Download Hi-Res ImageDownload to MS-PowerPointCite This:J. Chem. Inf. Model. 2023, 63, 11, 3227-3229
This publication is free to access through this site. Learn More
Editorial
Guidelines for Reporting Molecular Dynamics Simulations in JCIM PublicationsClick to copy article linkArticle link copied!
- Thereza A. Soares*
- Zoe Cournia
- Kevin Naidoo
- Rommie Amaro
- Habibah Wahab
- Kenneth M. Merz Jr.Kenneth M. Merz, Jr.More by Kenneth M. Merz, Jr.
This publication is licensed for personal use by The American Chemical Society.
Copyright © Published 2023 by American Chemical Society
Molecular dynamics (MD) simulation is a well-established methodology that studies the structure and dynamics of complex molecular systems at the microscopic level to provide mechanistic insights into the function of such systems. MD simulations capture the positions and motions of all particles comprising the system at every point in time based on Newtonian mechanics. Intermolecular interactions are calculated by a potential energy function (force field), and velocities are assigned through a Maxwell–Boltzmann distribution. Other conditions such as the initial conformation of the system, temperature, and pressure are assigned and controlled during the course of an MD simulation. MD simulations provide a robust approach to simultaneously probe temporal and spatial scales at the atomic to coarse-grained level, which are often difficult to obtain with other experimental techniques. It is also a standard component of the structural biologist’s toolbox, as the three-dimensional (3D) structures of biomolecules are models derived from physical signals obtained via experimental techniques (e.g., X-ray diffraction, cryo electron microscopy, and nuclear magnetic resonance spectroscopy), which need to be modeled and refined into a representative conformation expressed in atomic coordinates via MD simulations. However, as with every method, MD simulations come with limitations. Most often, errors in molecular simulations arise from inaccuracy in the chosen models to represent the system of interest, force field approximations, and/or insufficient sampling. Tackling these challenges via the improvement of algorithms and computer hardware that allows relevant timescale simulations to be run at a modest cost, as well as improvement of the accuracy of classical force-fields used to describe inter- and intramolecular interactions between atoms or particles, have made MD simulations more powerful and accurate over the past few years.
JCIM is one of the main venues for novel developments and exciting applications of MD methodologies associated with force field development, simulation protocols, and software. To enable the robustness and reproducibility of MD simulation studies published in JCIM, and in concert with our editorial on “Method and Data Sharing and Reproducibility of Scientific Results” (10.1021/acs.jcim.0c01389), herein, we provide guidelines for the “Methods” section of manuscripts submitted to JCIM that use MD to simulate biological, chemical, and physical systems (classical or hybrid, atomistic, or coarse-grained). The methodological section of papers presenting data from MD simulations should provide information accordingly with the following checklist.
Coordinates of the starting conformation
Click to copy section linkSection link copied!
If the starting coordinates are experimentally obtained, list the PDB code and appropriate references for the experimental data. Describe the methodologies to prepare the structures for MD simulations. Important points to consider to model a protein are, for instance, the addition of hydrogen atoms, the choice of protonation states with special attention to the choice of histidine protonation states, filling in missing side chains, loop regions and other missing molecular elements, solvation, salt concentration, inclusion of cofactors, substrates, etc.
If the starting coordinates are a computational model, provide a detailed, step-by-step, description of how the structural model was obtained. The homology modeling approach, which relies on the structural similarity of evolutionary related proteins, is often applied to generate structural models of a protein with unknown structure from the alignment with protein(s) of known structure(s). Because the correctness of sequence alignment and the choice of the best template structure(s) are critical for the quality of the model, authors should provide a detailed description of the main steps in homology modeling and use the Supporting Information section for additional data. Please describe the homologues used as template(s) for model building, the alignment of the target sequence to the template(s), model construction for the target protein based on the information from the alignment(s), and evaluation the structural quality of the model.
For AlphaFold2 protein models, the predicted local-distance difference test (pLDDT) should be >70 for all modeled residues. Disordered or long loops should be discarded, and caution should be taken to include native substrates, cofactors, and ligands that may be absent in the predicted structure. Moreover, important structural or functional elements should be cross-checked for validity with other experimental references (if available) to ensure that the proper conformation is being modeled.
