Pith. sign in

REVIEW 3 major objections 6 minor 172 references

Evaluating Electrostatic Embedding MLIP/MM for Relative Binding Free Energy Calculations

T0 review · 3 major / 6 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read This paper reports that electrostatic embedding of a machine-learned ligand potential halves the TYK2 relative binding free-energy error in a production alchemical benchmark, while leaving four other targets unchanged.

desk verdict First production alchemical test of electrostatic-embedding MLIP/MM, with an honest target-dependent result and a confounded headline gain. read the letter →

arxiv 2608.13355 v1 pith:MFQEWW5R submitted 2026-08-13 physics.chem-ph physics.comp-ph

Stephen E. Farr , Gianni De Fabritiis This is my paper · ORCID
classification physics.chem-phphysics.comp-ph
keywords relativebindingfreeenergyelectrostaticembeddingmachinelearninginteratomicpotentialMLIP/MMRESPchargesalchemicalprotein-ligandparticlemeshEwald
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper tries to establish that electrostatic embedding of a machine-learned ligand potential inside a classical environment works in production alchemical relative binding free energy (RBFE) calculations, not just in single-point QM/MM-style tests. The authors train one neural network, AceFF-2-RESP-1, to simultaneously predict energies, forces, and RESP partial charges, and feed those charges into the short-range direct-space part of the particle mesh Ewald sum with Thole damping. Across five benchmark targets, this roughly halves the TYK2 $\Delta\Delta G$ root-mean-square error (from 0.86 to 0.45 kcal/mol against GAFF2 and from 0.77 to 0.45 kcal/mol against mechanical-embedding AceFF-1.0) and improves every accuracy and correlation metric on that target, while performing comparably to both baselines on CDK2, thrombin, p38, and JNK1. If true, this is the first demonstration that replacing static ligand charges with geometry-dependent ones can pay off in a production alchemical campaign, and that the benefit is target-dependent rather than universal.

What carries the argument

The mechanism is a multi-task neural network, named AceFF-2-RESP-1, built on the TensorNet2 architecture and trained on $10^6$ conformations from the AceFF dataset to jointly predict potential energy, atomic forces, and RESP partial charges. The predicted charges replace static AM1-BCC charges in the ligand–environment Coulomb interaction, but only in the short-range, direct-space term of the particle mesh Ewald sum, implemented as an additive correction against the fixed baseline charges; a Thole-damping factor on the MM electric field prevents polarization catastrophes during alchemical transformations. This design keeps the calculation affordable while letting the ligand charges respond to geometry throughout the $\lambda$ schedule.

What would settle it

Recompute the TYK2 alchemical edges with the dynamic charges coupled into the full particle mesh Ewald sum, including reciprocal-space gradients, or alternatively with static RESP charges replacing AM1-BCC while keeping the original MLIP/MM scheme; if the $\Delta\Delta G$ RMSE reverts toward 0.86 kcal/mol, the reported improvement is an artifact of the short-range-only approximation or of the charge definition change, not of geometry-dependent electrostatics.

Watch

Extended reading notes

Core claim

On the authors' own terms, the discovery is a demonstrated, reproducible target-dependent effect: electrostatic embedding improves RBFE accuracy on TYK2, where static charges appear to be the dominant error source, and is neutral on four more flexible or complex targets. The improvement is consistent across all eight metrics and both baselines, with the TYK2 $\Delta\Delta G$ RMSE falling by a factor of 1.7–2.2. The paper also shows that standard single-molecule energy, force, and charge benchmarks do not predict which targets benefit: TYK2 is the only clear winner and also has the lowest force error on the 650-molecule ligand benchmark, but that pattern does not hold for the other four targets. This is the first assessment of an electrostatic-embedding MLIP/MM scheme against experimental binding affinities on a standard congeneric-series benchmark.

Load-bearing premise

The load-bearing premise is that the long-range, reciprocal-space part of the electrostatic interaction differs negligibly between the dynamic RESP charges and the static AM1-BCC baseline, so correcting only the short-range direct-space term is enough; if that premise fails, every free-energy correction in the paper is systematically biased.

Editorial extensions

If this is right

  • If the scheme is correct, replacing static AM1-BCC ligand charges with predicted conformation-dependent RESP charges reduces TYK2 $\Delta\Delta G$ RMSE from 0.86 to 0.45 kcal/mol against GAFF2 and from 0.77 to 0.45 kcal/mol against AceFF-1.0, improving all eight reported accuracy and correlation metrics.
  • On CDK2, thrombin, p38, and JNK1, electrostatic embedding matches both baselines within bootstrap intervals, so the method should not be expected to improve every target.
  • Joint training of energy, forces, and charges costs accuracy on standard benchmarks: AceFF-2-RESP-1 trails the single-task AceFF-2 on the Wiggle150 and torsion benchmarks while staying close on forces.
  • Standard single-molecule force, energy, and charge errors do not predict which targets will benefit, so model selection for RBFE cannot rely on those benchmarks alone.
  • The alchemical workflow is numerically stable across the full $\lambda$ schedule when Thole damping is applied, meaning the scheme is deployable in production campaigns.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Not claimed by the paper: the TYK2 gain could come from using RESP rather than AM1-BCC charges even without the charges being dynamic; a control experiment with static RESP charges in the classical baseline would separate these two variables.
  • If target rigidity is what gates the benefit, the natural testable extension is to run the same protocol on other rigid congeneric series; the authors' own explanation predicts gains where the pocket is rigid and static charges dominate.
  • The short-range-only PME approximation is a sensitivity that could be checked by reweighting the existing trajectories with full Ewald electrostatics, at low cost, before any new simulation is run.
  • The neutral results on four targets suggest pragmatic deployment: use electrostatic embedding after a rapid diagnostic of whether the ligand series has charge-sensitive interactions, rather than as a default replacement for mechanical embedding.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 6 minor

Summary. The paper trains a TensorNet2-based machine learning potential, AceFF-2-RESP-1, that jointly predicts energies, forces, and conformation-dependent RESP partial charges, and embeds it electrostatically in an MLIP/MM scheme following Semelak et al. The method is tested in production alchemical RBFE calculations on five targets from the Wang et al. benchmark set, with three replicates per edge and the QuantumBind-RBFE protocol. The central results are that electrostatic embedding improves essentially every reported accuracy and correlation metric for TYK2 (ΔΔG RMSE 0.86→0.45 kcal/mol against GAFF2 and 0.77→0.45 against the mechanical-embedding AceFF-1.0 baseline) while giving neutral results for CDK2, thrombin, p38, and JNK1. The paper also reports single-molecule benchmarks (Wiggle150, Sellers torsion scans, Schrödinger ligand benchmark) and finds that static benchmarks do not predict the target-dependent RBFE outcome.

