Protein and Ligand Preparations and Analyses using Docking and Molecular Dynamics

Check out our Demo Here!

I. Introduction

In the complex realm of drug discovery, understanding how small molecules interact with biological targets at the atomic level is crucial. While AI-driven predictions have revolutionized this field, grounding these predictions in biochemical and thermodynamic principles enhances their accuracy and applicability. While physics-based approaches in drug discovery are powerful, they’ve traditionally required deep expertise and long runtimes. Revilico aims to break that barrier by making these advanced insights available instantly—through an intuitive interface that fuses physics-based rigor with AI acceleration.

II. Molecular Docking: The First Step in Physics-Based Modeling

Molecular docking serves as the initial phase of physics-based modeling, aiming to predict the optimal binding pose of a ligand within a target’s active site. Key concepts include:

• Rigid vs. Flexible Docking: Rigid docking assumes fixed structures, while flexible docking accommodates conformational changes.
  
  

• Scoring Functions: These estimate binding affinity based on:

  • Electrostatic interactions

  • Van der Waals (VDW) forces

  • Hydrogen bonding

  • Hydrophobic interactions

Revilico automates this step by using hybrid AI-physics models to predict binding poses with high fidelity. We have an end to end suite using a variety of scoring functions including genetic algorithms, MM-GBSA inspired energy decomposition, random search, and gradient based search algorithms. We optimize for singular analyses as well as hundred-million bulk high throughput screens.

III. Equations and Physical Forces at Play

Central to physics-based modeling are the equations and forces governing molecular interactions:

• Energy Minimization: Utilizes force fields (e.g., CHARMM, AMBER, OPLS) to remove steric clashes and identify local minima in potential energy.
  
  
  

• Lennard-Jones Potential (LJ): Models VDW attraction and Pauli repulsion.
  
  
  

• Coulomb's Law (Electrostatics): Governs electrostatic interactions between charged particles.
  
  
  

• Bond Types: Includes covalent bonds and non-covalent interactions such as ionic bonds, hydrogen bonds, VDW interactions, and π-π stacking.

Revilico’s back-end applies validated force fields and approximations—such as Lennard-Jones potentials and Coulombic interactions—without the user needing to touch a single equation. Users simply upload or generate a structure, and Revilico handles the rest.

IV. The Dynamic Nature of Proteins

Proteins are dynamic entities influenced by factors like water content, electrolyte concentration, pH, and temperature. Molecular Dynamics (MD) simulations capture these dynamics by simulating atomic motions over time using Newton’s laws.

Unlike traditional rigid models, Revilico accounts for protein flexibility by incorporating molecular dynamics-informed ensembles into its predictions. These account for environmental conditions like solvent effects, temperature, and ion concentration, modeled automatically in the background. This allows the user to gauge protein flexibility, deviations of the protein based on ligand interactions, and intrinsic energies associated with binding.

V. Molecular Dynamics Simulations

MD simulations observe stability, binding evolution, and conformational changes of protein-ligand interactions. Key outputs include RMSD, Radius of Gyration (Rg), hydrogen bonding profiles, and binding energy calculations (ΔG).

With Revilico, users get metrics typically derived from lengthy MD simulations—such as RMSD, Radius of Gyration, and binding energy estimates—delivered in seconds to minutes, without any hardware setup or scripting.

VI. The Bottleneck: Traditional MD is Slow

Traditional MD simulations often require extensive compute time on clusters or supercomputers, posing challenges in configuration and interpretation for non-experts.

VII. Revilico’s Innovation: Instant, No-Code Physics-Based Outcomes

Revilico introduces a groundbreaking approach with AI-driven physics-based pipelines:

• Instant Outcomes: Start from a structure or sequence, generating binding energies, poses, and biophysical metrics like RMSD/Rg without coding or long waits.

• Democratized Access: Revolutionizes access to physics-driven insights, making them accessible without expert simulation knowledge.

By integrating physics principles with AI capabilities, Revilico empowers drug discovery by providing rapid, accurate, and interpretable results. This innovation marks a significant step towards enhancing efficiency and efficacy in pharmaceutical research and development.