9 Structure-based Drug Design with Active Learning

 

This chapter covers

  • How three-dimensional protein structures can be leveraged to guide rational drug design through computational modeling of protein-ligand interactions.
  • Implementing a complete protein-ligand docking workflow to predict and evaluate how small molecules bind to target proteins.
  • Using deep learning approaches to create surrogate models that dramatically accelerate virtual screening of ultra-large compound libraries.
  • Building active learning systems that efficiently identify promising drug candidates while minimizing computational resources.
  • Extending active learning techniques to more rigorous binding affinity predictions with free energy perturbation for lead optimization.

Structure-based drug design (SBDD) leverages knowledge of the three-dimensional structure of a biological target, such as a protein involved in a disease, to guide the design and selection of molecules that can interact with and modulate its function. A key computational technique within SBDD is molecular docking, which simulates the interaction between a small molecule (i.e., a potential drug) and the target protein. Docking predicts how well a molecule binds to the protein and can be used to screen large libraries of compounds to identify promising drug candidates.

9.1 Docking: A Core SBDD Technique

9.1.1 Protein-Ligand Docking

9.1.2 Minimal Protein-Ligand Docking Workflow

9.1.3 Prepare the Protein & Ligand Structures

9.1.4 Run A Docking Experiment

9.1.5 Interaction Fingerprints

9.2 Active Learning for Hit Identification: Deep Docking

9.2.1 Active Learning: Smart Choices with Limited Resources

9.2.2 The Deep Learning Surrogate Model

9.2.3 Training the Surrogate Model

9.2.4 Initial Sampling

9.2.5 Acquisition Functions for Active Learning

9.2.6 The Oracle

9.2.7 The Active Learning Loop

9.3 Active Learning for Lead Optimization: Free Energy Perturbation Experiments

9.3.1 The Role of Free Energy Calculations

9.4 Summary

9.5 References