Significance. If the attribution to electrostatic embedding is accepted, this is the first production-scale alchemical RBFE campaign with electrostatic-embedding MLIP/MM, and the TYK2 result is a notable, target-dependent improvement with a plausible mechanistic story. The paper has concrete strengths: the target set and edges were fixed in advance by prior work, three independent replicates per edge were run, bootstrap confidence intervals are reported, the model weights and modified software are publicly released, and the Limitations section is unusually candid about untested embedding parameters and the short-range PME hypothesis. The main weakness is that the experimental design does not isolate the electrostatic embedding from changes in the base potential and the charge model, so the paper's headline causal claim is not uniquely supported by the data. The TYK2 gain remains a real empirical result for the combined AceFF-2-RESP-1 electrostatically embedded model, but the specific mechanism is not established by the current comparison.

major comments (3)
  1. [Introduction; Table 2] The claim that fixing systems, edges, protocol, and sampling 'lets us separate the effect of the electrostatic coupling from the effect of the underlying potential' is not supported by the design, because AceFF-2-RESP-1 differs from the mechanical-embedding baseline AceFF-1.0 in three ways at once: the embedding (electrostatic vs mechanical), the base network architecture and training (TensorNet2 with 2 interaction layers and 128-dimensional embedding vs the older AceFF-1.0 model), and the charge model (RESP vs AM1-BCC). A mechanical-embedding run using the same AceFF-2-RESP-1 network with fixed charges is a necessary control to attribute the TYK2 improvement specifically to electrostatic coupling. Without it, the improvement could plausibly stem from the better base potential energy surface of AceFF-2 or from the switch to RESP charges. The paper should either add such a control or substantially soften the causal language throughout the abstract, results, and conclusion.
  2. [Methods: Short-Range PME Modification and Thole Damping; Limitations] The approximation that only the short-range direct-space part of the PME sum uses the dynamic MLIP charges is justified by the hypothesis that 'long-range electrostatics vary negligibly between RESP and AM1-BCC parameterizations.' This hypothesis is load-bearing for the reported free energies, since the electrostatic correction in Eq. (6) is only a short-range difference term. The manuscript does not test the sensitivity of any RBFE result to this approximation, despite the Limitations section correctly flagging it. A concrete test would be to compute full-PME dynamic-charge energies (or at least the reciprocal-space contribution) for representative end-state and intermediate lambda configurations on one target, and report the magnitude of the neglected long-range term relative to the observed TYK2 free-energy changes.
  3. [Limitations; Results: RBFE benchmarks] The Limitations paragraph lists untested embedding parameters (polarizabilities, epsilon, Thole exponent) and the short-range PME hypothesis, but does not list the missing same-potential mechanical-embedding control. As a result, the interpretation that 'static ligand charges were a primary limiting factor' for TYK2 and the conclusion that 'electrostatic embedding roughly halved the error' overstate the evidence. The authors should either run the missing control or reframe the main claim as an evaluation of the complete AceFF-2-RESP-1 electrostatically embedded scheme against two baselines, without asserting that the improvement is caused by electrostatic embedding per se.
minor comments (6)
  1. [Abstract; Section: Quantum Chemical Dataset and Neural Network Training] The dataset size '10 6' should be typeset as 10^6; the missing superscript appears in both the abstract and the Methods section.
  2. [Throughout] The model name appears inconsistently as AceFF-2-RESP-1 and AceFF-2-resp-1 (Table 1); please unify to the capitalization used in the title and abstract.
  3. [Tables 1-3] The text contains typographical spacing errors in table captions ('T able 1', 'T able 2', 'T able 3'); these should be corrected to 'Table'.
  4. [Introduction] The phrase 'hybrid MLIP/MM (or MLIP/MM) schemes' is redundant; the intended distinction between 'MLIP/MM' and 'NNP/MM' should be clarified or the parenthetical removed.
  5. [Results: Figure 2] Panel (b) shows a parity plot for predicted vs. reference RESP charges that the text calls 'tight, linear,' but no correlation coefficient or R^2 is reported; adding the numerical value would make the claim verifiable.
  6. [Methods: Short-Range PME Modification and Thole Damping, Eq. (7)] The notation 'ˆ⃗ rij' is not defined before use; please define the unit vector symbol and the meaning of the over-hat arrow notation explicitly.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the model and all embedding parameters are fixed a priori, and the RBFE benchmark is external.

full rationale

The paper's derivation chain is self-contained and does not reduce to its inputs. AceFF-2-RESP-1 was trained on a separate 10^6-conformer subset of the AceFF dataset with QM energies, forces, and RESP charges; no experimental RBFE values and no target-specific data from TYK2, CDK2, thrombin, p38, or JNK1 entered the training or the embedding-parameter choice. The embedding parameters (ANI-MBIS/Thole polarizabilities, epsilon = 2, Thole exponent a = 1.3, RESP charge model) are fixed from prior literature and not fitted to the benchmark. The five targets and alchemical edges were fixed in advance by the prior QuantumBind-RBFE study and reused unchanged. The predictive claim is therefore not a fitted input renamed as a prediction. The comparison against the authors' own AceFF-1.0 baseline and AceFF dataset is a self-citation, but it is an appropriate control and the final metric is agreement with experimental binding affinities, so the central conclusion does not depend on the self-citation for its content. The paper's stated limitation that only short-range direct-space PME is corrected by dynamic charges, and the confound that AceFF-2-RESP-1 differs from AceFF-1.0 in both base network and charge model, are correctness/attribution concerns, not circularity: they do not make the TYK2 improvement true by construction. No equation in the paper is defined in terms of the quantity it is said to predict, and no benchmark value is reused as an input to the model.

Assumptions & free parameters 4 free parameters · 6 assumptions · 0 invented entities

The central evaluation rests on a standard MLIP/MM energy decomposition, a short-range-only electrostatic correction with an untested long-range assumption, and fixed embedding hyperparameters whose sensitivity is not characterized. No new physical entities are introduced.

free parameters (4)
  • Loss weights w_e, w_f, w_q = Not reported
    Joint training loss in Eq. 1 balances energy, force, and charge errors; weights chosen by hand without a reported sensitivity analysis, so the energy-charge trade-off is not characterized.
  • Effective dielectric constant epsilon = 2
    Fixed at 2 in the polarization term of Eq. 5; the paper states it was not optimized and sensitivity was not tested.
  • Thole damping exponent a = 1.3
    Dimensionless damping exponent in Eq. 8; fixed, and the paper states the sensitivity of free energies to this value was not checked.
  • Atomic polarizabilities alpha_i = From ANI-MBIS table and van Duijnen-Swart
    Taken unchanged from prior literature; the paper notes these were not re-optimized for the alchemical setting.
assumptions (6)
  • domain assumption Long-range electrostatics differ negligibly between RESP and AM1-BCC charges
    Methods, Short-Range PME Modification: only the short-range direct-space term is corrected and the reciprocal-space difference is hypothesized to be negligible; restated in Limitations.
  • domain assumption RESP charges predicted by the network remain commensurable with the AMBER-family environment at every lambda value
    Introduction and Methods: RESP is chosen for compatibility with AMBER-family parameterization; no validation that charges stay consistent along the alchemical path beyond the compatibility argument.
  • domain assumption Thole damping is negligible at physical end states
    Methods, Thole Damping: damping acts only in alchemical intermediate states; the physical end-state polarization is assumed unaffected, but this is not directly tested.
  • domain assumption The MLIP/MM energy decomposition with additive electrostatic coupling is a valid representation of the ligand-environment interaction
    Eq. 2 and Eq. 3 in Methods are adopted from Semelak et al.; the paper evaluates the scheme but does not re-derive or benchmark the decomposition itself against QM/MM in this setting.
  • domain assumption The linear-response induction relation E_distortion = -0.5 E_pol holds
    Methods, after Eq. 5: follows a standard linear-response argument cited to Stone; the fixed factor is not independently verified here.
  • domain assumption The five Wang et al. targets and their alchemical edges are representative enough to draw general conclusions
    Targets fixed by the prior QuantumBind-RBFE study; with improvement on only one target, the conclusion that static charges are the dominant error source for TYK2 is a single-target inference.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Evaluating Electrostatic Embedding MLIP/MM for Relative Binding Free Energy Calculations." pith.science (2026). https://pith.science/paper/MFQEWW5R

