Pith. sign in

REVIEW 4 major objections 8 minor 55 references

EnerBridge-DPO: Energy-Guided Protein Inverse Folding with Markov Bridges and Direct Preference Optimization

T0 review · 4 major / 8 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read The paper claims that energy-guided preference fine-tuning of a Markov bridge sequence generator yields low-energy protein designs without sacrificing recovery fidelity.

desk verdict A sensible integration of bridge inverse folding, DPO, and energy loss, but the central DPO derivation has an unproven L2 proxy and the energy gains are not cleanly attributed to preference learning. read the letter →

arxiv 2506.09496 v1 pith:FCGTJVN6 submitted 2025-06-11 cs.LG cs.AI

Dingyi Rong , Haotian Lu , Wenzhuo Zheng , Fan Zhang , Shuangjia Zheng , Ning Liu This is my paper
classification cs.LGcs.AI
keywords proteininversefoldingMarkovbridgedirectpreferenceoptimizationenergy-guidedsequencedesignbindingfreeenergypredictionstabilitycomplexgenerative
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

Protein inverse folding asks which amino-acid sequences will fold into a given backbone; most learned solvers optimize sequence recovery and ignore the energetic stability of what they generate. EnerBridge-DPO is proposed as a way to make energy itself part of the objective: a Markov bridge generator is first pretrained to recover native-like sequences from a structure-derived prior, then fine-tuned with a preference loss in which lower-energy sequences are treated as winners and a supervised energy-constraint loss that teaches the model to predict binding free-energy changes. The paper reports that on protein complexes the designed sequences have lower predicted energy than those of all compared models, while recovery and perplexity stay comparable to the strongest baseline, and that the same model predicts $\Delta\Delta G$ at a level comparable to a dedicated predictor. A sympathetic reader would take away that inverse folding can be steered toward thermodynamic stability directly, without a separate post-hoc filtering step.

What carries the argument

The central object is the Bridge-DPO loss, an adaptation of direct preference optimization to the Markov bridge process. The Markov bridge is a generative process that starts from a structure-conditioned prior sequence and iteratively refines it toward a target distribution; the DPO step uses energy-based winner/loser pairs and a frozen reference model, and replaces the intractable path-likelihood ratio with the difference in squared prediction errors between the fine-tuned and reference models. Alongside it, an energy-constraint loss derived from a Boltzmann-aligned inverse folding predictor regresses predicted onto experimental $\Delta\Delta G$, forcing the model to learn quantitative energy values. A pretrained structure encoder supplies the informative prior sequence that anchors both stages.

What would settle it

Compute the exact log-likelihood ratio of the bridge process for a sample of preference pairs and compare it with the squared-error proxy used in the Bridge-DPO loss; if the two disagree in sign on a non-negligible fraction of pairs, the objective is not actually direct preference optimization. A simpler check is to train the model with only the energy-constraint loss and see whether it reproduces the reported energy improvement on the 26-complex benchmark.

Watch

Extended reading notes

Core claim

On its own terms, the paper's central claim is that EnerBridge-DPO simultaneously optimizes sequence-structure compatibility, low predicted energy, and quantitative binding-energy prediction in one inverse folding model. The discovery is that an energy oracle can replace human feedback in direct preference optimization for a Markov bridge generative model: preference pairs are built by ranking candidate sequences by measured or predicted energy, and the DPO objective is adapted so that the bridge model is pushed to assign lower reconstruction error to the low-energy winner than to the higher-energy loser, relative to the frozen pretrained reference. The explicit energy-constraint loss adds regression to experimental $\Delta\Delta G$ values, which the paper argues makes the model capture quantitative energy features rather than only relative rankings. The evidence offered is lower predicted energy on 26 protein complexes according to three computational energy predictors, recovery and perplexity on the MPNN benchmark comparable to the leading bridge baseline, and three-fold cross-validated $\Delta\Delta G$ predictions comparable to a dedicated predictor.

Load-bearing premise

The load-bearing premise is that a model's squared prediction error correctly represents how likely it is to generate a sequence, so that ranking sequences by these errors is the same as ranking them by preference; if that substitution fails, the claimed preference optimization may reduce to the supervised energy loss.

