spark-cpvs

The software tool 'spark-cpvs' proposes a strategy for structure-based virtual screening of ligands against target proteins, which is based on iteratively docking a set of ligands, training a ligand-based model on this set, and predicting the remainder of the ligands to exclude those predicted as 'low-scoring' ligands. The tool uses SVM and conformal prediction to deliver valid prediction intervals for ranking the predicted ligands and Apache Spark to parallelize both the docking and the modeling. The results show that conformal prediction based virtual screening (CPVS) is able to reduce the number of docked molecules by 62.61%, retain an accuracy for the top 30 hits of 94% on average and have a speedup of 3.7.