@misc{pith2026260813355,
  author       = {Pith},
  title        = {Pith review of: Evaluating Electrostatic Embedding MLIP/MM for Relative Binding Free Energy Calculations},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MFQEWW5R}},
  note         = {Machine review of arXiv:2608.13355}
}
abstract

Alchemical relative binding free energy (RBFE) calculations are limited by the fixed-charge approximation of classical force fields. Hybrid machine learning interatomic potential/molecular mechanics (MLIP/MM) schemes correct ligand strain, but under mechanical embedding still describe ligand--environment electrostatics with static point charges. Electrostatic embedding schemes coupling machine-learned charges to the MM environment have been proposed and validated against QM/MM for simple systems, but not tested in a production alchemical workflow. We take the electrostatic embedding scheme of Semelak et al.\ and evaluate it on protein--ligand RBFE. We trained a TensorNet2 model, \texttt{AceFF-2-RESP-1}, on $10^{6}$ conformations from the AceFF dataset, jointly predicting energies, forces and Restrained Electrostatic Potential (RESP) charges. We chose RESP over MBIS for commensurability with the AMBER-family force field it couples to. The predicted charges enter the short-range direct-space part of the particle mesh Ewald sum, with Thole damping to prevent polarization catastrophes during alchemical transformations. We tested the scheme across five targets from the Wang et al.\ benchmark set, fixed in advance by a prior study, with three replicates per edge and matched protocols. Electrostatic embedding improved every accuracy and correlation metric for TYK2 ($\Delta\Delta G$ RMSE $0.86 \rightarrow 0.45$~kcal/mol against GAFF2), but performed comparably to the classical and mechanical-embedding baselines for CDK2, thrombin, p38 and JNK1. Standard single-molecule energy and charge benchmarks were not good predictors of this target-dependent outcome. TYK2 combined good $\Delta\Delta G$ accuracy with the lowest force error on the Schr\"odinger benchmark, but this pattern did not hold for the other targets.

Figures

Figures reproduced from arXiv: 2608.13355 by the authors.

Figure 1
Figure 1. Sellers et al. torsion scan benchmark. 77 The orange lines are the median value. The methods are ordered from left to right by median MAE compared to the coupled cluster reference data. We performed relaxed torsion scans as done in. 61,79 Out-of-Distribution Generalizability: Schr¨odinger Ligand Bench￾mark To assess generalization to complex, drug-like chemical space, we evaluated our model on the Schr¨odinger ligan… view at source ↗
Figure 2
Figure 2. Multi-objective model evaluation on the Schr¨odinger ligand benchmark (650 conform￾ers). (a) Comparative force error distributions between AceFF-2-RESP-1 and the single-task AceFF-2 reference. (b) Parity plot showing correlation between MLIP-predicted, conformation￾dependent partial charges and reference ωB97M-V/def2-TZVPPD RESP charges. Relative Binding Free Energy (RBFE) Benchmarks We benchmarked the AceFF-2-RESP-… view at source ↗
Figure 3
Figure 3. Calculated vs experimental relative binding free energies (∆∆G) on alchemical edges. Each marker shows the mean from the three repeats and the error bars are the standard deviation. RMSE, MAE, Kendall τ and Spearman ρ metrics are annotated in the top-left of each panel. 14 [PITH_FULL_IMAGE:figures/full_fig_p014_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Calculated vs experimental absolute binding free energies (∆G). Each marker shows the mean from the three repeats and the error bars are the standard deviation. RMSE, MAE, Kendall τ and Spearman ρ metrics are annotated in the top-left of each panel. 15 [PITH_FULL_IMAG…
Figure 5
Figure 5. Figure 5: Correlation between per-target Schr¨odinger benchmark error metrics (force, energy, and RESP charge MAE, reported as mean ± SD) and downstream mean absolute error in alchemical ∆∆G predictions across the five benchmark systems. 17 [PITH_FULL_IMAGE:figures/full_fig_p01…
Figure 6
Figure 6. Figure 6: Variation of predicted partial charges for TYK2 ligand ejm54. A 2D depiction of the molecule is shown, the atoms are colored by the mean charge over the course of the simulation and the range of partial charge variation is depicted by the size of the shaded region and …

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

172 extracted references · 54 canonical work pages

  1. [1]

    and Cieplak, Piotr and Cornell, Wendy and Kollman, Peter A

    Bayly, Christopher I. and Cieplak, Piotr and Cornell, Wendy and Kollman, Peter A. , title =. The Journal of Physical Chemistry , volume =. 1993 , doi =

    1993

  2. [2]

    and Jack, David B

    Jakalian, Araz and Bush, Bruce L. and Jack, David B. and Bayly, Christopher I. , title =. Journal of Computational Chemistry , volume =. 2000 , doi =

    2000

  3. [3]

    and Bayly, Christopher I

    Jakalian, Araz and Jack, David B. and Bayly, Christopher I. , title =. Journal of Computational Chemistry , volume =. 2002 , doi =

    2002

  4. [4]

    , title =

    Verstraelen, Toon and Vandenbrande, Steven and Heidar-Zadeh, Farnaz and Vanduyfhuys, Louis and Van Speybroeck, Veronique and Waroquier, Michel and Ayers, Paul W. , title =. Journal of Chemical Theory and Computation , volume =. 2016 , doi =

    2016

  5. [5]

    and Cieplak, Piotr and Bayly, Christopher I

    Cornell, Wendy D. and Cieplak, Piotr and Bayly, Christopher I. and Gould, Ian R. and Merz, Kenneth M. and Ferguson, David M. and Spellmeyer, David C. and Fox, Thomas and Caldwell, James W. and Kollman, Peter A. , title =. Journal of the American Chemical Society , volume =. 1995 , doi =

    1995

  6. [6]

    Journal of Chemical Theory and Computation , volume =

    Zinovjev, Kirill , title =. Journal of Chemical Theory and Computation , volume =. 2023 , doi =

    2023

  7. [7]

    and Montagud Andreu, Rub

    Zinovjev, Kirill and Hedges, Lester O. and Montagud Andreu, Rub. emle-engine: A Flexible Electrostatic Machine Learning Embedding Package for Multiscale Molecular Dynamics Simulations , journal =. 2024 , doi =

    2024

  8. [8]

    Enhancing Electrostatic Embedding for

    Morado, Jo. Enhancing Electrostatic Embedding for. Journal of Chemical Theory and Computation , volume =. 2025 , doi =

    2025

Show all 172 references
  1. [9]

    and Gonz

    Grassano, Juan Santiago and Pickering, Ignacio and Roitberg, Adrian E. and Gonz. Assessment of Embedding Schemes in a Hybrid Machine Learning/Classical Potentials. Journal of Chemical Information and Modeling , volume =. 2024 , doi =

    2024

  2. [10]

    Machine Learning in

    B. Machine Learning in. Journal of Chemical Theory and Computation , volume =. 2021 , doi =

    2021

  3. [11]

    Learning Atomic Multipoles: Prediction of the Electrostatic Potential with Equivariant Graph Neural Networks , journal =

    Th. Learning Atomic Multipoles: Prediction of the Electrostatic Potential with Equivariant Graph Neural Networks , journal =. 2022 , doi =

    2022

  4. [12]

    Machine Learning of Solvent Effects on Molecular Spectra and Reactions , journal =

    Gastegger, Michael and Sch. Machine Learning of Solvent Effects on Molecular Spectra and Reactions , journal =. 2021 , doi =

    2021

  5. [13]

    and Meuwly, Markus , title =

    Unke, Oliver T. and Meuwly, Markus , title =. Journal of Chemical Theory and Computation , volume =. 2019 , doi =

    2019

  6. [14]

