fdMD

fdMD identifies ligand binding sites in proteins by running molecular dynamics on solvated small-fragment boxes with a repulsive Lennard-Jones potential to support fragment-based drug discovery (FBDD) and to reduce ligand aggregation and protein denaturation.

Key Features:

• Simulation of Solvated Fragments: Uses solvated small fragments in simulation boxes and applies a repulsive Lennard-Jones potential term to prevent ligand aggregation and maintain a minimal number of ligands around the target.
• Database Generation: Produces a database of ligand-solvated boxes for downstream analyses.
• Systematic Analysis Tools: Provides scripts to analyze simulation results and identify or discard spurious binding sites.
• Robust Descriptors: Employs descriptors including average MMGBSA (Molecular Mechanics/Generalized Born Surface Area) and KDEEP energies as indicators for binding site validity.

Scientific Applications:

• Fragment-Based Drug Discovery (FBDD): Supports FBDD campaigns by locating and characterizing fragment binding sites on protein targets.
• Binding Site Resolution: Identifies binding sites across diverse protein structures, including cases with multiple close-binding sites for the same ligand.

Methodology:

Targets are solvated with small fragment boxes and subjected to molecular dynamics simulations using a repulsive Lennard-Jones potential; four test cases were run for 200 ns (extended to 1 μs when necessary), and the selected number of replicas was reported sufficient to identify correct binding sites regardless of initial protein structures.