set you can see in which database the chemical substance appears.
DCR - Derwent Chemistry Resource
DCR chemicals are manually select by subject matter experts when deemed relevant for the patent. This means that the substance is either claimed or carries specific information such as data that is important for this patent. Indexing has started in the 1980's. Today 37 patent authorities are covered. At least one patent per patent family is being indexed.
Chemicals are being indexed from patents when claimed or exemplified and if their size is < 255 atoms excluding H-atoms
SureChEMBL
SureChEMBL is a publicly available large-scale resource containing compounds extracted from the full text, images and attachments of patent documents. The data are extracted from the patent literature according to an automated text and image-mining pipeline. https://chembl.gitbook.io/surechembl/
Structures are being indexed if the size of a chemical structure is below around 4800 Daltons independent of where in the patent they are mentioned.
Note that this indexing ids done in an automated process.
PubChem
PubChem chemicals are linked to DCR and SureChEMBL chemicals, but only the information for chemicals names is being used. (see chemical search - adding names to database search)
Content from those 3 databases is highly complementary and harmonized into one Chemical Master Record (CMR) which is maintained and regularly updated together with the Substance Patent Linking.
For WoS literature an own chemical database is being used. Mapping this information to CMR is in progress.