    Karwounopoulos, Johannes and Bieniek, Mateusz and Wu, Zhiyi and Baskerville, Adam L. and K. Evaluation of Machine Learning/Molecular Mechanics End-State Corrections with Mechanical Embedding to Calculate Relative Protein--Ligand Binding Free Energies , journal =. 2025 , doi =

    2025

  7. [15]

    The Journal of Physical Chemistry B , volume =

    Karwounopoulos, Johannes and Wu, Zhiyi and Tkaczyk, Sara and Wang, Shuzhe and Baskerville, Adam and Ranasinghe, Kavindri and Langer, Thierry and Wood, Geoffrey and Wieder, Marcus and Boresch, Stefan , title =. The Journal of Physical Chemistry B , volume =. 2024 , doi =

    2024

  8. [16]

    Journal of Chemical Theory and Computation , volume =

    Tkaczyk, Sara and Langer, Thierry and Wieder, Marcus and Rizzi, Andrea and Boresch, Stefan , title =. Journal of Chemical Theory and Computation , volume =. 2026 , doi =

    2026

  9. [17]

    Journal of Molecular Biology , volume =

    Warshel, Arieh and Levitt, Michael , title =. Journal of Molecular Biology , volume =. 1976 , doi =

    1976

  10. [18]

    and Bash, Paul A

    Field, Martin J. and Bash, Paul A. and Karplus, Martin , title =. Journal of Computational Chemistry , volume =. 1990 , doi =

    1990

  11. [19]

    Angewandte Chemie International Edition , volume =

    Senn, Hans Martin and Thiel, Walter , title =. Angewandte Chemie International Edition , volume =. 2009 , doi =

    2009

  12. [20]

    Thole, B. T. , title =. Chemical Physics , volume =. 1981 , doi =

    1981

  13. [21]

    and Swart, Marcel , title =

    van Duijnen, Piet Th. and Swart, Marcel , title =. The Journal of Physical Chemistry A , volume =. 1998 , doi =

    1998

  14. [22]

    , title =

    Stone, Anthony J. , title =. 2013 , doi =

    2013

  15. [23]

    and Wu, Chuanjie and Ren, Pengyu and Pande, Vijay S

    Ponder, Jay W. and Wu, Chuanjie and Ren, Pengyu and Pande, Vijay S. and Chodera, John D. and Schnieders, Michael J. and Haque, Imran and Mobley, David L. and Lambrecht, Daniel S. and DiStasio, Robert A. and Head-Gordon, Martin and Clark, Gary N. I. and Johnson, Margaret E. and...

    2010

  16. [24]

    and Huang, Jing and Roux, Beno

    Lemkul, Justin A. and Huang, Jing and Roux, Beno. An Empirical Polarizable Force Field Based on the Classical Drude Oscillator Model: Development History and Recent Applications , journal =. 2016 , doi =

    2016

  17. [25]

    and Qi, Rui and Walker, Brandon D

    Jing, Zhifeng and Liu, Chengwen and Cheng, Sara Y. and Qi, Rui and Walker, Brandon D. and Piquemal, Jean-Philip and Ren, Pengyu , title =. Annual Review of Biophysics , volume =. 2019 , doi =

    2019

  18. [26]

    The Journal of Chemical Physics , volume =

    Darden, Tom and York, Darrin and Pedersen, Lee , title =. The Journal of Chemical Physics , volume =. 1993 , doi =

    1993

  19. [27]

    and Darden, Tom and Lee, Hsing and Pedersen, Lee G

    Essmann, Ulrich and Perera, Lalith and Berkowitz, Max L. and Darden, Tom and Lee, Hsing and Pedersen, Lee G. , title =. The Journal of Chemical Physics , volume =. 1995 , doi =

    1995

  20. [28]

    and Pal, Rajat K

    Azimi, Solmaz and Khuttan, Sheenam and Wu, Joe Z. and Pal, Rajat K. and Gallicchio, Emilio , title =. Journal of Chemical Information and Modeling , volume =. 2022 , doi =

    2022

  21. [29]

    and Azimi, Solmaz and Khuttan, Sheenam and Deng, Nanjie and Gallicchio, Emilio , title =

    Wu, Joe Z. and Azimi, Solmaz and Khuttan, Sheenam and Deng, Nanjie and Gallicchio, Emilio , title =. Journal of Chemical Theory and Computation , volume =. 2021 , doi =

    2021

  22. [30]

    and Chodera, John D

    Shirts, Michael R. and Chodera, John D. , title =. The Journal of Chemical Physics , volume =. 2008 , doi =

    2008

  23. [31]

    , title =

    Tan, Zhiqiang and Gallicchio, Emilio and Lapelosa, Mauro and Levy, Ronald M. , title =. The Journal of Chemical Physics , volume =. 2012 , doi =

    2012

  24. [32]

    and Bayly, Christopher I

    Hahn, David F. and Bayly, Christopher I. and Boby, Melissa L and Bruce Macdonald, Hannah E. and Chodera, John D. and Gapsys, Vytautas and Mey, Antonia S. J. S. and Mobley, David L. and Benito, Laura Perez and Schindler, Christina E. M. and Tresadern, Gary and Warren, Gregory L...

  25. [33]

    Schindler, Christina E. M. and Baumann, Hannah and Blum, Andreas and B. Large-Scale Assessment of Binding Free Energy Calculations in Active Drug Discovery Projects , journal =. 2020 , doi =

    2020

  26. [34]

    Journal of Chemical Information and Modeling , volume =

    Cournia, Zoe and Allen, Bryce and Sherman, Woody , title =. Journal of Chemical Information and Modeling , volume =. 2017 , doi =

    2017

  27. [35]

    and Beuming, Thijs and Pearlman, David A

    Cournia, Zoe and Allen, Bryce K. and Beuming, Thijs and Pearlman, David A. and Radak, Brian K. and Sherman, Woody , title =. Journal of Chemical Information and Modeling , volume =. 2020 , doi =

    2020

  28. [36]

    Mey, Antonia S. J. S. and Allen, Bryce K. and Macdonald, Hannah E. Bruce and Chodera, John D. and Hahn, David F. and Kuhn, Maximilian and Michel, Julien and Mobley, David L. and Naden, Levi N. and Prasad, Samarjeet and Rizzi, Andrea and Scheen, Jenke and Shirts, Michael R. and...

    2020

  29. [37]

    and Aldeghi, Matteo and Gapsys, Vytautas and Ntekoumes, Dimitris and Bosisio, Stefano and Papadourakis, Michail and Henriksen, Niel M

    Rizzi, Andrea and Jensen, Travis and Slochower, David R. and Aldeghi, Matteo and Gapsys, Vytautas and Ntekoumes, Dimitris and Bosisio, Stefano and Papadourakis, Michail and Henriksen, Niel M. and de Groot, Bert L. and Cournia, Zoe and Dickson, Alex and Michel, Julien and Gilso...

