Implementation of PLS discriminant analysis to rank indirubin derivatives against decoys

摘要

Par tial Least Squares Discriminant Analysis (PLS-DA) is employed to obtain novel combinations of energetic terms present in classical scoring functions, which exceed and compensates the "traditional" consensus scheme. These novel scoring functions were involved to rank the database of indirubin inhibitors of glycogen synthase kinase-3β and cyclin dependent kinase-2 decoys from Directory of Useful Decoys. The ability of docking-scoring algorithm to prioritize the actives is assessed by means of several metrics. The best classifcation function includes donor component of Chemgauss2, steric contribution from Chemgauss3 and rotatable bond term of ScreenScore and provide signifcant improvement of enrichment factor at 5% of database.