Parameter refinement for molecular docking

Finding the optimal parameter values for any computer program with adjustable parameters can be very time consuming. In this paper, we introduce the use of the Plackett-Burman and the central composite designs with the aid of the partial least squares method to tackle this problem. Using DOCK3.5 as a test case, we also show a four-step procedure for sequential docking utilizing two parameter sets, both effecting a different level of accuracy. The DOCK parameter values were refined for protein kinase C regulatory domain yielding an orientation at the global "energy" minimum, which is in very good agreement with the experimental protein kinase C regulatory domain-phorbol 13-acetate complex. The scheme is now being used for screening molecular databases to find putative protein kinase C inhibitors.