Summary:
The advent of technological advances of computer-aided drug design has streamlined the drug design process, rendering it more cost- and time-efficient. Nowadays, rational structure-based drug design may quantify underlying molecular interactions involved in ligand-protein binding by utilizing the 3D structure of the therapeutic target in the process. Accurate quantification of these interactions can aid the optimization of binding affinity,selectivity, and other off -target interactions, which are a critical part of hit-to-lead and lead optimization efforts in drug discovery. One of the most important tasks in the lead optimization phase of the drug design process is to predict, among a series of lead candidates, which ones will bind more strongly to the therapeutic target. In this direction, relative binding free energy methodologies have been developed, which rely on physics-based molecular simulations and rigorous statistical mechanics to calculate the differences in the free energy of binding between a parent candidate drug and analogues. For example, Free Energy Perturbation (FEP) calculations coupled with Molecular Dynamics (MD) simulations calculate the free energy difference between an initial (reference) and an analog (target) molecule to an average of a function of their energy difference evaluated by sampling for the initial state.
Such methodologies have shown significant potential in the lead optimization process, however, they have been limited by technical challenges such as manual creation of large numbers of input files to setup/run/analyze free energy simulations. Automating free energy perturbation calculations would streamline the use of FEP calculations and would be a step forward to delivering high throughput calculations for accurate predictions of relative binding affinities before a compound is synthesized, and consequently save enormous experimental and human resources.
In this thesis, an algorithm called FEPrepare, which automates the set up procedure for relative binding free energy simulations has been designed and implemented as the first web-based server. The web server automates the set-up procedure for FEP calculations within the context of NAMD, one of the major MD engines. The user has to upload the structure files to the web-server. The algorithm is written in Python, utilizes the structure files uploaded by the user in order to perform atom renaming, and partial charge redistribution and create the necessary input files for VMD, a molecular viewer program, that can be used to help set up NAMD simulations and to help analyse and visualize NAMD output, to generate all needed files for the calculations. After the algorithm confirms compatibility of the required files with NAMD, it provides the user with everything needed to run a simulation.
Relative binding free energy calculations in drug design have proven very effective in facilitating the lead optimization process both time and cost efficient. The automation of Free Energy Perturbation calculations to provide access to large-scale simulations for lead optimization has been presented in this thesis.
Keywords:
free energy, drug design, NAMD, dual topology file, ligands, web development
