Increasing the virtual screening quality to identify adrenergic β2 receptor ligands using classification trees

The structure-based Virtual Screening (SBVS) campaign has been performed on an enhanced dataset of ligands and decoys for adrenergic β2 receptor using PLANTS1.2 as the molecular docking software and PyPLIF as an alternative post-docking scoring function. These approaches resulted in enrichment factor of true positives at 1% false positives (EF1%) values of 24.24 and 8.22 after ranked by using ChemPLP scores from PLANTS1.2 and by using Tc-PLIF values from PyPLIF, respectively. The attempts have also offered possibilities to explore the use of protein-ligand interaction fingerprint bitstrings resulted from rescoring using PyPLIF. In this article, the construction of classification trees employing the ChemPLP scores from PLANST1.2 and the protein-ligand interaction fingerprint bitstrings from PyPLIF as predictors to identify adrenergic β2 receptor ligands is presented. The best classification tree resulted in enrichment factor value of 201.64, which was significantly better at a 95% level of confidence compa...