Table of Contents

 

Chapter 1 Atomistic force fields for molecular simulations with emphasis on the CHARMM additive and Drude polarizable models

1.1.       Introduction

1.2.       Additive and Drude Potential energy functions

1.3.       Water

1.4.       Proteins

1.5.       Nucleic Acids

1.6.       Carbohydrates

1.7.       Lipids

1.8.       Atomic Ions

1.9.       Small Molecules

1.10.   Summary and Future Directions

 

Chapter 2 Treating Non-covalent Interactions in Biomolecules with QM and QM/MM models

1. 1.      Introduction
2. 2.      General Computational Models

2.1.1.       Full QM and QM/MM Methods

2.1.1         QM Methods

2.1.2         MM Force Fields

2.1.3         QM/MM Interactions

2.2   Machine learning potential functions

2.2.1         Standalone ML Potentials            

2.2.2         ?-learning ML potentials

2.2.3         Hybrid ML/MM Potentials

2.3   Computation of the Binding Energies and Free Energies

2.3.1         Docking and Minimization

2.3.2         Basic Binding Free Energy Simulations

2.3.3         Multi-level End-point Methods for Improved Accuracy

2.3.4         Hydration of the Binding Sites

1. 3.      Applications and Discussions

3.1   Characterization of Protein-Ligand Interactions Using Full QM Calculations

3.2   QM/MM Simulations for Protein-Ligand Interactions

3.3   ML/MM Binding Free Energy Simulations

1. 4.      Concluding Remarks

 

Chapter 3 QM/MM Simulations on Catalytic Mechanisms of Metalloenzymes

1.1.             Introduction

1.2.             QM/MM Simulations on Catalytic Mechanisms of Metalloenzymes

1.2.1.       The QM/MM Models

1.2.1         Simulation Techniques based on QM/MM Models

1.3.             Applications of QM/MM Methods

1.3.1.       Exchange and Superexchange Enhanced Reactivity

1.4.             Preorganized Local Electric Fields Effect

1.5.             Dynamic Effects in Modulating the Reactivity and Selectivity

1.4. Conclusions

1.5. Acknowledgements

1.6. References

 

Chapter 4 Integrative Multiscale Modeling of Biomolecular Interactions: from Mechanistic Understanding to Design

1. Introduction
2. Integrative Multiscale Molecular Modelling of Biomolecular Interactions and Mechanisms

2.1. Computational Approaches to Explore Conformational Dynamics and Binding

2.1.1.        Classical Molecular Dynamics Simulations

2.1.2.        Enhanced sampling methods

2.1.3.        Advanced MD Techniques for Environment-Sensitive Biomolecular Modeling

2.1.4.        Quantum and Semi-empirical Methods for Protein-Ligand Interaction Analysis

2.1.5.        Modeling Reactivity in Enzymatic Catalysis

2.1.6.        Quantum Chemical methods and QM/MM approaches

2.1.7.        Direct quasi classical molecular dynamics

2.1.8.        Post-Simulation Analysis of Biomolecular Interactions

2.1.9.        Mapping Non-Covalent Interactions

2.1.10.    Protein-Communication networks

1. Case Study I: Time-evolution of the Millisecond Allosteric Activation of IGPS

3.1. Effect of PRFAR binding in IGPS: transient oxyanion hole formation and interface closure

3.2. Modeling Spontaneous Substrate Binding and Formation of the Ternary Complex

3.3. Time evolution toward the active ternary complex: IGPS caught in the allosterically active state

3.4. Free-energy Landscape of the IGPS Allosteric Activation

3.5. Activation of correlated motions: unraveling the allosteric activation mechanism of IGPS

3.6. Mechanistic Insights into IGPS Catalysis: Impact of Active Site Preorganization

3.7. Integrative Multiscale Modeling of Allosteric Activation: Strategy and Implications

1. Case Study II: Protein Energy Networks as a Framework for Allosteric Pathway Discovery and Drug Design in GPCRs

4.1   Mapping the Free Energy Landscape of A1R Activation

4.2   Decoding Allosteric Communication through Protein Energy Networks

4.3   Integrating Molecular Dynamics and Protein Energy Networks for Allosteric Drug Design

1. 5.      Case Study III: Mechanistically-Guided Design of Enzymes for C-N Bond Formation via Multiscale Modeling