    2020

  30. [38]

    and Le Grand, Scott and Walker, Ross C

    Hopkins, Chad W. and Le Grand, Scott and Walker, Ross C. and Roitberg, Adrian E. , title =. Journal of Chemical Theory and Computation , volume =. 2015 , doi =

    2015

  31. [39]

    and Doerr, Stefan and Mirarchi, Antonio and Saban

    Farr, Stephen E. and Doerr, Stefan and Mirarchi, Antonio and Saban. arXiv preprint arXiv:2601.00581 , year =. doi:10.48550/arXiv.2601.00581 , url =. 2601.00581 , archivePrefix =

    doi

  32. [40]

    Multiscale Neural Network Potential with Anisotropic Message Passing for the Fast and Accurate Simulation of Protein Dynamics and Enzymatic Reactions , journal =

    Th. Multiscale Neural Network Potential with Anisotropic Message Passing for the Fast and Accurate Simulation of Protein Dynamics and Enzymatic Reactions , journal =. 2026 , doi =

    2026

  33. [41]

    2025 , howpublished =

    Acellera , title =. 2025 , howpublished =

    2025

  34. [42]

    2023 , keywords =

    Scientific Data , author =. 2023 , keywords =. doi:10.1038/s41597-022-01882-6 , abstract =

    2023 doi

  35. [43]

    and Wagen, Corin C

    Mann, Elias L. and Wagen, Corin C. and Vandezande, Jonathon E. and Wagen, Arien M. and Schneider, Spencer C. , year =. Egret-1:. doi:10.48550/ARXIV.2504.20955 , abstract =

    doi arXiv

  36. [44]

    Journal of Chemical Theory and Computation , author =

    Nutmeg and. Journal of Chemical Theory and Computation , author =. 2024 , pages =. doi:10.1021/acs.jctc.4c00794 , language =

    2024 doi

  37. [45]

    Journal of Chemical Theory and Computation , author =

    Broadening the. Journal of Chemical Theory and Computation , author =. 2025 , pages =. doi:10.1021/acs.jctc.4c01625 , language =

    2025 doi

  38. [46]

    Journal of Physics: Condensed Matter , author =

    The atomic simulation environment—a. Journal of Physics: Condensed Matter , author =. 2017 , pages =. doi:10.1088/1361-648X/aa680e , number =

    2017 doi

  39. [47]

    Brew, R. R. and Nelson, I. A. and Binayeva, M. and Nayak, A. S. and Simmons, W. J. and Gair, J. J. and Wagen, C. C. , journal=

  40. [48]

    2025 , pages =

    Journal of Chemical Information and Modeling , author =. 2025 , pages =. doi:10.1021/acs.jcim.5c00033 , language =

    2025 doi

  41. [49]

    The Journal of Chemical Physics , author =

    Recent developments in the PySCF program package , volume =. The Journal of Chemical Physics , author =. 2020 , pages =. doi:10.1063/5.0006074 , abstract =

    2020 doi

  42. [50]

    2016 , pages =

    The Journal of Chemical Physics , author =. 2016 , pages =. doi:10.1063/1.4952647 , language =

    2016 doi

  43. [51]

    Neumann, Mark and Gin, James and Rhodes, Benjamin and Bennett, Steven and Li, Zhiyi and Choubisa, Hitarth and Hussey, Arthur and Godwin, Jonathan , month = oct, year =. Orb:. doi:10.48550/arXiv.2410.22570 , abstract =

    doi arXiv

  44. [52]

    and others , month = jun, year =

    Wood, Brandon M. and others , month = jun, year =. doi:10.48550/arXiv.2506.23971 , publisher =

    doi

  45. [53]

    and Kuryla, Domantas and Hart, Joseph and Kasoar, Elliott and Elena, Alin M

    Batatia, Ilyes and Baldwin, William J. and Kuryla, Domantas and Hart, Joseph and Kasoar, Elliott and Elena, Alin M. and Moore, Harry and Gawkowski, Miko. 2026 , note =. doi:10.48550/arXiv.2602.19411 , abstract =

    2026 doi

  46. [54]

    Batatia, I. and Kov. Advances in Neural Information Processing Systems , volume=

  47. [55]

    Nature Communications , author =

    E(3)-equivariant graph neural networks for data-efficient and accurate interatomic potentials , volume =. Nature Communications , author =. 2022 , keywords =. doi:10.1038/s41467-022-29939-5 , abstract =

    2022 doi

  48. [56]

    2025 , eprint=

    The Open Molecules 2025 (OMol25) Dataset, Evaluations, and Models , author=. 2025 , eprint=

    2025

  49. [57]

    Physical Review Letters , volume =

    Generalized Neural-Network Representation of High-Dimensional Potential-Energy Surfaces , author =. Physical Review Letters , volume =. 2007 , doi =

    2007

  50. [58]

    Physical Review Letters , volume =

    Gaussian Approximation Potentials: The Accuracy of Quantum Mechanics, without the Electrons , author =. Physical Review Letters , volume =. 2010 , doi =

    2010

  51. [59]

    Chemical Science , volume =

    ANI-1: an extensible neural network potential with DFT accuracy at force field computational cost , author =. Chemical Science , volume =. 2017 , doi =

    2017

  52. [60]

    Chemical Reviews , volume =

    Machine Learning Force Fields , author =. Chemical Reviews , volume =. 2021 , doi =

    2021

  53. [61]

    ChemRxiv , year =

    Plé, Thomas and Adjoua, Olivier and El Khoury, Léa and Alferkh, Lina and Jing, Zhifeng and Villot, Corentin and Benali, Anouar and Posenitskiy, Evgeny and Lagardère, Louis and Piquemal, Jean-Philip , title =. ChemRxiv , year =. doi:10.26434/chemrxiv-2025-f1hgn-v5 , URL =

    doi

  54. [62]

    and Gokey, Trevor and Dotson, David L

    Behara, Pavan Kumar and Jang, Hyesu and Horton, Joshua T. and Gokey, Trevor and Dotson, David L. and Boothroyd, Simon and Bayly, Christopher I. and Cole, Daniel J. and Wang, Lee-Ping and Mobley, David L. , title =. The Journal of Physical Chemistry B , volume =. 2024 , doi =

    2024

  55. [63]

    2025 , note =

    OrbMol , howpublished =. 2025 , note =

    2025

  56. [64]

    2025 , eprint=

    Orb-v3: atomistic simulation at scale , author=. 2025 , eprint=

    2025

  57. [65]

    2024 , eprint=

    Orb: A Fast, Scalable Neural Network Potential , author=. 2024 , eprint=

    2024

  58. [66]

    , title =

    Riniker, Sereina and Landrum, Gregory A. , title =. Journal of Chemical Information and Modeling , volume =. 2015 , doi =

    2015

  59. [67]

    Journal of Chemical Information and Modeling , volume =

    Friedrich, Nils-Ole and Meyder, Agnes and de Bruyn Kops, Christina and Sommer, Kai and Flachsenberg, Florian and Rarey, Matthias and Kirchmair, Johannes , title =. Journal of Chemical Information and Modeling , volume =. 2017 , doi =

    2017

  60. [68]

    doi:10.1145/3620665.3640366 , month = apr, publisher =

    Ansel, Jason and Yang, Edward and He, Horace and Gimelshein, Natalia and Jain, Animesh and Voznesensky, Michael and Bao, Bin and Bell, Peter and Berard, David and Burovski, Evgeni and Chauhan, Geeta and Chourdia, Anjali and Constable, Will and Desmaison, Alban and DeVito, Zach...

