Significance. Conifer Point is compiling a vast array of data describing the binding of small chemical fragments to proteins. We use an innovative Monte Carlo simulator to produce this rigorous data, and use a powerful proprietary search algorithm to mine it. This information uncovers idiosyncratic details about protein binding sites that empower medicinal chemists to design therapeutic compounds which bind in a highly specific fashion to achieve cross-protein selectivity, mutation avoidance, and high ligand efficiency (affinity/molecular weight). Our goal is to provide chemists with affordable, ready access to powerful, predictive tools specifically designed to help them generate and prioritize ideas for progressing their hit compounds ? designs resulting in molecules that optimally bind into the protein target. Fragment binding data also enables projects to get started on difficult targets where screening has limited success. This can greatly expand the number of addressable disease targets, such as protein-protein interactions. Our team has uniquely demonstrated the ability to produce fragment maps on a large scale, generating >100,000 such maps in the last several years. More fragment maps enable broader searches to find more chemical possibilities. Our long-term goal is to make 100?s of fragment maps on a few 1,000 therapeutically-important proteins. Such a huge compilation of fragment binding data will be unprecedented. Innovation. Over the last decade, the PI and participants on this grant have been employing these fragment/water binding maps on dozens of successful pre-clinical lead identification projects, but they were restricted to proprietary commercial efforts. Our goal now is to make the information derived from fragment maps available at low cost to a broad community of researchers in drug discovery. We do this through a novel Web application that provides fragment-based design services leveraging the latest Web technologies for browser-based computation and integrates with other chemist-requested capabilities. Included are water maps ranked by binding free energy that are inexpensive to produce (<$1/protein) yet uniquely describe multi-body water configurations. Computational chemistry software is poised for a fundamental transition from desktop software to Web- based applications with low-cost subscriptions and a broader audience. Our Web application represents the vanguard of this transition.
Specific Aims. We have received feedback from evaluations of our prototype fragment-based chemistry design service regarding issues which have led to three specific aims.
Aim 1. Expand the collection of maps to include proteins from 20% of the Therapeutic Target Database, including a core set of fragments covering commercially- available fragment compound libraries.
Aim 2. Build interfaces in the Web application that enable users to define new fragments, and run incremental fragment map simulations on existing protein structures.
Aim 3. Dramatically improve the performance of the computations for energy minimization to achieve immediate feedback during design. Overall Impact. Researchers need alternative approaches to drug lead development. Mining fragment maps is a fertile source of new chemistry ideas. When widely adopted, the comprehensive Web service would provide the platform to significantly advance drug discovery.
Our mission is to provide easy access via the Web to a vast array of data describing the binding of small chemical fragments to proteins, leveraging this data to empower medicinal chemists in addressing the core challenges of drug discovery. Specificity, potency, and mutation avoidance of drugs require that compounds bind to the target protein in a way that exploits the highly idiosyncratic details of the 3D structure of that protein which are uncovered by the binding of the chemical fragments. This provides chemists with a powerful new tool for generating ideas to resolve conflicting drug optimization goals, and can greatly expand the number of addressable disease targets, such as protein-protein interactions, as has been demonstrated.