5.1   Elucidating the mechanism of (S)-Selective C-N bond formation in P411 Variants

5.2   Mechanistically-guided Design of (R)-Selective p411 Variants for C-N Bond Formation

5.3   Integrative Multiscale Modeling for Enzyme Design

1. 6.      Perspective and outlook

 

 

Chapter 5 Methodological Advances in Computational Biomodelling: End-point Free Energy Approaches

1. 1.      End-point Free Energy Approaches: An Introductory Tale
2. 2.      Methodological Advances

2.1. Naïve implementations: MM/GBSA and MM/PBSA

2.2. The entropic component

2.3. The interior dielectric constant

2.4. The sampling strategy

2.5. Elevation of the core Hamiltonian

2.6. The mutational scanning derivatives

1. 3.      Practical Applications

3.1. Protein-X binding

3.2. Host-guest binding

3.3. Multimeric systems

References

 

Chapter 6 Computational modeling of protein-protein/peptide/RNA interactions

1. Introduction
2. Prediction of protein-protein complexes by traditional docking methods
3. Structure prediction of biomolecular complexes using deep learning methods
4. Prediction of protein-peptide and protein-RNA complexes using deep learning methods
5. Molecular dynamics simulation of proteins in complex with proteins, peptides and RNA
6. Calculation of Binding Free Energies of biomolecular complexes
7. Peptide and protein interaction design
8. Conclusions

 

Chapter 7 Modeling Peptide–Protein Interactions with MELD: A Physics-Based Framework for Structure, Affinity, and Design

1.1. Introduction: The Biological and Clinical Relevance of Peptides
and Peptide Epitopes

1.1.1.       Opportunities and Potential of Peptides as Therapeutic Agents

1.1.2.       Risks and Challenges in Peptide-Based Therapeutics

1.1.3.       The Impact of Computational Chemistry on Peptide Drug Discovery Pipelines

1.2.     MELD: a versatile computational tool for modeling molecular recognition

1.2.1.       The MELD approach

1.2.2.       MELD-Adapt learns dataset accuracy

1.3.     Protocols

1.3.1.       MELD Binding protocols

1.3.2.       MELD Binding Affinity

1.3.3.       Quantifying binding receptor selectivity

1.3.4.       Quality control of MELD simulations

1.4.     Case studies

1.4.1.       Linear Peptides

1.4.2.       Cyclic peptides

1.4.3.       Design of Miniproteins to Compete Peptide Epitope Binders

1.5.     Conclusions

1.6.     Acknowledgements

1.7.     Bibliography

Chapter 8 Transforming Drug Binding with AI-Enhanced Computational Modeling

INTRODUCTION

COMPUTATIONAL STRATEGIES OPTIMIZING DRUG DEVELOPMENT

INNOVATIONS IN COMPUTATIONAL MODELING OF RECEPTOR-LIGAND BINDING DYNAMICS

Binding Kinetics as a Predictor of In Vivo Drug Effectiveness

Trypsin–Benzamidine as a Benchmark for Computational Methods

Precise kinetic rate estimates can be achieved within sub-microsecond simulation times

Improved Kinetic Predictions via Next-Gen Sampling Methods

Impact of Force Fields and Sampling Limitations on Kinetic Rate Accuracy

Ligand Gaussian accelerated molecular dynamics (LiGaMD)

