The concept of a global chemical resolver is a distributed system that will allow chemical researchers to easily ask the general question, “Which organizations around the world have information about this chemical?” – but without needing to query each resource individually or to be familiar with each site’s search interface. Rather than attempting to provide complete records for a chemical from every different database, such a search would simply provide links back to individual resources where users can get more detail. For example, a user could input a SMILES or an InChIKey, and in a single click of a button, quicky discover whether (hypothetically) PubChem has a record for that chemical, or does EPA, or ChEMBL, etc. The it would be up to the user to choose which organizations they’re interested in, and to follow links to those sites to get more information as desired.
By using a common and reasonably simple communications protocol, any number of separate organizations would be able to register themselves as part of this “global search” and be able to provide results back to a single central query host. Importantly, this frees the central query manager from doing any actual cheminformatics work (e.g. parsing SMILES) or needing to know the details of each resource’s search API; it would be up to each individual resource to implement the search within their own infrastructure, presumably using methodology they already have internally. This simplifies the cross-site query system and makes it easy to add new databases to the search. It also spreads the computational burden across all organizations, rather than putting a heavy load on the central query server.
The following section further describes the conceptual data model, but is not intended as a detailed technical specification. The implementation examples provided are demonstrations only, not exemplars of a final product.