  61. [69]

    PubChem Compound Summary for CID 5291, Imatinib , year =

  62. [70]

    npj Computational Materials , author =

    Latent. npj Computational Materials , author =. 2025 , pages =. doi:10.1038/s41524-025-01577-7 , abstract =

    2025 doi

  63. [71]

    Nature Communications , author =

    A fourth-generation high-dimensional neural network potential with accurate electrostatics including non-local charge transfer , volume =. Nature Communications , author =. 2021 , pages =. doi:10.1038/s41467-020-20427-2 , abstract =

    2021 doi

  64. [72]

    Journal of Chemical Theory and Computation , author =

    Advancing. Journal of Chemical Theory and Computation , author =. 2025 , pages =. doi:10.1021/acs.jctc.4c01792 , language =

    2025 doi

  65. [73]

    2025 , pages =

    Nucleic Acids Research , author =. 2025 , pages =. doi:10.1093/nar/gkae1059 , abstract =

    2025 doi

  66. [74]

    Froitzheim, Thomas and Müller, Marcel and Hansen, Andreas and Grimme, Stefan , month = jun, year =. g-. doi:10.26434/chemrxiv-2025-bjxvt , abstract =

    2025 doi

  67. [75]

    2019 , pages =

    Journal of Chemical Theory and Computation , author =. 2019 , pages =. doi:10.1021/acs.jctc.8b01176 , language =

    2019 doi

  68. [76]

    2025 , pages =

    Journal of Chemical Information and Modeling , author =. 2025 , pages =. doi:10.1021/acs.jcim.5c01853 , language =

    2025 doi

  69. [77]

    2024 , eprint=

    Introducing GPU-acceleration into the Python-based Simulations of Chemistry Framework , author=. 2024 , eprint=

    2024

  70. [78]

    2024 , eprint=

    Enhancing GPU-acceleration in the Python-based Simulations of Chemistry Framework , author=. 2024 , eprint=

    2024

  71. [79]

    2020 , pages =

    Journal of Chemical Information and Modeling , author =. 2020 , pages =. doi:10.1021/acs.jcim.0c00451 , language =

    2020 doi

  72. [80]

    2023 , eprint=

    A foundation model for atomistic materials chemistry , author=. 2023 , eprint=

    2023

  73. [81]

    2025 , eprint=

    The dark side of the forces: assessing non-conservative force models for atomistic machine learning , author=. 2025 , eprint=

    2025

  74. [82]

    ACS Medicinal Chemistry Letters , author =

    Transforming. ACS Medicinal Chemistry Letters , author =. 2018 , pages =. doi:10.1021/acsmedchemlett.8b00437 , language =

    2018 doi

  75. [83]

    Nature Communications , author =

    Approaching coupled cluster accuracy with a general-purpose neural network potential through transfer learning , volume =. Nature Communications , author =. 2019 , keywords =. doi:10.1038/s41467-019-10827-4 , abstract =

    2019 doi

  76. [84]

    Journal of the American Chemical Society , author =

    Molecular. Journal of the American Chemical Society , author =. 2025 , pages =. doi:10.1021/jacs.5c09558 , language =

    2025 doi

  77. [85]

    and Mishra, Adesh Rohan and Liu, Elliott and Ceder, Gerbrand and Miret, Santiago and Ong, Shyue Ping , title=

    Ko, Tsz Wai and Deng, Bowen and Nassar, Marcel and Barroso-Luque, Luis and Liu, Runze and Qi, Ji and Thakur, Atul C. and Mishra, Adesh Rohan and Liu, Elliott and Ceder, Gerbrand and Miret, Santiago and Ong, Shyue Ping , title=. npj Computational Materials , year=. doi:10.1038/...

    doi

  78. [86]

    Ko, Tsz Wai and Nassar, Marcel and Qi, Ji and Miret, Santiago and Ong, Shyue Ping , title =

  79. [87]

    Science , volume=

    The many roles of computation in drug discovery , author=. Science , volume=. 2004 , publisher=

    2004

  80. [88]

    Advancing drug discovery through enhanced free energy calculations , author=. Acc. Chem. Res. , volume=. 2017 , publisher=

    2017

  81. [89]

    and Riniker, Sereina and Cournia, Zoe , title =

    Armacost, Kira A. and Riniker, Sereina and Cournia, Zoe , title =. J. Chem. Inf. Model. , volume =. 2020 , doi =

    2020

  82. [90]

    Blinded prediction of protein--ligand binding affinity using Amber thermodynamic integration for the 2018 D3R grand challenge 4 , author=. J. Comput.-Aided Mol. Des. , volume=. 2019 , publisher=

    2018

  83. [91]

    pmx: Automated protein structure and topology generation for alchemical perturbations , author=. J. Comput. Chem. , volume=. 2015 , publisher=

    2015

  84. [92]

    TIES 20: Relative binding free energy with a flexible superimposition algorithm and partial ring morphing , author=. J. Chem. Theory Comput. , volume=. 2021 , publisher=

    2021

  85. [93]

    Accurate and reliable prediction of relative ligand binding potency in prospective drug discovery by way of a modern free-energy calculation protocol and force field , author =. J. Am. Chem. Soc. , volume=. 2015 , publisher=

    2015

  86. [94]

    Assessment of binding affinity via alchemical free-energy calculations , author=. J. Chem. Inf. Model. , volume=. 2020 , publisher=

    2020

  87. [95]

    Alchemical transfer approach to absolute binding free energy estimation , author=. J. Chem. Theory Comput. , volume=. 2021 , publisher=

    2021

  88. [96]

    Relative binding free energy calculations for ligands with diverse scaffolds with the alchemical transfer method , author=. J. Chem. Inf. Model. , volume=. 2022 , publisher=

    2022

  89. [97]

    PLoS Comput

    OpenMM 7: Rapid development of high performance algorithms for molecular dynamics , author=. PLoS Comput. Biol. , volume=. 2017 , publisher=

    2017

  90. [98]

    Asynchronous replica exchange software for grid and heterogeneous computing , author=. Comput. Phys. Commun. , volume=. 2015 , publisher=

    2015

  91. [99]

    ff14SB: improving the accuracy of protein side chain and backbone parameters from ff99SB , author=. J. Chem. Theory Comput. , volume=. 2015 , publisher=

    2015

  92. [100]

    Automatic atom type and bond type perception in molecular mechanical calculations , author=. J. Mol. Graphics Modell. , volume=. 2006 , publisher=

    2006

  93. [101]

    ACS Omega , volume=

    Fast, accurate, and reliable protocols for routine calculations of protein--ligand binding affinities in drug design projects using AMBER GPU-TI with ff14SB/GAFF , author=. ACS Omega , volume=. 2020 , publisher=

    2020

  94. [102]

    The Journal of Chemical Physics , volume=

    Theory of binless multi-state free energy estimation with applications to protein-ligand binding , author=. The Journal of Chemical Physics , volume=. 2012 , publisher=

    2012

  95. [103]

    Large-scale assessment of binding free energy calculations in active drug discovery projects , author=. J. Chem. Inf. Model. , volume=. 2020 , publisher=

    2020

  96. [104]

    2022 , howpublished =

    Gallicchio, Emilio , title =. 2022 , howpublished =

    2022

  97. [105]

    GitHub repository , howpublished =

    AToM-OpenMM , year =. GitHub repository , howpublished =

  98. [106]

    Case, David and Aktulga, H. Metin and Belfon, Kellon and Ben-Shalom, Ido and Brozell, Scott and Cerutti, David and Cheatham, Thomas and Cruzeiro, Vinícius and Darden, Tom and Duke, Robert and Giambasu, George and Gilson, Michael and Gohlke, Holger and Götz, Andreas and Harris,...

    2019

  99. [107]

    Case, H.M

    D.A. Case, H.M. Aktulga, K. Belfon, I.Y. Ben-Shalom, J.T. Berryman, S.R. Brozell, D.S. Cerutti, T.E. Cheatham, III, G.A. Cisneros, V.W.D. Cruzeiro, T.A. Darden, R.E. Duke, G. Giambasu, M.K. Gilson, H. Gohlke, A.W. Goetz, R. Harris, S. Izadi, S.A. Izmailov, K. Kasavajhala, M.C....

