State-of-the-art computational methods and protocols are at disposal to unravel the molecular mechanism of complex biological systems involved in pathogen infections. Specifically, Molecular Dynamics (MD) based on classical force fields and Hybrid Quantum Mechanics/Molecular Mechanics (QM/MM) simulations are performed to characterize the structure, the dynamics and the function and the molecular mechanism of biomolecules, at an atomic-level of detail.
The simulations are performed with including GROMACS and Amber2024 for the classical MD simulations, and Cp2k for QM/MM MD. To sample slow events occurring on time scales longer than 10 of μs, routinely performed in standard MD simulations, biased simulations can be applied. These include metadynamics and related methods, thermodynamic integration, umbrella sampling, free energy perturbation. Many of these simulations can be run with the PLUMED plugin.
As an example, this set of computational tools can be leveraged to explore how pathogens interact and corrupt human or plant membranes, the molecular mechanism of enzymatic processes, the molecular determinants of protein/protein or protein/nucleic acid interactions, slow functional motions of biomolecules, allosteric regulatory mechanisms.

In silico drug discovery in the field of pathogen infection involves the use of computational docking algorithms to screen vast libraries of commercially available molecules to identify potential drug-candidates targeting a specific user-selected biomolecular target involved in bacterial, viral and plant infections. The libraries of commercial molecules commonly used to identify hit compound are TargetMol, Molport, Enamine, Cambridge, and ZINC20.
Molecular docking simulations are routinely performed using the Glide package within Schrödinger’s program, to predict how molecules interact with the target and rank them on the basis of their binding free energy and pharmacological properties. Molecular Dynamics (MD) simulations can then be applied to investigate the behavior of drug-target complexes over time to evaluate stability and efficacy. The integration of bioinformatics and machine learning further enhances the efficiency and accuracy of in silico drug discovery pipelines. In this context molecular docking complemented by all atom simulations can also be used to unravel the mechanism of known inhibitors.

With the advent of single-particle cryogenic electron microscopy (Cryo-EM), the ability to examine large DNA, RNA, and protein macromolecules at near-atomic resolution has become a reality, facilitating the analysis of their structural and functional attributes. However, traditional refinement tools typically assume that all acquired images represent a singular structure. This assumption presents significant hurdles in resolving highly disordered and flexible regions. To tackle this challenge, all atoms simulations of experimentally resolved structure can be run to refine the structure resulting from Cryo-EM data. More advanced approaches can be applied:

• Molecular Dynamics (MD) simulations can be used to flexibly fit atomic structures into density maps by adding external forces proportional to the gradient of the density map.
• Methods that enforce the average structures derived multi-replica MD simulations to the experimental density maps can be run thus extracting an ensemble of structure that better represent the cryo-EM derived electron density.

These methods can be applied to a diverse set of biological systems, ranging from single proteins and RNA macromolecules to large molecular machines.