Speed of calculations

Speed of ligand docking calculations depends on the settings of genetic algorithm search. Currently two types of settings are predefined in Lead-Finder: more accurate and exhaustive search in the default docking regime and faster search in the screening regime. While default docking settings were adjusted to yield maximum docking accuracy, settings for the screening regime were ´lightened‘ to achieve maximum speed of calculations at the cost of small (~5%) decrease in Docking Success Rate.

To benchmark speed of docking calculations a set of 407 diverse protein-ligand complexes (used also for Docking Success Rate measurements) was used. All calculations were performed in two regimes: default docking and screening. Results of current benchmarking studies providing average time needed to dock ligand with a certain degrees of freedom (number of freely rotatable bonds) are presented on the figure. As can be seen from the figure, screening regime performs 2–4 times faster than default docking, taking on average 30 seconds for docking ligand with 6 freely rotatable bonds.

a) Time pre ligand indicates average time (in seconds) of docking one compound from the library (on the single processor core).
b) Total screening time  total time (in hours) of virtual library screening on the computer cluster.
c) Volume of the energy grid  indicates the volume of the energy grid (for more details on grid definition see section Technology).

Speed of ligand docking was also benchmarked by us in a number of real-life virtual screening studies, in which commercial library of 300000 compounds (STK library by Vitas-M Laboratory) was screened against a number of protein targets. Using a cluster of 16 computational nodes (dual Xeon 5150 2.6 GHz) we were able to screen the whole library in 12-24 hours depending on the protein target. Average time for docking on compound from the library for particular protein is provided in table above.

To extrapolate speed of docking calculations for other machine architectures, the following table can be used:

Copyright  2008  MolTech Ltd