ML enables reliable predictions and reduced computational cost

Increasing simulation time can enhance predicted kinetic rates accuracy

AI IN TRANSFORMING PRECLINICAL DRUG DEVELOPMENT

COMPUTATIONAL AI RESOURCES FOR DRUG DEVELOPMENT 

Notable Achievements of AI in Drug Discovery

Target Protein Conformation and Binding Region Exploration

Supplementing AI-based Compound Hunt

Compound-based Search

Conformational-based Search  

Active feedback-informed virtual docking

RECEPTOR-LIGAND INTERACTIONS PREDICTION USING AI

Traditional ML strategies for predicting binding affinity

Forecasting Binding Affinity through DL

Conventional ML and DL Models for ligand binding

CONCLUSION AND FUTURE DIRECTION

REFERENCES

 

Chapter 9 Molecular simulations-based predictions of drug’s residence times in the Exascale era: current status and recent advances

1. Introduction

2. MD-based methods to estimate RTs

3. Qualitative and fast estimates of RTs

4. Statistical mechanics-based methods for quantitative estimates of RTs

4.1 Methods that use bias potentials or forces

4.1.1 Ligand Gaussian accelerated molecular dynamics (LiGaMD)

4.1.2 Constant-Force Steered MD

4.1.3 Dissipation-corrected Targeted MD

4.1.4 Metadynamics-based methods

4.2 Markov state models

4.3 Methods that sample the transition path ensemble

4.3.1 Weighted Ensemble (WE)

4.3.2 Milestoning

4.3.3 Transition State-Partial Path Transition Interface Sampling (TS-PPTIS)

4.3.4 Adaptive Multilevel Splitting (AMS)

5. Predicting RTs in the exascale era: a Julich perspective

5.1 Sampling stochastic trajectories via an Onsager-Machlup action-derived molecular dynamics algorithm

5.2 Sampling the transition path ensemble and computing kinetic rates by combining MD in trajectory space with metadynamics

5.3 GROMACS implementation of the path-MD algorithm

5.4 Weak scaling of the path-MD algorithm on a pre-exascale supercomputer

6. Conclusions

7. Code repository

8. Acknowledgments

 

Chapter 10 Binding thermodynamics and kinetics of host-guest systems determined from long-timescale molecular dynamics simulations

1. Introduction

1.1 Host-Guest Systems: A Basic Research Cornerstone

1.2 Structures of Host-Guest Systems

2. Background on Molecular Recognition 

2.1 Non-Covalent Binding Thermodynamics 

2.2 Host-Guest Binding Kinetics 

2.3 Coupling of Thermodynamics and Kinetics 

3 Molecular Dynamics Techniques 

3.1 Why Use Molecular Dynamics? 

3.2 Molecular Dynamics 

3.3 Classical MD 

3.4 cMD Post-Analysis Techniques 

3.4 Enhanced Sampling Techniques 

3.5 Umbrella Sampling 

3.6 Milestoning Theory 

3.7 Density Functional Theory 

3.8 Brownian Dynamics 

4. Using Long Classical MD to Investigate Chemical Host-Guest Systems 

4.1 βcyclodextrin via Long MD

4.2 Melatonin in β-CD Nanosponges

4.3 CB7 Host–Guest Calorimetry Simulations

4.4 Aspirin–β-CD Protonation Study

4.5 β-CD Modifications Tune Curcumin Binding Energetics

5. Other Molecular Modeling Tools to Study Host-Guest Systems

5.1 Umbrella Sampling in Host-Guest Systems 

5.2 Metadynamics in Host-Guest Systems 

5.3 Milestoning?in Host-Guest Systems 

6. Final Perspective

 

Chapter 11 Supervised molecular dynamics approaches to protein-ligand (un)binding

1. 1.         Introduction

1.1. The importance of the (un)binding pathways

1.2. The molecular dynamics sampling issue and the (un)binding processes 

1.3. The challenge of identifying pathways: roots in classic physics

1.4. The use of enhanced and adaptive methods to increase sampling 

1. 2.      Supervised MD (SuMD) accelerates ligand binding sampling

2.1. The SuMD original algorithm

2.1.1.       System Initialization

2.1.2.       Trajectory Segmentation 

2.2.1.       Progress Evaluation 

2.2.2.       Supervision Adjustment 

2.2.3.       Termination Criteria 

2.3.       Adaptation to the first SuMD algorithm 

2.3.1.       SuMD for Ligand Unbinding Simulations 

2.3.2.       Multiple Walker SuMD (mwSuMD) 

1. 3.            SuMD approaches to provide functional hypotheses for binding and unbinding interactions