    2019

  100. [108]

    Alchemical transformations for concerted hydration free energy estimation with explicit solvation , author=. J. Chem. Phys. , volume=. 2021 , publisher=

    2021

  101. [109]

    Alchemical binding free energy calculations in AMBER20: Advances and best practices for drug discovery , author=. J. Chem. Inf. Model. , volume=. 2020 , publisher=

    2020

  102. [110]

    Large scale relative protein ligand binding affinities using non-equilibrium alchemy , journal=

    Gapsys, V and P. Large scale relative protein ligand binding affinities using non-equilibrium alchemy , journal=. 2020 , publisher=

    2020

  103. [111]

    The Statistical-Thermodynamic Basis for Computation of Binding Affinities: A Critical Review , Author =. Biophys. J. , Year =

  104. [112]

    Avoiding singularities and numerical instabilities in free energy calculations based on molecular simulations , author=. Chem. Phys. Lett. , volume=. 1994 , publisher=

    1994

  105. [113]

    The Journal of Chemical Physics , volume=

    Separation-shifted scaling, a new scaling method for Lennard-Jones interactions in thermodynamic integration , author=. The Journal of Chemical Physics , volume=. 1994 , publisher=

    1994

  106. [114]

    New soft-core potential function for molecular dynamics based alchemical free energy calculations , author=. J. Chem. Theory Comput. , volume=. 2012 , publisher=

    2012

  107. [115]

    Mey, Antonia S. J. S. and Allen, Bryce K. and Bruce McDonald, Hannah E. and Chodera, John D. and Hahn, David F. and Kuhn, Maxmillian and Michel, Julien and Mobley, David L. and Naden, Levi N. and Prasad, Samarjeet and Rizzi, Andrea and Scheen, Jenke and Shirts, Michael R. and ...

    doi

  108. [116]

    Accurate calculation of relative binding free energies between ligands with different net charges , author=. J. Chem. Theory Comput. , volume=. 2018 , publisher=

    2018

  109. [117]

    The Journal of Chemical Physics , volume=

    Calculating the binding free energies of charged species based on explicit-solvent simulations employing lattice-sum methods: An accurate correction scheme for electrostatic finite-size effects , author=. The Journal of Chemical Physics , volume=. 2013 , publisher=

    2013

  110. [118]

    The Journal of Chemical hysics , volume=

    Separated topologies—A method for relative binding free energy calculations using orientational restraints , author=. The Journal of Chemical hysics , volume=. 2013 , publisher=

    2013

  111. [119]

    Broadening the scope of binding free energy calculations using a Separated Topologies approach , author=

  112. [120]

    and Rufa, Dominic and Zhang, Ivy and Pulido, Iván and Henry, Mike and Bruce Macdonald, Hannah E

    Wang, Yuanqing and Fass, Josh and Kaminow, Benjamin and Herr, John E. and Rufa, Dominic and Zhang, Ivy and Pulido, Iván and Henry, Mike and Bruce Macdonald, Hannah E. and Takaba, Kenichiro and Chodera, John D. End-to-end differentiable construction of molecular mechanics force...

    2022 doi

  113. [121]

    Journal of chemical information and modeling , volume=

    NNP/MM: Accelerating molecular dynamics simulations with machine learning potentials and molecular mechanics , author=. Journal of chemical information and modeling , volume=. 2023 , publisher=

    2023

  114. [122]

    and Tresadern, Gary and Mobley, David L

    Gapsys, Vytautas and Hahn, David F. and Tresadern, Gary and Mobley, David L. and Rampp, Markus and de Groot, Bert L. , title =. Journal of Chemical Information and Modeling , volume =. 2022 , doi =

    2022

  115. [123]

    Schindler, Christina E. M. and Baumann, Hannah and Blum, Andreas and Böse, Dietrich and Buchstaller, Hans-Peter and Burgdorf, Lars and Cappel, Daniel and Chekler, Eugene and Czodrowski, Paul and Dorsch, Dieter and Eguida, Merveille K. I. and Follows, Bruce and Fuchß, Thomas an...

    2020

  116. [124]

    Journal of Chemical Information and Modeling , volume =

    Saban\'es Zariquiey, Francesc and P\'esrez, Adrià and Majewski, Maciej and Gallicchio, Emilio and De Fabritiis, Gianni , title =. Journal of Chemical Information and Modeling , volume =. 2023 , doi =

    2023

  117. [125]

    and Huddleston, Kate K

    Devereux, Christian and Smith, Justin S. and Huddleston, Kate K. and Barros, Kipton and Zubatyuk, Roman and Isayev, Olexandr and Roitberg, Adrian E. , title =. Journal of Chemical Theory and Computation , volume =. 2020 , doi =

    2020

  118. [126]

    0: The Sage Small Molecule Force Field , author=

    Development and Benchmarking of Open Force Field 2.0. 0: The Sage Small Molecule Force Field , author=. Journal of Chemical Theory and Computation , year=

  119. [127]

    Journal of chemical theory and computation , volume=

    HTMD: high-throughput molecular dynamics for molecular discovery , author=. Journal of chemical theory and computation , volume=. 2016 , publisher=

    2016

  120. [128]

    Journal of computational chemistry , volume=

    Development and testing of a general amber force field , author=. Journal of computational chemistry , volume=. 2004 , publisher=

    2004

  121. [129]

    Journal of computational chemistry , volume=

    CHARMM general force field: A force field for drug-like molecules compatible with the CHARMM all-atom additive biological force fields , author=. Journal of computational chemistry , volume=. 2010 , publisher=

    2010

  122. [130]

    Journal of chemical information and modeling , volume=

    Automation of the CHARMM General Force Field (CGenFF) II: assignment of bonded parameters and partial atomic charges , author=. Journal of chemical information and modeling , volume=. 2012 , publisher=

    2012

  123. [131]

    Journal of chemical theory and computation , volume=

    OPLS3e: Extending force field coverage for drug-like small molecules , author=. Journal of chemical theory and computation , volume=. 2019 , publisher=

    2019

  124. [132]

    Advances in protein chemistry , volume=

    Force fields for protein simulations , author=. Advances in protein chemistry , volume=. 2003 , publisher=

    2003

  125. [133]

    Journal of computer-aided molecular design , volume=

    Biomolecular force fields: where have we been, where are we now, where do we need to go and how do we get there? , author=. Journal of computer-aided molecular design , volume=. 2019 , publisher=

    2019

  126. [134]

    Journal of computer-aided molecular design , volume=

    Force field development phase II: Relaxation of physics-based criteria… or inclusion of more rigorous physics into the representation of molecular energetics , author=. Journal of computer-aided molecular design , volume=. 2019 , publisher=

    2019

  127. [135]

    The Journal of Physical Chemistry B , volume=

    A simple QM/MM approach for capturing polarization effects in protein- ligand binding free energy calculations , author=. The Journal of Physical Chemistry B , volume=. 2011 , publisher=

    2011

  128. [136]

    Journal of chemical information and modeling , volume=

    A scalable molecular force field parameterization method based on density functional theory and quantum-level machine learning , author=. Journal of chemical information and modeling , volume=. 2019 , publisher=

    2019

  129. [137]

    2023 , eprint=

    Performance and Analysis of the Alchemical Transfer Method for Binding Free Energy Predictions of Diverse Ligands , author=. 2023 , eprint=

    2023

  130. [138]

    Rufa and Hannah E

    Dominic A. Rufa and Hannah E. Bruce Macdonald and Josh Fass and Marcus Wieder and Patrick B. Grinaway and Adrian E. Roitberg and Olexandr Isayev and John D. Chodera , title =. 2020 , doi =. https://www.biorxiv.org/content/early/2020/07/30/2020.07.29.227959.full.pdf , journal =

    2020

  131. [139]

    0] , author=

    Best practices for constructing, preparing, and evaluating protein-ligand binding affinity benchmarks [article v1. 0] , author=. Living journal of computational molecular science , volume=. 2022 , publisher=

    2022

  132. [140]

    doi:10.5281/zenodo.6210305 , url =

    Hannah Bruce Macdonald and dfhahn and Mike Henry and John Chodera and David Dotson and William Glass and Iván Pulido , title =. doi:10.5281/zenodo.6210305 , url =

    doi

  133. [141]

    Communications Chemistry , volume=

    The maximal and current accuracy of rigorous protein-ligand binding free energy calculations , author=. Communications Chemistry , volume=. 2023 , publisher=

    2023

  134. [142]

    Journal of Chemical Information and Modeling , volume=

    Enhancing Protein--Ligand Binding Affinity Predictions Using Neural Network Potentials , author=. Journal of Chemical Information and Modeling , volume=. 2024 , publisher=

    2024

  135. [143]

    Journal of Chemical Information and Modeling , volume=

    Current State of Open Source Force Fields in Protein--Ligand Binding Affinity Predictions , author=. Journal of Chemical Information and Modeling , volume=. 2024 , publisher=

    2024

  136. [144]

    Journal of chemical theory and computation , volume=

    OPLS4: Improving force field accuracy on challenging regimes of chemical space , author=. Journal of chemical theory and computation , volume=. 2021 , publisher=

    2021

  137. [145]

    Mitchell and Lily Wang and Jeff Wagner and trevorgokey , title =

    SimonBoothroyd and Matt Thompson and Josh A. Mitchell and Lily Wang and Jeff Wagner and trevorgokey , title =. doi:10.5281/zenodo.8118623 , url =

    doi

  138. [146]

    Journal of Chemical Information and Modeling , volume=

    Alchemical Transformations and Beyond: Recent Advances and Real-World Applications of Free Energy Calculations in Drug Discovery , author=. Journal of Chemical Information and Modeling , volume=. 2024 , publisher=

    2024

  139. [147]

    ACS Physical Chemistry Au , volume=

    Modern alchemical free energy methods for drug discovery explained , author=. ACS Physical Chemistry Au , volume=. 2023 , publisher=

    2023

  140. [148]

    0: the Sage small molecule force field , author=

    Development and benchmarking of open force field 2.0. 0: the Sage small molecule force field , author=. Journal of chemical theory and computation , volume=. 2023 , publisher=

    2023

  141. [149]

    0.0—the parsley small-molecule force field , author=

    Development and benchmarking of open force field v1. 0.0—the parsley small-molecule force field , author=. Journal of chemical theory and computation , volume=. 2021 , publisher=

    2021

  142. [150]

    The Journal of Physical Chemistry B , volume=

    OpenMM 8: Molecular Dynamics Simulation with Machine Learning Potentials , author=. The Journal of Physical Chemistry B , volume=. 2024 , publisher=

    2024

  143. [151]

    Journal of chemical information and modeling , volume=

    TorchANI: a free and open source PyTorch-based deep learning implementation of the ANI neural network potentials , author=. Journal of chemical information and modeling , volume=. 2020 , publisher=

    2020

  144. [152]

    Journal of chemical theory and computation , volume=

    GFN2-xTB—An accurate and broadly parametrized self-consistent tight-binding quantum chemical method with multipole electrostatics and density-dependent dispersion contributions , author=. Journal of chemical theory and computation , volume=. 2019 , publisher=

    2019

  145. [153]

    arXiv preprint arXiv:2312.07511 , year=

    A Hitchhiker's Guide to Geometric GNNs for 3D Atomic Systems , author=. arXiv preprint arXiv:2312.07511 , year=

    arXiv

  146. [154]

    GPUGRID - Volunteer Computing for Molecular Simulations , howpublished =

  147. [155]

    TensorNet: Cartesian Tensor Representations for Efficient Learning of Molecular Potentials , url =

    Simeon, Guillem and De Fabritiis, Gianni , booktitle =. TensorNet: Cartesian Tensor Representations for Efficient Learning of Molecular Potentials , url =

  148. [156]

    Journal of chemical information and modeling , volume=

    A comparison of quantum and molecular mechanical methods to estimate strain energy in druglike fragments , author=. Journal of chemical information and modeling , volume=. 2017 , publisher=

    2017

  149. [157]

    The Journal of chemical physics , volume=

    Geometry optimization made simple with translation and rotation coordinates , author=. The Journal of chemical physics , volume=. 2016 , publisher=

    2016

  150. [158]

    2024 , howpublished =

    Acellera , title =. 2024 , howpublished =

    2024

  151. [159]

    Journal of the American Chemical Society , volume=

    Kov. Journal of the American Chemical Society , volume=

  152. [160]

    Anstine, D. M. and Zubatyuk, R. and Isayev, O. , journal=

  153. [161]

    arXiv preprint arXiv:2408.12625 , year=

    Machine Learning Potentials: A Roadmap Toward Next-Generation Biomolecular Simulations , author=. arXiv preprint arXiv:2408.12625 , year=

    arXiv

  154. [162]

    Journal of Chemical Theory and Computation , volume=

    TorchMD-Net 2.0: Fast Neural Network Potentials for Molecular Simulations , author=. Journal of Chemical Theory and Computation , volume=. 2024 , publisher=

    2024

  155. [163]

    arXiv preprint arXiv:2202.02541 , year=

    Torchmd-net: equivariant transformers for neural network based molecular potentials , author=. arXiv preprint arXiv:2202.02541 , year=

    arXiv

  156. [164]

    Journal of chemical theory and computation , volume=

    TorchMD: A deep learning framework for molecular simulations , author=. Journal of chemical theory and computation , volume=. 2021 , publisher=

    2021

  157. [165]

    Journal of Chemical Theory and Computation , year=

    Evaluation of Machine Learning/Molecular Mechanics End-State Corrections with Mechanical Embedding to Calculate Relative Protein--Ligand Binding Free Energies , author=. Journal of Chemical Theory and Computation , year=

  158. [166]

    Journal of chemical theory and computation , volume=

    QD : A quantum deep potential interaction model for drug discovery , author=. Journal of chemical theory and computation , volume=. 2023 , publisher=

    2023

  159. [167]

    The Journal of Chemical Physics , volume=

    Amber free energy tools: Interoperable software for free energy simulations using generalized quantum mechanical/molecular mechanical and machine learning potentials , author=. The Journal of Chemical Physics , volume=. 2024 , publisher=

    2024

  160. [168]

    Journal of chemical theory and computation , volume=

    Development of range-corrected deep learning potentials for fast, accurate quantum mechanical/molecular mechanical simulations of chemical reactions in solution , author=. Journal of chemical theory and computation , volume=. 2021 , publisher=

    2021

  161. [169]

    The Journal of Physical Chemistry B , year=

    Software Infrastructure for Next-Generation QM/MM- MLP Force Fields , author=. The Journal of Physical Chemistry B , year=

  162. [170]

    The Journal of Physical Chemistry Letters , volume=

    Alchemical Free-Energy Calculations at Quantum-Chemical Precision , author=. The Journal of Physical Chemistry Letters , volume=. 2025 , publisher=

    2025

  163. [171]

    Molecular Modeling of Proteins , pages=

    Calculation of binding free energies , author=. Molecular Modeling of Proteins , pages=. 2015 , publisher=

    2015

  164. [172]

    Current topics in medicinal chemistry , volume=

    The significance of chirality in drug design and development , author=. Current topics in medicinal chemistry , volume=. 2011 , publisher=

    2011

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.