Annual subscription of Chemical Abstracts Service (CAS) BioFinder Discovery Platform Software License

The FDA is seeking to purchase an annual software license for the Chemical Abstracts Service (CAS) BioFinder Discovery Platform. This platform is a computational tool that can predict thousands of biological targets for chemicals and interfaces with the CAS SciFinder database, which is already used by the FDA. The platform must meet specific technical requirements, including comprehensive curated data, multiple independent prediction algorithms, and extensive biological target coverage. The FDA believes that CAS is the sole source capable of meeting these requirements. Interested parties may submit a capability statement, proposal, or quotation, which will be considered by the agency. • The platform must predict biological targets for chemicals based on a training set of over 1 million compounds. • The platform must provide known and predicted quantitative binding and functional activity for at least 3,600 targets. • The platform must use at least 6 independent methods for target binding prediction. • The platform must provide confidence scores for predicted interactions using a consensus approach. • The platform must include metabolite prediction using a manually curated training dataset. • The platform must allow for direct comparison of biological target predictions of metabolites with substrates. • The platform must include interactive tabular and graphical dashboards for data visualization. • The platform must integrate with the CAS SciFinder database. • The platform must be accessible via the internet with user account credentials. • The platform must allow for downloadable search results in a structured data format. • The platform must include an AI chatbot interface for data queries. • The platform must have matched molecular pair analysis capabilities.