3.1. Adenosine receptor agonist binding, activation, and the role of extracellular loops in selectivity

3.2. From stability to dynamics in Class B GPCRs drug design

3.3. Towards understanding the G protein activation mechanism with mwSuMD

1. 4.       Conclusion
2. 5.       Bibliography

 

Chapter 12 Dissipation-corrected targeted Molecular Dynamics

1. 1.       Introduction
2. 2.       Theory

2.1. Free energies and friction

2.2. Fluctuation theorems, Jarzynski’s equality and its cumulant approximation

2.3. Targeted MD simulations

2.4. Markovian Langevin equations

2.5. Dissipation correction and non-equilibrium friction factors

2.6. Temperature-boosted Langevin simulations

2.7. Paths and path-collective variables

1. 3.      Applications

3.1. Benchmark system: sodium chloride

3.2. Velocity-dependent friction in lubricants

3.3. Predicting ligand (un)binding paths and kinetics

3.4. The emergence of anisotropic friction

3.5. Driven transport through functional materials and ion channels

3.6. Inducing conformational changes and allosteric information transfer in proteins

1. 4.      Practical considerations

4.1   Setting up dcTMD simulations

4.2   On finding pathways

4.3   On the use of Langevin equation integrators

4.4   Theoretical limits and error sources

1. 5.      Conclusion
2. 6.      Acknowledgements
3. 7.      References

 

 

Chapter 13 Hybrid Gaussian accelerated molecular dynamics and the weighted ensemble methods for biomolecular simulations

1. 1.         Introduction
2. 2.         Methods

1.2.1.      Gaussian accelerated molecular dynamics (GaMD)

1.2.2.      Weighted ensemble method (WE)

1.2.3.      Gaussian accelerated molecular dynamics-weighted ensemble method (GaMD-WE)

1.2.4.      Parallelizable Gaussian accelerated molecular dynamics (ParGaMD)

1. 3.         Applications of GaMD-WE

3.1. Applications of ParGaMD

1. Conclusions
2. Bibliography

 

Chapter 14 Replica Exchange Gaussian Accelerated Molecular Dynamics for Enhanced Sampling and Free Energy Calculations of Biomolecular Interactions

