FRASE-bot is a computational platform enabling de novo construction of small-molecule ligands directly in the binding pocket of a target protein. It makes use of machine learning to distill 3D information relevant to the protein of interest from thousands of 3D protein-ligand complexes in the Protein Data Bank (PDB) and respective structure-activity relationships (SAR). At the first step of FRASE-based design, the target protein is seeded with ligand fragments through screening of FRASE database. DeepSPLIF, an original graph-convolutional neural network model, is used to recongnize most relevant, native-like fragment poses. Next, the seeded fragments are exploited to inform either (i) de novo design of 3D ligand structures or (ii) ultra-large-scale virtual screening of commercially available compounds. Here, option 2 will be used. The fragments identified will be converted into sets of pharmacophore features. The features will be clustered by type and 3D corrdinates. Centroids of the largest clusters will be used to compose pharmacophore models for large scale-virtual screening (of collections recommended by the CACHE organizers) using Phase (Schrodinger). Eventually, the top pharmacophoric hits will be docked using Glide (Schrodinger) and the top hits will be rescored using MM-PBSA (implemented in GROMACS).
FRASE-bot can be considered as a step toward a "virtual medicinal chemist". It exploits the concept of FRAgments in Structural Environments (FRASE) developed by Kireev group and applied to design potent in vivo antitumor agents . Most recently, FRASE-bot was used to identify small-molecule ligands for Calcium and Integrin Binding protein 1 (CIB1) , a promising target against triple negative breast cancer (TNBC). The signaling function of CIB1 relies on protein-protein interactions and its structure does not feature any natural ligand-binding pocket. No small-molecule CIB1 inhibitors have been reported yet.
 Da, C.; Zhang, D.; Stashko, M.; Vasileiadi, E.; Parker, R. E.; Minson, K. A.; Huey, M. G.; Huelse, J. M.; Hunter, D.; Gilbert, T. S. K.; Norris-Drouin, J.; Miley, M.; Herring, L. E.; Graves, L. M.; Deryckere, D.; Earp, H. S.; Graham, D. K.; Frye, S. V.; Wang, X.; Kireev, D. Data-Driven Construction of Antitumor Agents with Controlled Polypharmacology. J. Am. Chem. Soc. 2019, 141, 15700–15709.
 Leisner, T. M.; Freeman, T. C.; Black, J. L.; Parise, L. V. CIB1: A Small Protein with Big Ambitions. FASEB J. 2016, 30, 2640–2650.
Schrodinger, Pipeline Pilot
1) Da et al, Data-Driven Construction of Antitumor Agents with Controlled Polypharmacology. J. Am. Chem. Soc. 2019, 141, 15700–15709
2) Da, C.; Kireev, D. Structural Protein–Ligand Interaction Fingerprints (SPLIF) for Structure-Based Virtual Screening: Method and Benchmark Study. J. Chem. Inf. Model. 54, 2555–2561.
3) Da, C.; Stashko, M.; Jayakody, C.; Wang, X.; Janzen, W.; Frye, S.; Kireev, D. Discovery of Mer Kinase Inhibitors by Virtual Screening Using Structural Protein–Ligand Interaction Fingerprints. Bioorg. Med. Chem. 2015, 23, 1096–1101.
4) Kireev, D. ChemNet: A Novel Neural Network Based Method for Graph/Property Mapping. J. Chem. Inf. Comput. Sci. 1995, 35, 175–180.