<![CDATA[Structure-Based Virtual Screening (SBVS) protocols targeting C-X-C chemokine receptor type 4 (CXCR4) have been constructed by employing PLANTS 1.2 to perform molecular docking simulations and PyPLIF 0.1.1 to identify Protein-Ligand Interaction Fingerprint (PLIF). By using ChemPLP score from PLANTS 1.2 and Tc-PLIF from PyPLIF 0.1.1 to select best pose in the retrospective SBVS showed Enrichment Factor (EF) values of less than the EF value of the reference SBVS (17.5). Nevertheless, the retrospective SBVS campaigns have also resulted in PLIF bitstrings for all poses resulted from the molecular docking simulations.In this article, binary Quantitative Structure-Activity Relationship (QSAR) analysis employing new predictors ensemble PLIF resulted from the retrospective SBVS campaings, instead of using PLIF bitsrings from the best pose only, are presented. The ensemble PLIF as predictors were calculated by taking into account all poses with ChemPLP score lower than a certain ChemPLP score as the predefi ned cutoff , and subsequently for every compound the percentage of “on” interactions was calculated for every PLIF bitstring. The predefi ned cutoff was selected by performing systematic trials to obtain a ChemPLP score as the cutoff with the highest F-score and Matthews correlation coeffi cient (MCC) value.The results showed that the F-score and MCC values could reach 0.58 and 0.61, respectively with EF value of 323.47, which was much better than the EF value of the reference SBVS protocol.]]>
Copyrights © 2017