Cartesian coordinates for the structural models, sequence alignments, topology files, Ramachandran plots, and other validation analyses should be made available as Supporting Information and/or deposited in a public data repository (see section entitled “Input/output files and data availability”).
JCIM will not consider straightforward applications of homology modeling combined with molecular docking methods to a single target system without adequate experimental validation.
Force field
Click to copy section linkSection link copied!
Provide information about the force field used in the simulations with attention to the parameter set version (major force fields have evolved as better computational and experimental data become available for parameter refinement). JCIM will not accept simulations with old force field versions unless a clear justification is provided. Specialized force fields should be used when available (e.g., for intrinsically disordered proteins). In the case of the reparametrization or development of new parameters compatible with a given force field, please provide benchmark data to support the actual need for reparameterization, proper validation of novel parameters against experimental or high-level QM data, full description of bonded and nonbonded potential parameters with the provision of atomic parameters and topology files as supplementary data, and provision of topology files if construction of new topologies was required.
Simulation conditions
Click to copy section linkSection link copied!
Describe the thermodynamic ensembles used, algorithms and simulation parameters used in the energy minimization, equilibration, and production phases. Please provide a detailed account of the following descriptors.
| a. | Energy minimization algorithm, MD integrator, time-step, length of simulations (sampling), number of replicas. | ||||
| b. | Constraint algorithms for bond lengths/water geometry if they are applied to solute and/or solvent. | ||||
| c. | Thermostat algorithm, target temperature, temperature coupling constant. | ||||
| d. | Barostat algorithm, target pressure, pressure coupling constant, compressibility, pressure treatment (e.g., isotropic, semi-isotropic, anisotropic). | ||||
| e. | Long-range interaction treatment (e.g., cutoffs, PME with spacing for the FFT grid and interpolation order, RF with dielectric constant) and approximation methods (e.g., switch, shift). | ||||
| f. | MD engine used with attention to the software version. | ||||
| g. | Simulation conditions (box shape and dimensions, total number of atoms, number of water molecules, implicit or explicit membranes or solvents, salt concentration, lipid composition, cofactors, metals, etc.). | ||||
| h. | Convergence metrics. | ||||
Authors should justify that the chosen model, force field, simulation conditions, sampling, and algorithms are suitable to study the specific phenomenon.
Replica simulations and convergence
Click to copy section linkSection link copied!
Molecular systems often present a rugged energy landscape with many local minima separated by high energy barriers. For this reason, small differences in the initial velocities, floating-point precision, or hardware can lead to different paths when starting from identical simulations. JCIM requires that studies reporting on MD simulations should be of suitable length and include at least three replica copies, ideally starting from different coordinates and velocites, with discussion of the statistical variance observed in the calculations. The latter should be accounted into proper context when discussing the data and conclusions from simulations. The random number generator seed values used for the replicas should be disclosed to ensure data reproducibility.
Enhanced sampling simulations
Click to copy section linkSection link copied!
MD simulations are not suitable to sample events occurring between free energy barriers such as conformational changes of biomolecular systems. In such cases enhanced sampling simulations should be used such as nonadaptive biasing potential methods (e.g., Gaussian-accelerated MD), adaptive-bias simulations (e.g., metadynamics), adaptive biasing force methods, replica exchange methods, etc. If such simulations are used, then the choice of enhanced sampling methodology, the parameters, and convergence criteria for the enhanced sampling method should be clearly described in the manuscript methods section and Supporting Information.
Hybrid quantum classical simulations
Click to copy section linkSection link copied!