1. 1.      Introduction
2. 2.      Theory

2.1.Conventional GaMD Theory

2.2. Conventional Rex-GaMD Theory

2.3.Multiple Parameter Rex-GaMD (MP-Rex-GaMD) Theory

1. 3.      Simulation Workflow
2. 4.      Applications of Rex-GaMD

4.1.Dialanine

4.1.1.      Design of Acceleration Parameters in Conventional GaMD

4.1.2.      Replica Exchange Strategy for Dialanine

4.1.3.      Performance of Force Constant Rex-GaMD (σ-Rex-GaMD)

4.1.4.      Performance of Threshold Energy Rex-GaMD (E-Rex-GaMD)

4.1.5.      Performance of Multi-Parameter Replica Exchange GaMD (MP-Rex-GaMD)

4.1.6.      Free Energy Landscapes and Convergence

4.1.7.      Summary

4.2.Chignolin

4.2.1.      Design of Acceleration Parameters in Conventional GaMD

4.2.2.      Replica Exchange Strategy for Chignolin

4.2.3.      Performance of Force Constant Rex-GaMD (σ-Rex-GaMD)

4.2.4.      Performance of Threshold Energy Rex-GaMD (E-Rex-GaMD

4.2.5.      Performance of Multi-Parameter Replica Exchange GaMD (MP-Rex-GaMD)

4.2.6.      Free Energy Landscapes and Convergence

4.2.7.      Summary

4.3.HIV Protease

4.3.1.      Design of Acceleration Parameters in Conventional GaMD

4.3.2.      Replica Exchange Strategy for HIV Protease

4.3.3.      Performance of Force Constant Rex-GaMD (σ-Rex-GaMD)

4.3.4.      Performance of Threshold Energy Rex-GaMD (E-Rex-GaMD

4.3.5.      Performance of Multi-Parameter Replica Exchange GaMD (MP-Rex-GaMD)

4.3.6.      Free Energy Landscapes and Convergence

4.3.7.      Summary

1. 5.      Challenges and Future Directions

5.1.Computational Cost and Scaling

5.2.Optimization of Parameters

5.3.Convergence Assessment

5.4.Choice of Exchanged Parameters

5.5.Analysis Complexity

5.6.Integration with Experimental Data

5.7.Integration with Machine Learning and Artificial Intelligence

5.8.Development of User-Friendly Software and Tools

5.9.Development of More Robust Free Energy and Kinetic Analysis Methods

Permission

References

 

Chapter 15 Enhanced sampling of biomolecular interactions with Gaussian accelerated Molecular Dynamics

 

1. Introduction
2. Methods

2.1. Gaussian accelerated Molecular Dynamics (GaMD)

2.2. Selective GaMD

2.3. GaMD, Deep Learning and Free Energy prOfiling Workflow (GLOW)

1. Applications

3.1.      Proteins

3.1.1.       Dynamic allostery of GPCRs

3.1.2.       Elucidating pathways of peptide agonist dissociation and deactivation in class B2 G-protein-coupled receptors

3.1.3.       Understanding the mechanism of activation and substrate processing in γ-secretase complex

3.1.4.       Dynamic insights into Polycystin-1 (PC1) activation by stalk-derived peptide agonists

3.1.5.       Predicting the thermodynamics and kinetics parameters of small molecules, peptides and proteins binding to target proteins

3.1.6.       Understanding the molecular basis of single-stranded nucleic acids recognition in the presence of splicing modulators

3.1.7.       Predicting the thermodynamics and kinetic rate parameters of small molecule binding to an RNA aptamer

1. Conclusion
2. Acknowledgements
3. References

 

Chapter 16 Contact-Dynamics–Based Perturbation Analysis: Revealing Allosteric Communication Pathways in Proteins

1. Introduction
2. Historical Background
3. Beyond Classical Models: A Broader View of Allostery
4. Multi-Scale Views of Allostery
5. Methods for Determining Allostery
6. Experimental Methods
7. Computational Methods
8. Allostery Through the Lens of dCNA
9. Difference Contact Network Analysis (dCNA)
10. Core Concepts and Perspectives
11. Summary and Outlook
12. Acknowledgment
13. References

 

Chapter 17 Biophysical methods for studying the adamantyl amine-lipid-influenza A M2 protein channel system

1. 1.        Introduction

1.1.             Exploration of influenza A M2 channel function

1.2.             Biophysical methods used to study AM2 function

1. 2.        The pharmacological binding site of adamantyl amines in AM2
2. 3.        Interactions of adamantyl amines with AM2 WT channel

3.1.             MD simulations of adamantyl amine – AM2 WT complexes

3.2.             BFE calculations and ITC measurements against AM2TM WT

3.3.             Binding of rimantadine analogs to AM2TM WT

3.4.             Binding of rimantadine enantiomers to AM2TM WT 

1. 4.        Interactions of adamantyl amines with amantadine-resistant AM2 mutant channels

4.1.             Interactions with AM2CD and AM2TM V27A, L27F channels and structure-based drug design

4.2.             Interactions with AM2TM S31N channel

1. 5.        Interactions of AM2-adamantyl amine complexes with membranes  

5.1.             Interactions of adamantyl amines with AM2TM in single phospholipid membranes

5.2.             Interactions of AM2TM and AM2CD with cholesterol

1. 6.        Challenges for design of new antivirals
2. 7.        References

Index