Abstract:We study methods of non-linear global optimization in the problem of molecular docking and present a new implementation of protein-ligand docking and preparation of molecules for docking in a new open source software Hess. It uses BFGS for local optimization and two different forms of Monte Carlo method for global optimization. We compare Hess to another open-source package, Smina, and find that the chosen methods give better results in both performance and binding energy (using the same Vinardo scoring function). Also, during preparation of some molecules Hess provides a more accurate computation of implicit hydrogens thanks to the dearomatization algorithm provided by the Indigo cheminformatics package.
What problem does this paper attempt to address?
The main focus of this paper is the new implementation of molecular docking software, particularly the improved methods for molecular conformation optimization. The study presents a new open-source software called Hess, which utilizes BFGS for local optimization and two different Monte Carlo methods for global optimization. Compared to another open-source software Smina, Hess shows better performance and binding energy estimation. Additionally, Hess provides more accurate handling of implicit hydrogen calculations for certain molecules, thanks to the dearomatization algorithm provided by the Indigo cheminformatics package.
The paper introduces the molecular docking problem, which aims to predict the optimal pose and conformation of a drug ligand molecule in the active site of a protein to achieve the lowest binding energy. The ligand is considered flexible and can undergo rotations, translations, and internal conformational changes, while the protein is treated as rigid. Various energy estimation methods are mentioned in the article, including quantum mechanics, molecular mechanics, and empirical scoring functions such as Vinardo.
Hess software processes the molecules during the molecular preparation stage, which includes steps such as restoring chemical bonds, considering charge states, and adding hydrogen atoms. The algorithm specifically highlights the superiority of the dea