Editorial extensions

If this is right

  • If EnerBridge-DPO works as reported, protein designers can fold energy optimization into the generative model itself instead of screening or rescoring designs after generation.
  • The same checkpoint can serve two purposes: produce low-energy sequences and predict $\Delta\Delta G$ for mutations, potentially removing the need for a separate predictor.
  • Because recovery and perplexity stay at the level of the base bridge model, energy alignment does not have to trade away the model's ability to find sequences compatible with the backbone.
  • The preference-pair construction from mutational datasets suggests that any data source with energy or fitness rankings could be used to steer the generator toward more stable designs.
  • The approach is in principle agnostic to the energy oracle, so swapping in a more accurate or faster scorer would directly retarget the model.

Reading between the lines

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

  • Editorial inference: the reported energy gains are measured with computational energy predictors, not experimental stability; wet-lab validation on a handful of designed complexes would be the decisive next test.
  • Editorial inference: because the bridge-DPO loss substitutes squared prediction error for the true bridge log-likelihood ratio, it is plausible that similar energy reductions could be achieved by the supervised energy loss alone; an ablation that isolates the preference term on the energy benchmark would clarify the mechanism.
  • Editorial inference: the energy-as-preference recipe could generalize to other structure-conditioned generative models, since the preference signal is carried by the energy oracle rather than by any bridge-specific property.
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

4 major / 8 minor

Summary. The paper presents EnerBridge-DPO, an inverse folding model that fine-tunes a Markov-bridge pretrained model (Bridge-IF style) with a Direct Preference Optimization (DPO) objective using energy-based preference pairs, together with an explicit energy-constraint loss. The authors report that EnerBridge-DPO maintains perplexity and recovery rates comparable to Bridge-IF, produces lower predicted energies (FoldX, Rosetta, BA-Cycle) on 26 selected protein complexes, and achieves Delta-Delta-G prediction performance comparable to BA-DDG. The central methodological novelty is the Bridge-DPO loss in Eq. (7), whose derivation in Appendix B substitutes L2 prediction error for the DPO log-likelihood ratio.

Significance. If the central derivation were valid and the ablations clean, the paper would be a useful step toward energy-aware inverse folding with generative models, with practical value for protein engineering. The paper includes comparisons against many baselines, evaluations with three independent energy functions, and ablation studies. The pretraining stage itself is a reasonable use of Markov bridges. However, the validity of the Bridge-DPO objective and the attribution of the energy improvements to the DPO mechanism are not established by the present evidence; these points are load-bearing for the paper's claimed contribution.

major comments (4)
  1. [Section 3.3.2, Eq. (7), Appendix B.4] The Bridge-DPO loss in Eq. (7) replaces the DPO log-likelihood ratio with differences of L2 prediction errors ||Y - phi_theta(z_t,t)||^2, while the pretraining objective in Eq. (6) is a cross-entropy negative log-likelihood. For a probability-vector output, L2 error is not proportional to negative log-likelihood, so the substitution does not preserve the DPO gradient structure. Appendix B.4 explicitly calls this a 'simplification' but provides no quantitative or empirical justification. Because the paper's headline energy improvement is attributed to this preference fine-tuning, this unsupported proxy is load-bearing for the paper's central claim.
  2. [Section 3.3.2, Eq. (7), Algorithm 1] Equation (7) samples z_t^w ~ q(z_t^w|X,S) and z_t^l ~ q(z_t^l|X,S), i.e., without conditioning on the target sequences Y_w and Y_l. Algorithm 1 lines 7-8, however, sample z_t^w ~ q(z_t|X,S,Y_w) and z_t^l ~ q(z_t|X,S,Y_l). For a Markov bridge, these are different distributions: dropping the endpoint changes the noisy state distribution at each timestep. Appendix B.4 acknowledges the simplification but does not argue that the resulting loss preserves the intended preference ordering. The measured energy reductions in Table 2 could therefore be an artifact of evaluating winner/loser errors at mismatched bridge states.
  3. [Section 3.4, Section 4.3, Table 3] The Delta-Delta-G prediction claim in Section 4.3 is not evidence of emergent physical understanding. Equation (8) uses a learnable k_BT scaling factor and Equation (9) trains the model with an explicit MAE loss against experimental Delta-Delta-G labels. Table 3 therefore largely measures supervised fitting ability. Moreover, the construction of the experimental label for a pair (Y_w,Y_l) in Eq. (9) is not specified: SKEMPI provides per-mutation Delta-Delta-G values, and the paper does not state whether the label is the difference between the two mutants' values or some other quantity. Both the interpretation and the reproducibility of Table 3 need to be addressed.
  4. [Section 4.4, Tables 4 and 5, Table 2] The ablation study in Section 4.4 does not report FoldX, Rosetta, or BA-Cycle energies for the 26 complexes used in Table 2. The w/o DPO and w/o Energy variants are evaluated only on perplexity/recovery and Delta-Delta-G metrics. As a result, the central claim that the DPO preference mechanism, rather than the explicit energy loss L_energy, is responsible for the lower predicted energies in Table 2 is not supported by the presented experiments. An ablation reporting the Table 2 energy metrics for all three variants is needed.
minor comments (8)
  1. [Section 3.3.2, Eq. (7)] Equation (7) writes phi_theta(z_t,t) without the structure argument S, whereas Algorithm 1 and Eq. (6) use phi_theta(z_t,S,t); please make the notation consistent.
  2. [Appendix D.2] The text reports '47,297 million preference pairs' for BindingGym, which appears to be a typo; the intended scale should be clarified.
  3. [Section 4.1.4] The ZScore formula is garbled and should be rewritten clearly so that the normalization across models and methods is unambiguous.
  4. [Table 2] The claim that EnerBridge-DPO produces 'significantly lower' energies is not backed by statistical tests; please report paired significance tests or confidence intervals given the high standard deviations.
  5. [Section 4.2] The criteria for selecting the 26 protein complexes are not described; please specify the selection protocol or report results on the full test set to rule out cherry-picking.
  6. [Figure 2] Figure 2 shows only four complexes; it would be helpful to report aggregate structural validity metrics (e.g., TM-score/RMSD distributions) for the designed sequences across all 26 complexes.
  7. [Tables 4 and 5] The checkmark notation in the ablation tables is ambiguous; please spell out the exact model configuration for each row.
  8. [General] The paper does not include a code or data availability statement; adding one would aid reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found; the low-energy and ΔΔG claims are grounded in external predictors or held-out supervised folds, and the self-cited BindingGym dataset is not load-bearing.

full rationale

The derivation chain is not circular. The central energy-design claim in Table 2 is evaluated with FoldX, Rosetta, and BA-Cycle, none of which were used to construct the DPO preference pairs, so the reported energy reductions are not baked into the training objective by construction. The ΔΔG results in Table 3 are produced by an explicitly supervised energy-constraint loss (Eq. 9) trained on experimental labels and evaluated by three-fold cross-validation; this is ordinary supervised prediction, not a fitted parameter renamed as an emergent first-principles result. The only self-citation, BindingGym (Ref. [40], whose authors include S. Zheng), supplies experimental DMS preference data but is supplemented by SKEMPI and is not load-bearing for the main claims. The main validity risk in the paper is non-circular: Eq. (7) substitutes an L2 prediction error for a log-likelihood ratio, and Appendix B.4 admits this is a simplification and proxy, while Eq. (7)'s sampling notation q(z_t|X,S) conflicts with Algorithm 1's endpoint-conditioned sampling q(z_t|X,S,Y_w/Y_l). These are unsupported approximations that could undermine the DPO interpretation, but they are not an identity between inputs and outputs and therefore do not constitute circularity.

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

The central method rests on four main assumptions: the Markov bridge formulation, the L2 proxy for DPO log-likelihoods, the Boltzmann-aligned ΔΔG formula, and the Bradley-Terry preference model. The most fragile is the L2 proxy, which is introduced ad hoc and is not derived from the cross-entropy pretraining objective. Free parameters include the learnable kBT, the energy loss weight λ, and unspecified DPO hyperparameters β and ω(λ_t). No new physical entities are invented.

free parameters (4)
  • kBT scaling factor = not reported (learnable)
    In Eq 8, kBT converts log-likelihood ratios into ΔΔG units and is learned by fitting Lenergy on SKEMPI labels. Without this fitted scale, the predicted ΔΔG values have no physical units.
  • λ energy-loss weight = 0.5 (Eq 10)
    Chosen by hand to balance the Bridge-DPO loss and the energy constraint loss. No sensitivity analysis or tuning procedure is reported, so the effect of this choice on the results is unknown.
  • β DPO divergence penalty = not reported
    Controls how far the fine-tuned model may deviate from the reference model in Eq 7. The value is never given, leaving the reported loss under-specified.
  • ω(λ_t) timestep weighting = not reported
    Appears inside the DPO loss in Eq 7 but is never defined or assigned a value, making the exact objective used in training unclear.
assumptions (4)
  • domain assumption The Markov bridge transition in Eq 5 with categorical states is a valid generative model for protein sequences.
    Adopted from Bridge-IF (Section 3.2.2). If this transition model is inappropriate for protein sequence space, the pretrained reference model and the DPO loss built on it are both compromised.
  • ad hoc to paper The L2 prediction error ||Y − φθ(zt, S, t)||^2 acts as a valid proxy for the bridge log-likelihood ratio in the DPO objective.
    Appendix B.4 introduces this substitution without proof, despite the pretraining objective in Eq 6 being cross-entropy. The entire Bridge-DPO loss in Eq 7 depends on this unverified equivalence.
  • domain assumption ΔΔG can be computed from inverse-folding log-likelihood ratios via Boltzmann alignment as in Eq 8.
    Borrowed from BA-DDG [39]. If the log-likelihood-to-free-energy mapping is not physically valid, the energy constraint loss and the ΔΔG prediction results lose their meaning.
  • standard math Energy-based preference pairs follow a Bradley-Terry model of rewards.
    Standard DPO assumption from [17], used in Section 3.3.1. It is not tested on these protein sequence preferences.

how reviews work

0 comments
Cite this review

Pith. "Pith review of EnerBridge-DPO: Energy-Guided Protein Inverse Folding with Markov Bridges and Direct Preference Optimization." pith.science (2026). https://pith.science/paper/FCGTJVN6

@misc{pith2026250609496,
  author       = {Pith},
  title        = {Pith review of: EnerBridge-DPO: Energy-Guided Protein Inverse Folding with Markov Bridges and Direct Preference Optimization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FCGTJVN6}},
  note         = {Machine review of arXiv:2506.09496}
}
abstract

Designing protein sequences with optimal energetic stability is a key challenge in protein inverse folding, as current deep learning methods are primarily trained by maximizing sequence recovery rates, often neglecting the energy of the generated sequences. This work aims to overcome this limitation by developing a model that directly generates low-energy, stable protein sequences. We propose EnerBridge-DPO, a novel inverse folding framework focused on generating low-energy, high-stability protein sequences. Our core innovation lies in: First, integrating Markov Bridges with Direct Preference Optimization (DPO), where energy-based preferences are used to fine-tune the Markov Bridge model. The Markov Bridge initiates optimization from an information-rich prior sequence, providing DPO with a pool of structurally plausible sequence candidates. Second, an explicit energy constraint loss is introduced, which enhances the energy-driven nature of DPO based on prior sequences, enabling the model to effectively learn energy representations from a wealth of prior knowledge and directly predict sequence energy values, thereby capturing quantitative features of the energy landscape. Our evaluations demonstrate that EnerBridge-DPO can design protein complex sequences with lower energy while maintaining sequence recovery rates comparable to state-of-the-art models, and accurately predicts $\Delta \Delta G$ values between various sequences.

Figures

Figures reproduced from arXiv: 2506.09496 by the authors.

Figure 1
Figure 1. Overview of EnerBridge-DPO. (a) In Stage 1, we pre-train the model on the structure [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Folding comparison of designed sequences (in red) and the native sequences (in blue). [PITH_FULL_IMAGE:figures/full_fig_p016_2.png] view at source ↗
Figure 3
Figure 3. Comparison of correlations between experimental [PITH_FULL_IMAGE:figures/full_fig_p016_3.png] view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

55 extracted references · 36 canonical work pages

  1. [1]

    A new age in protein design empowered by deep learning,

    H. Khakzad, I. Igashov, A. Schneuing, C. Goverde, M. Bronstein, and B. Correia, “A new age in protein design empowered by deep learning,” Cell Systems, vol. 14, no. 11, pp. 925–939, 2023

    work page 2023

  2. [2]

    Sparks of function by de novo protein design,

    A. E. Chu, T. Lu, and P.-S. Huang, “Sparks of function by de novo protein design,”Nature biotechnology, vol. 42, no. 2, pp. 203–215, 2024

    work page 2024

  3. [3]

    Highly accurate protein structure prediction with alphafold,

    J. Jumper, R. Evans, A. Pritzel, T. Green, M. Figurnov, O. Ronneberger, K. Tunyasuvunakool, R. Bates, A. Žídek, A. Potapenko et al., “Highly accurate protein structure prediction with alphafold,” nature, vol. 596, no. 7873, pp. 583–589, 2021

    work page 2021

  4. [4]

    Representing long-range context for graph neural networks with global attention,

    Z. Wu, P. Jain, M. Wright, A. Mirhoseini, J. E. Gonzalez, and I. Stoica, “Representing long-range context for graph neural networks with global attention,” Advances in neural information processing systems , vol. 34, pp. 13 266–13 279, 2021

    work page 2021

  5. [5]

    Robust deep learning–based protein sequence design using proteinmpnn,

    J. Dauparas, I. Anishchenko, N. Bennett, H. Bai, R. J. Ragotte, L. F. Milles, B. I. Wicky, A. Courbet, R. J. de Haas, N. Bethel et al., “Robust deep learning–based protein sequence design using proteinmpnn,” Science, vol. 378, no. 6615, pp. 49–56, 2022

    work page 2022

  6. [6]

    Pifold: Toward effective and efficient protein inverse folding,

    Z. Gao, C. Tan, P. Chacón, and S. Z. Li, “Pifold: Toward effective and efficient protein inverse folding,” arXiv preprint arXiv:2209.12643, 2022

    arXiv 2022

  7. [7]

    The graph neural network model,

    F. Scarselli, M. Gori, A. C. Tsoi, M. Hagenbuchner, and G. Monfardini, “The graph neural network model,” IEEE transactions on neural networks, vol. 20, no. 1, pp. 61–80, 2008

    2008

  8. [8]

    Structure-informed language models are protein designers,

    Z. Zheng, Y . Deng, D. Xue, Y . Zhou, F. Ye, and Q. Gu, “Structure-informed language models are protein designers,” in International conference on machine learning. PMLR, 2023, pp. 42 317–42 338

    work page 2023

Show all 55 references
  1. [9]

    Knowledge-design: Pushing the limit of protein design via knowledge refinement,

    Z. Gao, C. Tan, and S. Z. Li, “Knowledge-design: Pushing the limit of protein design via knowledge refinement,” arXiv preprint arXiv:2305.15151, 2023

    2023 arXiv

  2. [10]

    Graph denoising diffusion for inverse protein folding,

    K. Yi, B. Zhou, Y . Shen, P. Liò, and Y . Wang, “Graph denoising diffusion for inverse protein folding,” Advances in Neural Information Processing Systems, vol. 36, pp. 10 238–10 257, 2023

    2023

  3. [11]

    Bridge-if: Learning inverse protein folding with markov bridges,

    Y . Zhu, J. Wu, Q. Li, J. Yan, M. Yin, W. Wu, M. Li, J. Ye, Z. Wang, and J. Wu, “Bridge-if: Learning inverse protein folding with markov bridges,”arXiv preprint arXiv:2411.02120, 2024

    2024 arXiv

  4. [12]

    Diffusion model with representation alignment for protein inverse folding,

    C. Wang, Y . Zhou, Z. Zhai, J. Shen, and K. Zhang, “Diffusion model with representation alignment for protein inverse folding,”arXiv preprint arXiv:2412.09380, 2024

    2024 arXiv

  5. [13]

    Cath–a hierarchic classification of protein domain structures,

    C. A. Orengo, A. D. Michie, S. Jones, D. T. Jones, M. B. Swindells, and J. M. Thornton, “Cath–a hierarchic classification of protein domain structures,” Structure, vol. 5, no. 8, pp. 1093–1109, 1997

    1997

  6. [14]

    Protein sequence design by conformational landscape optimization,

    C. Norn, B. I. Wicky, D. Juergens, S. Liu, D. Kim, D. Tischer, B. Koepnick, I. Anishchenko, F. Players, D. Baker et al., “Protein sequence design by conformational landscape optimization,” Proceedings of the National Academy of Sciences, vol. 118, no. 11, p. e2017228118, 2021

    2021

  7. [15]

    Retrobridge: Modeling retrosynthesis with markov bridges,

    I. Igashov, A. Schneuing, M. Segler, M. Bronstein, and B. Correia, “Retrobridge: Modeling retrosynthesis with markov bridges,” arXiv preprint arXiv:2308.16212, 2023

    2023 arXiv

  8. [16]

    Disco: Diffusion schrödinger bridge for molecular conformer optimization,

    D. Lee, D. Lee, D. Bang, and S. Kim, “Disco: Diffusion schrödinger bridge for molecular conformer optimization,” in Proceedings of the AAAI Conference on Artificial Intelligence, vol. 38, no. 12, 2024, pp. 13 365–13 373

    2024

  9. [17]

    Direct preference opti- mization: Your language model is secretly a reward model,

    R. Rafailov, A. Sharma, E. Mitchell, C. D. Manning, S. Ermon, and C. Finn, “Direct preference opti- mization: Your language model is secretly a reward model,”Advances in Neural Information Processing Systems, vol. 36, pp. 53 728–53 741, 2023. 10

    2023

  10. [18]

    Generative models for graph-based protein design,

    J. Ingraham, V . Garg, R. Barzilay, and T. Jaakkola, “Generative models for graph-based protein design,” Advances in neural information processing systems, vol. 32, 2019

    2019

  11. [19]

    Alphadesign: A graph protein design method and benchmark on alphafolddb,

    Z. Gao, C. Tan, and S. Z. Li, “Alphadesign: A graph protein design method and benchmark on alphafolddb,” arXiv preprint arXiv:2202.01079, 2022

    2022 arXiv

  12. [20]

    Generative de novo protein design with global context,

    C. Tan, Z. Gao, J. Xia, B. Hu, and S. Z. Li, “Generative de novo protein design with global context,”arXiv preprint arXiv:2204.10673, 2022

    2022 arXiv

  13. [21]

    Structgnn: an efficient graph neural network framework for static structural analysis,

    Y .-T. Chou, W.-T. Chang, J. G. Jean, K.-H. Chang, Y .-N. Huang, and C.-S. Chen, “Structgnn: an efficient graph neural network framework for static structural analysis,” Computers & Structures , vol. 299, p. 107385, 2024

    2024

  14. [22]

    Learning inverse folding from millions of predicted structures,

    C. Hsu, R. Verkuil, J. Liu, Z. Lin, B. Hie, T. Sercu, A. Lerer, and A. Rives, “Learning inverse folding from millions of predicted structures,” in International conference on machine learning. PMLR, 2022, pp. 8946–8970

    2022

  15. [23]

    Illuminating protein space with a programmable generative model,

    J. B. Ingraham, M. Baranov, Z. Costello, K. W. Barber, W. Wang, A. Ismail, V . Frappier, D. M. Lord, C. Ng-Thow-Hing, E. R. Van Vlacket al., “Illuminating protein space with a programmable generative model,” Nature, vol. 623, no. 7989, pp. 1070–1078, 2023

    2023

  16. [24]

    Accurate and robust protein sequence design with carbondesign,

    M. Ren, C. Yu, D. Bu, and H. Zhang, “Accurate and robust protein sequence design with carbondesign,” Nature Machine Intelligence, vol. 6, no. 5, pp. 536–547, 2024

    2024

  17. [25]

    Structured denoising diffusion models in discrete state-spaces,

    J. Austin, D. D. Johnson, J. Ho, D. Tarlow, and R. Van Den Berg, “Structured denoising diffusion models in discrete state-spaces,” Advances in neural information processing systems, vol. 34, pp. 17 981–17 993, 2021

    2021

  18. [26]

    Protein design using physics informed neural networks,

    S. I. Omar, C. Keasar, A. J. Ben-Sasson, and E. Haber, “Protein design using physics informed neural networks,” Biomolecules, vol. 13, no. 3, p. 457, 2023

    2023

  19. [27]

    Machine learning for evolutionary- based and physics-inspired protein design: Current and future synergies,

    C. Malbranke, D. Bikard, S. Cocco, R. Monasson, and J. Tubiana, “Machine learning for evolutionary- based and physics-inspired protein design: Current and future synergies,” Current Opinion in Structural Biology, vol. 80, p. 102571, 2023

    2023

  20. [28]

    Protein structure prediction using rosetta,

    C. A. Rohl, C. E. Strauss, K. M. Misura, and D. Baker, “Protein structure prediction using rosetta,” in Methods in enzymology. Elsevier, 2004, vol. 383, pp. 66–93

    2004

  21. [29]

    Deep reinforcement learning from human preferences,

    P. F. Christiano, J. Leike, T. Brown, M. Martic, S. Legg, and D. Amodei, “Deep reinforcement learning from human preferences,” Advances in neural information processing systems, vol. 30, 2017

    2017

  22. [30]

    Fine-tuning language models from human preferences,

    D. M. Ziegler, N. Stiennon, J. Wu, T. B. Brown, A. Radford, D. Amodei, P. Christiano, and G. Irving, “Fine-tuning language models from human preferences,” arXiv preprint arXiv:1909.08593, 2019

    1909 arXiv

  23. [31]

    Training language models to follow instructions with human feedback,

    L. Ouyang, J. Wu, X. Jiang, D. Almeida, C. Wainwright, P. Mishkin, C. Zhang, S. Agarwal, K. Slama, A. Ray et al., “Training language models to follow instructions with human feedback,” Advances in neural information processing systems, vol. 35, pp. 27 730–27 744, 2022

    2022

  24. [32]

    Safe rlhf: Safe reinforcement learning from human feedback,

    J. Dai, X. Pan, R. Sun, J. Ji, X. Xu, M. Liu, Y . Wang, and Y . Yang, “Safe rlhf: Safe reinforcement learning from human feedback,” arXiv preprint arXiv:2310.12773, 2023

    2023 arXiv

  25. [33]

    Rlaif: Scaling reinforcement learning from human feedback with ai feedback,

    H. Lee, S. Phatale, H. Mansoor, K. R. Lu, T. Mesnard, J. Ferret, C. Bishop, E. Hall, V . Carbune, and A. Rastogi, “Rlaif: Scaling reinforcement learning from human feedback with ai feedback,” 2023

    2023

  26. [34]

    Dpok: Reinforcement learning for fine-tuning text-to-image diffusion models,

    Y . Fan, O. Watkins, Y . Du, H. Liu, M. Ryu, C. Boutilier, P. Abbeel, M. Ghavamzadeh, K. Lee, and K. Lee, “Dpok: Reinforcement learning for fine-tuning text-to-image diffusion models,” Advances in Neural Information Processing Systems, vol. 36, pp. 79 858–79 885, 2023

    2023

  27. [35]

    Training diffusion models with reinforcement learning,

    K. Black, M. Janner, Y . Du, I. Kostrikov, and S. Levine, “Training diffusion models with reinforcement learning,” arXiv preprint arXiv:2305.13301, 2023

    2023 arXiv

  28. [36]

    Diffusion model alignment using direct preference optimization,

    B. Wallace, M. Dang, R. Rafailov, L. Zhou, A. Lou, S. Purushwalkam, S. Ermon, C. Xiong, S. Joty, and N. Naik, “Diffusion model alignment using direct preference optimization,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 8228–8238

    2024

  29. [37]

    Biological structure and function emerge from scaling unsupervised learning to 250 million protein sequences,

    A. Rives, J. Meier, T. Sercu, S. Goyal, Z. Lin, J. Liu, D. Guo, M. Ott, C. L. Zitnick, J. Maet al., “Biological structure and function emerge from scaling unsupervised learning to 250 million protein sequences,” Proceedings of the National Academy of Sciences, vol. 118, no. 15...

    2021

  30. [38]

    Scalable diffusion models with transformers,

    W. Peebles and S. Xie, “Scalable diffusion models with transformers,” inProceedings of the IEEE/CVF international conference on computer vision, 2023, pp. 4195–4205

    2023

  31. [39]

    Boltzmann-aligned inverse folding model as a predictor of mutational effects on protein-protein interactions,

    X. Jiao, W. Mao, W. Jin, P. Yang, H. Chen, and C. Shen, “Boltzmann-aligned inverse folding model as a predictor of mutational effects on protein-protein interactions,” arXiv preprint arXiv:2410.09543, 2024

    2024 arXiv

  32. [40]

    Bindinggym: A large-scale mutational dataset toward deciphering protein-protein interactions,

    W. Lu, J. Zhang, M. Gu, and S. Zheng, “Bindinggym: A large-scale mutational dataset toward deciphering protein-protein interactions,” bioRxiv, pp. 2024–12, 2024

    2024

  33. [41]

    Skempi 2.0: an updated benchmark of changes in protein–protein binding energy, kinetics and thermodynamics upon mutation,

    J. Jankauskait˙e, B. Jiménez-García, J. Dapk ¯unas, J. Fernández-Recio, and I. H. Moal, “Skempi 2.0: an updated benchmark of changes in protein–protein binding energy, kinetics and thermodynamics upon mutation,” Bioinformatics, vol. 35, no. 3, pp. 462–469, 2019

    2019

  34. [42]

    Rotamer density estimator is an unsupervised learner of the effect of mutations on protein-protein interaction,

    S. Luo, Y . Su, Z. Wu, C. Su, J. Peng, and J. Ma, “Rotamer density estimator is an unsupervised learner of the effect of mutations on protein-protein interaction,” bioRxiv, pp. 2023–02, 2023

    2023

  35. [43]

    Surface-vqmae: Vector-quantized masked auto-encoders on molecular surfaces,

    F. Wu and S. Z. Li, “Surface-vqmae: Vector-quantized masked auto-encoders on molecular surfaces,” in Forty-first International Conference on Machine Learning, 2024

    2024

  36. [44]

    Improved denoising diffusion probabilistic models,

    A. Q. Nichol and P. Dhariwal, “Improved denoising diffusion probabilistic models,” in International conference on machine learning. PMLR, 2021, pp. 8162–8171

    2021

  37. [45]

    Adam: A method for stochastic optimization,

    D. P. Kingma and J. Ba, “Adam: A method for stochastic optimization,”arXiv preprint arXiv:1412.6980, 2014

    2014 arXiv

  38. [46]

    Attention is all you need,

    A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, Ł. Kaiser, and I. Polosukhin, “Attention is all you need,”Advances in neural information processing systems, vol. 30, 2017

    2017

  39. [47]

    Learning from protein structure with geometric vector perceptrons,

    B. Jing, S. Eismann, P. Suriana, R. J. Townshend, and R. Dror, “Learning from protein structure with geometric vector perceptrons,” arXiv preprint arXiv:2009.01411, 2020

    2009 arXiv

  40. [48]

    Global-context aware generative protein design,

    C. Tan, Z. Gao, J. Xia, B. Hu, and S. Z. Li, “Global-context aware generative protein design,” inICASSP 2023-2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2023, pp. 1–5

    2023

  41. [49]

    Deep learning guided optimization of human antibody against sars-cov-2 variants with broad neutralization,

    S. Shan, S. Luo, Z. Yang, J. Hong, Y . Su, F. Ding, L. Fu, C. Li, P. Chen, J. Maet al., “Deep learning guided optimization of human antibody against sars-cov-2 variants with broad neutralization,” Proceedings of the National Academy of Sciences, vol. 119, no. 11, p. e2122954119, 2022

    2022

  42. [50]

    The hdock server for integrated protein–protein docking,

    Y . Yan, H. Tao, J. He, and S.-Y . Huang, “The hdock server for integrated protein–protein docking,”Nature protocols, vol. 15, no. 5, pp. 1829–1852, 2020

    2020

  43. [51]

    Language models of protein sequences at the scale of evolution enable accurate structure prediction,

    Z. Lin, H. Akin, R. Rao, B. Hie, Z. Zhu, W. Lu, A. dos Santos Costa, M. Fazel-Zarandi, T. Sercu, S. Candido et al., “Language models of protein sequences at the scale of evolution enable accurate structure prediction,” BioRxiv, vol. 2022, p. 500902, 2022

    2022

  44. [52]

    Learning to predict mutation effects of protein-protein interactions by microenvironment-aware hierarchical prompt learning,

    L. Wu, Y . Tian, H. Lin, Y . Huang, S. Li, N. V . Chawla, and S. Z. Li, “Learning to predict mutation effects of protein-protein interactions by microenvironment-aware hierarchical prompt learning,”arXiv preprint arXiv:2405.10348, 2024

    2024 arXiv

  45. [53]

    Multi-level interaction modeling for protein mutational effect prediction,

    Y . Mo, X. Hong, B. Gao, Y . Jia, and Y . Lan, “Multi-level interaction modeling for protein mutational effect prediction,” arXiv preprint arXiv:2405.17802, 2024

    2024 arXiv

  46. [54]

    Foldx 5.0: working with rna, small molecules and a new graphical interface,

    J. Delgado, L. G. Radusky, D. Cianferoni, and L. Serrano, “Foldx 5.0: working with rna, small molecules and a new graphical interface,” Bioinformatics, vol. 35, no. 20, pp. 4168–4169, 2019

    2019

  47. [55]

    The rosetta all-atom energy function for macromolecular modeling and design,

    R. F. Alford, A. Leaver-Fay, J. R. Jeliazkov, M. J. O’Meara, F. P. DiMaio, H. Park, M. V . Shapovalov, P. D. Renfrew, V . K. Mulligan, K. Kappelet al., “The rosetta all-atom energy function for macromolecular modeling and design,” Journal of chemical theory and computation, vo...

    2017

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.