Authors: Nikola Minovski Andrej Perdih Tom Solmajer
Publish Date: 2011/08/12
Volume: 18, Issue: 5, Pages: 1735-1753
Abstract
The virtual combinatorial chemistry approach as a methodology for generating chemical libraries of structurallysimilar analogs in a virtual environment was employed for building a general mixed virtual combinatorial library with a total of 53871 6FQ structural analogs introducing the real synthetic pathways of three well known 6FQ inhibitors The druggability properties of the generated combinatorial 6FQs were assessed using an inhouse developed druglikeness filter integrating the Lipinski/Veber rulesets The compounds recognized as druglike were used as an external set for prediction of the biological activity values using a neuralnetworks NN model based on an experimentallydetermined set of active 6FQs Furthermore a subset of compounds was extracted from the pool of druglike 6FQs with predicted biological activity and subsequently used in virtual screening VS campaign combining pharmacophore modeling and molecular docking studies This complex scheme a powerful combination of chemometric and molecular modeling approaches provided novel QSAR guidelines that could aid in the further lead development of 6FQs agents
Keywords: