pith. machine review for the scientific record. sign in

arxiv: 2604.09540 · v1 · submitted 2026-04-10 · 💻 cs.ET · q-bio.BM

Recognition: unknown

A Physically-Informed Subgraph Isomorphism Approach to Molecular Docking Using Quantum Annealers

Francesco Micucci , Matteo Barbieri , Gabriella Bettonte , Domenico Bonanni , Anita Camillini , Anna Fava , Daniele Gregori , Andrea R. Beccari
show 1 more author
Gianluca Palermo
Authors on Pith no claims yet

Pith reviewed 2026-05-10 16:02 UTC · model grok-4.3

classification 💻 cs.ET q-bio.BM
keywords molecular dockingquantum annealingQUBOphysicochemical interactionssubgraph isomorphismD-Wavedrug design
0
0 comments X

The pith

Adding physical interaction terms to QUBO improves molecular docking accuracy on quantum annealers

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

The paper extends an earlier geometric-only method for molecular docking on D-Wave annealers by adding corrective terms for Coulomb forces, van der Waals forces, hydrogen bonds, and hydrophobic interactions. These terms are folded into the same QUBO that encodes subgraph isomorphism between the ligand graph and the discretized protein pocket grid. The authors test the augmented formulation on the hardware and report that the physical corrections change the selected poses in ways that raise accuracy relative to the geometry-only baseline. A reader would care because docking accuracy directly affects which candidate molecules advance in drug design pipelines, and the work tests whether quantum hardware can carry both shape and energy information in one optimization step.

Core claim

The central claim is that a QUBO formulation for ligand-protein docking, previously limited to geometric graph isomorphism, can be extended by adding corrective energy terms for Coulomb and van der Waals forces together with H-bond and hydrophobic components; when solved on quantum annealers the resulting poses show measurable accuracy gains over the purely geometric version.

What carries the argument

The QUBO that augments the prior subgraph-isomorphism objective with additive terms representing Coulomb, van der Waals, hydrogen-bond, and hydrophobic interactions between ligand and protein atoms.

If this is right

  • The physical corrections alter the energy landscape seen by the annealer and therefore change which ligand poses are returned.
  • The same corrective-term approach can be applied to other subgraph-isomorphism problems that also carry interaction energies.
  • Hardware results on D-Wave quantify the accuracy difference between the two QUBO versions for the tested cases.
  • The formulation keeps the problem size compatible with current annealer embedding limits while adding physics.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • If the added terms remain linear or quadratic in the binary variables, the method could be ported to other quantum optimization platforms without changing the solver.
  • The approach suggests that hybrid geometric-plus-energy QUBOs might be useful for related placement problems such as protein-ligand or protein-protein docking variants.
  • Scaling the grid resolution or the number of interaction types would test how far the current embedding overhead grows before the annealer can no longer solve the instance.

Load-bearing premise

That the physical interaction energies can be discretized and mapped onto the existing QUBO variables without introducing approximation errors large enough to cancel the claimed accuracy improvement.

What would settle it

Direct comparison of root-mean-square deviation or binding-pose success rate for the same ligand-protein pairs solved once with the geometric QUBO and once with the full physicochemical QUBO on the same D-Wave device; if accuracy does not rise or falls, the central claim is refuted.

Figures

Figures reproduced from arXiv: 2604.09540 by Andrea R. Beccari, Anita Camillini, Anna Fava, Daniele Gregori, Domenico Bonanni, Francesco Micucci, Gabriella Bettonte, Gianluca Palermo, Matteo Barbieri.

Figure 1
Figure 1. Figure 1: Hydrogen bond example. The distance between D and A should be ˚ [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Comparison of the fully geometric approach and the proposed [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Comparison of the fully geometric approach and the proposed method [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
read the original abstract

Molecular docking is a crucial step in the development of new drugs as it guides the positioning of a small molecule (ligand) within the pocket of a target protein. In the literature, a feasibility study explored the potential of D-Wave quantum annealers for purely geometric molecular docking, neglecting physicochemical interactions between the protein and the ligand and focusing solely on their simplified geometries. To achieve this, the ligands were represented as graphs incorporating their geometric properties and then mapped onto a grid that discretized the three-dimensional space of the protein pocket. The quality of the ligand pose on the protein pocket was evaluated through the isomorphism between the ligand graph and the spatial grid. This paper builds on the previous study by introducing physicochemical interactions between the protein-ligand pair into the QUBO problem to improve the accuracy of the docking results. This paper presents a novel QUBO formulation that includes Coulomb and van der Waals forces, together with components representing H-bond and hydrophobic interactions. We integrate these physical interactions as corrective terms to the previous purely geometric QUBO formulation, and provide experimental results using the D-Wave quantum annealers to demonstrate their impact on the accuracy of the docking results.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The manuscript extends a prior geometric subgraph-isomorphism QUBO for molecular docking on D-Wave quantum annealers by adding corrective terms that encode Coulomb and van der Waals forces together with H-bond and hydrophobic interactions. It reports experimental runs on D-Wave hardware intended to show that these physically-informed terms improve pose accuracy relative to the purely geometric baseline.

Significance. If the mapping of continuous potentials to the QUBO is free of uncontrolled discretization artifacts or post-hoc scaling and if quantitative improvements over the geometric baseline are demonstrated with proper controls, the work would provide a concrete example of incorporating domain physics into quantum-annealing formulations for structure-based drug design.

major comments (2)
  1. [Abstract / Methods] The abstract states that Coulomb, van der Waals, H-bond and hydrophobic terms are integrated as corrective terms, yet no equations, discretization procedure, or explicit QUBO matrix entries for these terms are supplied; without this information it is impossible to verify that the added interactions do not introduce new tunable weights or grid-mapping approximations that could account for any observed accuracy change.
  2. [Results / Experiments] The claim that the physical terms improve docking accuracy rests on experimental results, but the manuscript provides neither quantitative metrics (e.g., RMSD distributions, success rates), nor a direct comparison against the purely geometric QUBO baseline, nor details on how the D-Wave embeddings and annealing schedules were configured; these omissions make it impossible to assess whether the reported impact is load-bearing or an artifact of the mapping.
minor comments (1)
  1. [Methods] Notation for the grid variables and the subgraph-isomorphism penalty terms should be defined consistently with the earlier geometric paper so that the corrective terms can be read as additive modifications.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments, which highlight important areas for clarification. We address each major point below and will revise the manuscript to supply the missing details.

read point-by-point responses

  1. Referee: [Abstract / Methods] The abstract states that Coulomb, van der Waals, H-bond and hydrophobic terms are integrated as corrective terms, yet no equations, discretization procedure, or explicit QUBO matrix entries for these terms are supplied; without this information it is impossible to verify that the added interactions do not introduce new tunable weights or grid-mapping approximations that could account for any observed accuracy change.

    Authors: We agree that the submitted version omits the explicit equations and QUBO construction details for the physicochemical terms. In the revised manuscript we will add a dedicated Methods subsection that (i) derives the corrective QUBO terms for Coulomb, van der Waals, hydrogen-bond and hydrophobic interactions, (ii) specifies the discretization grid and scaling procedure used to map the continuous potentials, and (iii) lists the resulting quadratic coefficients that are added to the original geometric QUBO. This will make clear that the added terms follow the same grid-mapping framework as the geometric baseline and do not introduce independent tunable weights beyond those already present in the prior formulation. revision: yes

  2. Referee: [Results / Experiments] The claim that the physical terms improve docking accuracy rests on experimental results, but the manuscript provides neither quantitative metrics (e.g., RMSD distributions, success rates), nor a direct comparison against the purely geometric QUBO baseline, nor details on how the D-Wave embeddings and annealing schedules were configured; these omissions make it impossible to assess whether the reported impact is load-bearing or an artifact of the mapping.

    Authors: We acknowledge that the current Results section is insufficiently quantitative. We will expand it to include (i) RMSD distributions and success-rate statistics at standard thresholds (e.g., RMSD < 2 Å), (ii) a direct, side-by-side comparison against the purely geometric QUBO baseline on the same ligand–protein pairs, and (iii) full experimental configuration details: embedding heuristics, chain-strength values, annealing times, and any post-processing used. These additions will allow readers to evaluate whether the observed accuracy gains are attributable to the physical terms rather than mapping artifacts. revision: yes

Circularity Check

0 steps flagged

No circularity: additive formulation with independent experimental validation

full rationale

The paper extends a prior geometric QUBO by adding explicit corrective terms for Coulomb, van der Waals, H-bond and hydrophobic interactions. This is an additive construction whose claimed improvement is validated by new D-Wave experiments rather than by re-deriving or fitting the baseline. The self-citation supplies only the geometric starting point; the central result (accuracy impact of the physicochemical terms) does not reduce to that citation by definition or by construction. No fitted parameters are renamed as predictions, no ansatz is smuggled, and no uniqueness theorem is invoked.

Axiom & Free-Parameter Ledger

1 free parameters · 2 axioms · 0 invented entities

The approach rests on the assumption that docking can be cast as subgraph isomorphism on a discretized grid and that physical interactions can be added as linear or quadratic corrective terms in the same QUBO without breaking the mapping to the annealer.

free parameters (1)
  • weights or scaling factors for physical interaction terms
    The relative strength of geometric versus physicochemical contributions must be balanced by parameters whose values are not specified in the abstract.
axioms (2)
  • domain assumption Subgraph isomorphism between ligand graph and protein-pocket grid can be encoded as a QUBO solvable by quantum annealing
    Inherited from the prior geometric feasibility study referenced in the abstract.
  • domain assumption Coulomb, van der Waals, H-bond, and hydrophobic interactions can be approximated as additive corrective terms in the QUBO objective
    Stated as the core modeling choice in the abstract.

pith-pipeline@v0.9.0 · 5540 in / 1497 out tokens · 46115 ms · 2026-05-10T16:02:07.769060+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

25 extracted references · 6 canonical work pages · 1 internal anchor

  1. [1]

    Protein-ligand docking: current status and future challenges,

    S. F. Sousa, P. A. Fernandes, and M. J. Ramos, “Protein-ligand docking: current status and future challenges,”Proteins, vol. 65, no. 1, pp. 15–26, Oct. 2006

    2006

  2. [2]

    A guide to in silico drug design,

    Y . Chang, B. A. Hawkins, J. J. Du, P. W. Groundwater, D. E. Hibbs, and F. Lai, “A guide to in silico drug design,”Pharmaceutics, vol. 15, no. 1, p. 49, Dec. 2022

    2022

  3. [3]

    Comparison of shape-matching and docking as virtual screening tools,

    P. C. D. Hawkins, A. G. Skillman, and A. Nicholls, “Comparison of shape-matching and docking as virtual screening tools,”J. Med. Chem., vol. 50, no. 1, pp. 74–82, Jan. 2007

    2007

  4. [4]

    Molecular Docking via Weighted Subgraph Isomorphism on Quantum Annealers

    E. Triuzzi, R. Mengoni, F. Micucci, D. Bonanni, D. Ottaviani, A. Beccari, and G. Palermo, “Molecular docking via weighted subgraph isomorphism on quantum annealers,”arXiv, 2025. [Online]. Available: https://arxiv.org/abs/2405.06657

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  5. [5]

    Encoding molecular docking for quantum computers,

    J. Zha, J. Su, T. Li, C. Cao, Y . Ma, H. Wei, Z. Huang, L. Qian, K. Wen, and J. Zhang, “Encoding molecular docking for quantum computers,” Journal of Chemical Theory and Computation, vol. 19, no. 24, pp. 9018– 9024, 2023

    2023

  6. [6]

    Molecular docking with gaussian boson sampling,

    L. Banchi, M. Fingerhuth, T. Babej, C. Ing, and J. M. Arrazola, “Molecular docking with gaussian boson sampling,”Science Advances, vol. 6, no. 23, p. eaax1950, 2020

    2020

  7. [7]

    Molecular docking via quantum approximate optimization algorithm,

    Q.-M. Ding, Y .-M. Huang, and X. Yuan, “Molecular docking via quantum approximate optimization algorithm,”arXiv, 2024. [Online]. Available: https://arxiv.org/abs/2308.04098

    work page arXiv 2024

  8. [8]

    Towards molecular docking with neutral atoms,

    M. Garrigues, V . Onofre, and N. Bosc-Haddad, “Towards molecular docking with neutral atoms,”arXiv, 2024. [Online]. Available: https://arxiv.org/abs/2402.06770

    work page arXiv 2024

  9. [9]

    Chemical function based pharmacophore models as suitable filters for virtual 3d-database screening,

    T. Langer, R. Hoffmann, F. Bachmair, and S. Begle, “Chemical function based pharmacophore models as suitable filters for virtual 3d-database screening,”Journal of Molecular Structure: THEOCHEM, vol. 503, no. 1, pp. 59–72, 2000. [Online]. Available: https://www.sciencedirect.com/science/article/pii/S0166128099003632

    2000

  10. [10]

    Computer-Aided drug design and drug discovery: A prospective analysis,

    S. K. Niazi and Z. Mariam, “Computer-Aided drug design and drug discovery: A prospective analysis,”Pharmaceuticals (Basel), vol. 17, no. 1, Dec. 2023

    2023

  11. [11]

    Molecular docking,

    G. M. Morris and M. Lim-Wilby, “Molecular docking,” inMethods in Molecular Biology, ser. Methods in molecular biology (Clifton, N.J.). Totowa, NJ: Humana Press, 2008, pp. 365–382

    2008

  12. [12]

