Selection of Docking Routine for Fast and Efficient Virtual Screening

Presentation: P01011

Session: Cheminformatics, Library Design and Virtual HTS - Poster Session

Dmytro Koavslkyy, Maxim Platonov, Galina Dubinina, Volodymyr Dubina and Galina Dubinina,
Kiev National Taras Schevchenko University

Presenting Author: Dmytro Kovalskyy, Kiev National Taras Schevchenko University - Ukraine

    Selecting combination of docking engine and scoring function remains the crucial step for the successful employment of the Virtual Screening approach in drug discovery. In present study we evaluate performance of QXP's docking routines in order to define that suitable for fast virtual screening of large database. The final goal of the study was to select and optimize docking procedure to obtain a reasonable pose prediction per complex within 1 minute. Thrombin was selected as test target enzyme and Enamine's stock collection as compounds set.

    Based on the results to be presented here we suggest to use 100 steps of sdock routine for fast virtual screening with satisfactory prediction of binding poses. Other routines either are slow and require more steps for convergence or give unacceptable results. To process docking results we have developed proprietary Multi-Filter program that allows smart and stepwise selection of those compounds that fits to a particular pharmacophore model criteria.


Close Window X