The hybrid quantum classical (QM/MM) MD approach has opened up a means of inquiry probing the electronic, configurational, and conformational fundamentals of chemical events that underlie material and biochemical properties. While the method has become increasingly accessible through open-access packages, performing a QM/MM MD simulation to study charge transfer, electron excitations, or chemical reactions require the appropriate selection of a QM level of theory ranging from ab initio molecular orbital through valence bond, empirical valence bond, and semi empirical methods. The choice of a boundary method communicating between the quantum and classical components will greatly affect the accuracy of QM/MM MD simulations. High-end ab initio QM regions will provide accurate computations at the cost of slower simulation and subsequently poorer sampling. Including significant electronic role players in the QM region further determines the accuracy of the model and its predictions. In addition, a complete description of the classical MD methodology should be provided (see sections 1 to 3). Studies detailing careful consideration of these factors are necessary for reproducible results. Benchmarking spectral, structural, and configurational properties among others against experimental measures inspires confidence in a QM/MM MD study. Publications that blindly repeat methods varying only the chemical system without considering levels of theory boundary choice etc. make a less convincing case to readers and reviewers alike.
Input/output files and data availability
Click to copy section linkSection link copied!
JCIM now requires depositing your input files (including topologies, parameters, input files, and initial configurations) and output trajectory files in a public repository providing a DOI and provide the accession link in the manuscript. Some available repositories are Zenodo (https://zenodo.org/), Dryad (https://datadryad.org/), Nomad (https://nomad-repository.eu/), and Figshare (https://figshare.com/). If the full trajectories are too large, a selection of clustered structure representing the respective trajectories should be made available in the public repository.
Experimental validation
Click to copy section linkSection link copied!
Ideally, computational predictions should be tested using experimental work. While it is not always possible to perform experimental studies to verify working hypotheses, we require that manuscripts using MD simulations validate their results using available experimental data from the literature. Examples include NMR NOE constraints, HDX-MS data, SAXS curves, FRET distances, diffusion coefficients, electron density profiles, structure factors, bending moduli, and other structural, mechanical, or dynamical properties.
Author Information
Click to copy section linkSection link copied!
- Thereza A. Soares,
https://orcid.org/0000-0002-5891-6906
- Zoe Cournia,https://orcid.org/0000-0001-9287-364X
- Kevin Naidoo,https://orcid.org/0000-0002-9898-3708
- Rommie Amaro,https://orcid.org/0000-0002-9275-9553
- Habibah Wahab,https://orcid.org/0000-0002-8353-8679
Cited By
Click to copy section linkSection link copied!
Citation Statements
beta
Smart citations by scite.ai include citation statements extracted from the full text of the citing article. The number of the statements may be higher than the number of citations provided by ACS Publications if one paper cites another multiple times or lower if scite has not yet processed some of the citing articles.
Explore this article's citation statements on scite.ai
This article is cited by 15 publications.
- Andreas Vitalis, Steffen Winkler, Yang Zhang, Julian Widmer, Amedeo Caflisch. A FAIR-Compliant Management Solution for Molecular Simulation Trajectories. Journal of Chemical Information and Modeling 2025, 65
(5)
, 2443-2455. https://doi.org/10.1021/acs.jcim.4c01301
- Christian Jorgensen, Raleigh M. Linville, Ian Galea, Edward Lambden, Martin Vögele, Charles Chen, Evan P. Troendle, Fiorella Ruggiu, Martin B. Ulmschneider, Birgit Schiøtt, Christian D. Lorenz. Permeability Benchmarking: Guidelines for Comparing in Silico, in Vitro, and in Vivo Measurements. Journal of Chemical Information and Modeling 2025, 65
(3)
, 1067-1084. https://doi.org/10.1021/acs.jcim.4c01815
- Thanh T. Lai, Charles L. Brooks III. Accuracy and Reproducibility of Lipari-Szabo Order Parameters From Molecular Dynamics. The Journal of Physical Chemistry B 2024, 128