    Quantum molecular unfolding,

    K. Mato, R. Mengoni, D. Ottaviani, and G. Palermo, “Quantum molecular unfolding,”arXiv, 2021. [Online]. Available: https://arxiv.org/abs/2107.13607

    work page arXiv 2021

  13. [13]

    Fast prediction and visualization of protein binding pockets with pass,

    G. P. Brady, Jr and P. F. W. Stouten, “Fast prediction and visualization of protein binding pockets with pass,”J. Comput. Aided Mol. Des., vol. 14, no. 4, pp. 383–401, 2000

    2000

  14. [14]

    T. Hou, J. Wang, Y . Li, and W. Wang, “Assessing the performance of the molecular mechanics/poisson boltzmann surface area and molecular mechanics/generalized born surface area methods. ii. the accuracy of ranking poses generated from docking,”Journal of Computational Chemistry, vol. 32, no. 5, pp. 866–877, 2011. [Online]. Available: https://onlinelibrary...

    work page doi:10.1002/jcc.21666 2011

  15. [15]

    Ueber die anwendung des satzes vom virial in der kinetischen theorie der gase,

    H. A. Lorentz, “Ueber die anwendung des satzes vom virial in der kinetischen theorie der gase,”Ann. Phys., vol. 248, no. 1, pp. 127–136, Jan. 1881

  16. [16]

    Further development and validation of empirical scoring functions for structure-based binding affinity predic- tion,

    R. Wang, L. Lai, and S. Wang, “Further development and validation of empirical scoring functions for structure-based binding affinity predic- tion,”J. Comput. Aided Mol. Des., vol. 16, no. 1, pp. 11–26, 2002

    2002

  17. [17]

    Merck molecular force field. II. MMFF94 van der waals and electrostatic parameters for intermolecular interactions,

    T. A. Halgren, “Merck molecular force field. II. MMFF94 van der waals and electrostatic parameters for intermolecular interactions,”J. Comput. Chem., vol. 17, no. 5-6, pp. 520–552, Apr. 1996

    1996

  18. [18]

    Grabowski,Hydrogen Bonding—New Insights, 01 2006, vol

    S. Grabowski,Hydrogen Bonding—New Insights, 01 2006, vol. 3

    2006

  19. [19]

    An overview of the amber biomolecular simulation package,

    R. Salomon-Ferrer, D. A. Case, and R. C. Walker, “An overview of the amber biomolecular simulation package,”WIREs Computational Molecular Science, vol. 3, no. 2, pp. 198–210, 2013. [Online]. Available: https://wires.onlinelibrary.wiley.com/doi/abs/10.1002/wcms.1121

    work page doi:10.1002/wcms.1121 2013

  20. [20]

    Charmm general force field: A force field for drug-like molecules compatible with the charmm all-atom additive biological force fields,

    K. Vanommeslaeghe, E. Hatcher, C. Acharya, S. Kundu, S. Zhong, J. Shim, E. Darian, O. Guvench, P. Lopes, I. V orobyov, and A. MacK- erell, “Charmm general force field: A force field for drug-like molecules compatible with the charmm all-atom additive biological force fields,” Journal of Computational Chemistry, vol. 31, no. 4, pp. 671–690, Mar. 2010

    2010

  21. [21]

    ProLIF: a library to encode molecular interactions as fingerprints,

    C. Bouysset and S. Fiorucci, “ProLIF: a library to encode molecular interactions as fingerprints,”J. Cheminform., vol. 13, no. 1, p. 72, Sep. 2021

    2021

  22. [22]

    Advantage processor overview,

    C. McGeoch and P. Farr ´e, “Advantage processor overview,” D-Wave Systems Inc., Tech. Rep., 2022. [Online]. Available: https://www.dwavequantum.com/resources/white-paper/the- d-wave-advantage-system-an-overview/

    2022

  23. [23]

    Forging the basis for developing protein–ligand interaction scoring functions,

    Z. Liu, M. Su, L. Han, J. Liu, Q. Yang, Y . Li, and R. Wang, “Forging the basis for developing protein–ligand interaction scoring functions,” Accounts of Chemical Research, vol. 50, no. 2, pp. 302–309, 2017

    2017

  24. [24]

    Pdbbind database,

    P. Database, “Pdbbind database,” 2025, accessed: 2025-02-08. [Online]. Available: http://www.pdbbind.org.cn/

    2025

  25. [25]

    Performance of domain- wall encoding for quantum annealing,

    J. Chen, T. Stollenwerk, and N. Chancellor, “Performance of domain- wall encoding for quantum annealing,”IEEE Transactions on Quantum Engineering, vol. 2, pp. 1–14, 2021

    2021