Most drugs work by binding specific, disease-related proteins;and molecules targeted to bind specific proteins are also used as chemical probes to elucidate pathways and biomolecular mechanisms. Years of drug discovery and chemical biology research have generated a large body of molecular recognition data, which are of value not only for drug discovery and chemical biology, but also in the emerging fields of systems pharmacology and pharmacogenomics. However, for many years, these data were hard to access and use because they are published almost exclusively in patents and the scientific literature, and hence have not been available in machine-readable form. Our BindingDB database came on line in late 2000 as the first public database collecting a broad set of these binding data from the scientific literature and offering it on the Web for query, download, and analysis. Today, BindingDB's web-site provides about 8,000 visitors a month with ready access to nearly 800,000 binding data for 340,000 different drug-like small molecules and 6,400 proteins, most of them candidate drug-targets. BindingDB's significance is enhanced by the NIH's intensifying focus on translational research and a growing interest in drug- discovery at many universities. In the coming grant period, we plan to further improve BindingDB's value to the biomedical research community. A central effort will be to continue expanding BindingDB's data collection, by collaboratively curating protein-small molecule binding data from the scientifi literature, initiating data curation from patents, and broadening our scope to include selected biopharmaceuticals. We will furthermore develop and deploy innovative tools for accessing, viewing, analyzing and applying the data contained within BindingDB. For example, we aim to create high-level browsers that will allow users to navigate the sea of data within BindingDB;connect BindingDB's data and functionalities with emerging workflow systems for information analysis and visualization;and expand BindingDB's tools in support of systems pharmacology and pharmacogenomics.
Most medications are molecules that stick to those protein molecules in the body which are involved in causing a given disease. A tremendous amount of research into which molecules will stick to which proteins has been carried out and published in scientific papers and patents, but it has been very difficult for scientists to locate and use the information they need because of the limitations of traditional paper publications. Here, we propose to continue and expand our project to collect as much of this information as possible and put it into a large, publicly accessible database on the world-wide web, along with search and analysis software, to help scientists find and use the data they need in their efforts to discover new medications.