CONTOUR® SBDD: Growth and Scoring
Structure-based drug design using the CONTOUR® platform
Vitae Pharmaceuticals has developed the CONTOUR® technology platform to rapidly design and optimize compounds within the binding site of a target protein and simultaneously predict their binding activity in silico. This technology uses proprietary algorithms and machine learning to achieve an unprecedented level of speed and accuracy. It is capable of generating tens of thousands of drug-like compounds while simultaneously predicting their three-dimensional structure and binding affinities. The CONTOUR® platform uses three differentiated approaches to grow potential structures directly within the active site – de novo growth, directed growth, and programmable growth coupled with a proprietary scoring function – essentially allowing for rational drug design.
CONTOUR® technology grows molecules within a protein binding site starting from a user positioned starting fragment. The coordinates for the starting fragment can originate from a ligand in a high resolution co-crystal structure or by docking a fragment to a defined functional interaction site in the binding cavity. The de novo mode generates structures by picking fragments from the fragment library and adds them to the starting fragment. A directed growth mode generates molecules with a specified sequence of fragments determined by the structure of molecules to be modeled. The third and most powerful option is the programmable growth mode which interpolates between the de novo and the directed modes. It allows users to specify the growth instructions from very simple ones to complex nested loops to customize it to their needs. These instructions can be executed in an exhaustive or a random fashion to evaluate the possible combinations of growth instructions in assembling fragments.
Once a compound is designed, the CONTOUR® empirical scoring function identifies molecular configurations and low energy conformations that are sterically allowed in a given pocket of the binding site. This process of attaching additional fragments coupled with conformational searches and conformational optimization continues until a defined number of fragments are assembled, or until a preset molecular weight, rotatable bond, heavy atom count or other limit is achieved. Grown molecules ideally assume the shape and size of the chosen cavity guided by the sterics containing linear and branched attachments. The energy minimized conformations are subsequently optimized in the CONTOUR® platform to fine tune the intermolecular interactions between the protein and the ligand.
One of the great challenges of computational drug design technologies to date has been the ability to both design novel drug-like compounds, and to correctly predict their in vitro activity. The accuracy of CONTOUR® structure based drug design allows it to closely predict the activity of in-house designed and literature compounds.