RASPD⁺ (RApid Screening with Physicochemical Descriptors + Machine Learning) is a computationally fast protocol for identifying lead-like molecules based on predicted binding free energy against a target protein with a 3D structure and a defined ligand binding pocket. RASPD was originally developed at the Supercomputing Facility for Bioinformatics and Computational Biology,(SCFBio), Indian Institutes of Technology Delhi (IITD) by Goutam Mukherjee and B. Jayaram, and development continued at Heidelberg Institute for Theoretical Studies (HITS) in the Molecular and Cellular Modeling group. In version 1.0 of the RASPD⁺ software, new feature like scaffold search was added and several machine learning algorithms were introduced. The model was trained on around 4000 non-metallo protein-ligand complexes retrieved from the PDBBIND refined data set.

Performance of RASPD⁺ A Pearson correlation coefficient of 0.74 and an RMSE ±1.86 kcal mol⁻¹ were obtained when predicting binding energies for test sets consisting of 493 completely unseen protein–ligand complexes. The performance of RASPD+ is comparable with that of other scoring functions like KDeep and other methods but does not require docking of ligands into protein binding sites. Using this method, it is possible to screen a million molecule library against a target protein of known binding pocket within a couple of minutes.

For more details, please visit our Documentation or page.