(44)
, 10813-10822. https://doi.org/10.1021/acs.jpcb.4c04895
- Shunzhou Wan, Agastya P. Bhati, Alexander D. Wade, Peter V. Coveney. Ensemble-Based Approaches Ensure Reliability and Reproducibility. Journal of Chemical Information and Modeling 2023, 63
(22)
, 6959-6963. https://doi.org/10.1021/acs.jcim.3c01654
- Shunzhou Wan, Agastya P. Bhati, Peter V. Coveney. Comparison of Equilibrium and Nonequilibrium Approaches for Relative Binding Free Energy Predictions. Journal of Chemical Theory and Computation 2023, 19
(21)
, 7846-7860. https://doi.org/10.1021/acs.jctc.3c00842
- Anju Choorakottayil Pushkaran, Alya A. Arabi. Accurate prediction of DNA-Intercalator binding energies: Ensemble of short or long molecular dynamics simulations?. International Journal of Biological Macromolecules 2025, 306 , 141408. https://doi.org/10.1016/j.ijbiomac.2025.141408
- Fatima Zahra Guerguer, Bouchra Rossafi, Oussama Abchir, Yasir S. Raouf, Dhabya Bakhit Albalushi, Abdelouahid Samadi, Samir Chtita, . Potential Azo-8-hydroxyquinoline derivatives as multi-target lead candidates for Alzheimer’s disease: An in-depth in silico study of monoamine oxidase and cholinesterase inhibitors. PLOS ONE 2025, 20
(1)
, e0317261. https://doi.org/10.1371/journal.pone.0317261
- Said El Rhabori, Marwa Alaqarbeh, Yassine El Allouche, Lhoucine Naanaai, Abdellah El Aissouq, Mohammed Bouachrine, Samir Chtita, Fouad Khalil. Exploring innovative strategies for identifying anti-breast cancer compounds by integrating 2D/3D-QSAR, molecular docking analyses, ADMET predictions, molecular dynamics simulations, and MM-PBSA approaches. Journal of Molecular Structure 2025, 1320 , 139500. https://doi.org/10.1016/j.molstruc.2024.139500
- Wouter Edeling, Maxime Vassaux, Yiming Yang, Shunzhou Wan, Serge Guillas, Peter V. Coveney. Global ranking of the sensitivity of interaction potential contributions within classical molecular dynamics force fields. npj Computational Materials 2024, 10
(1)
https://doi.org/10.1038/s41524-024-01272-z
- Michaela Neubergerová, Roman Pleskot, . Plant protein–lipid interfaces studied by molecular dynamics simulations. Journal of Experimental Botany 2024, 75
(17)
, 5237-5250. https://doi.org/10.1093/jxb/erae228
- Olena Lenchuk, Jochen Rohrer, Karsten Albe. Mo–Si Alloys Studied by Atomistic Computer Simulations Using a Novel Machine‐Learning Interatomic Potential: Thermodynamics and Interface Phenomena. Advanced Engineering Materials 2024, 26
(17)
https://doi.org/10.1002/adem.202302043
- Wei Qin, Xiang Guo, Jinwu Xiao, Zhihang Guan, Xianping Qiu, Fu-Quan Bai. Recent progress in model and dynamics research and its promotion to material development of energetic AlH3. International Journal of Hydrogen Energy 2024, 110 , 430-444. https://doi.org/10.1016/j.ijhydene.2025.02.165
- Sahaya Nadar, Maheshkumar R. Borkar, Tabassum Khan. Identification of potential focal adhesion kinase (FAK) inhibitors: a molecular modeling approach. Journal of Biomolecular Structure and Dynamics 2024, , 1-11. https://doi.org/10.1080/07391102.2024.2314266
- Massimiliano Tognolini, Alessio Lodola, Carmine Giorgio. Drug discovery: In silico dry data can bypass biological wet data?. British Journal of Pharmacology 2024, 181
(3)
, 340-344. https://doi.org/10.1111/bph.16266
- Ekaterina Shevchenko, Stefan Laufer, Antti Poso, Thales Kronenberger. Drug Design in Motion: Concepts and Applications of Classical Molecular Dynamics Simulations. 2024, 199-242. https://doi.org/10.1007/978-3-031-76718-0_8
Learn about these metrics
Article Views are the COUNTER-compliant sum of full text article downloads since November 2008 (both PDF and HTML) across all institutions and individuals. These metrics are regularly updated to reflect usage leading up to the last few days.
Citations are the number of other articles citing this article, calculated by Crossref and updated daily. Find more information about Crossref citation counts.
The Altmetric Attention Score is a quantitative measure of the attention that a research article has received online. Clicking on the donut icon will load a page at altmetric.com with additional details about the score and the social media presence for the given article. Find more information on the Altmetric Attention Score and how the score is calculated.
Recommended Articles
- This publication has no figures.
- This publication has no References.