Controllable protein design with particle-based Feynman-Kac steering

arxiv: 2511.09216 · v2 · submitted 2025-11-12 · 💻 cs.LG · q-bio.QM· stat.ML

Controllable protein design with particle-based Feynman-Kac steering

Erik Hartman , Jonas Wallin , Johan Malmstr\"om , Jimmy Olsson This is my paper

Pith reviewed 2026-05-17 23:29 UTC · model grok-4.3

classification 💻 cs.LG q-bio.QMstat.ML

keywords protein designdiffusion modelsFeynman-Kac steeringbinder designcontrollable generationRFdiffusiongenerative models

0 comments p. Extension

The pith

Particle-based Feynman-Kac steering lets diffusion protein models be guided by non-differentiable rewards like interface energetics.

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

This paper shows how to steer RFdiffusion using the Feynman-Kac framework by building guiding potentials from ProteinMPNN sequence predictions and structural relaxation. The steering process improves predicted interface energetics and raises binder designability by 89.5 percent compared with unguided sampling. A reader would care because it supplies a general way to impose arbitrary design goals on diffusion generators without requiring differentiable objectives or model retraining.

Core claim

By developing guiding potentials that leverage ProteinMPNN and structural relaxation, the authors demonstrate that Feynman-Kac steering can be applied to RFdiffusion to consistently improve predicted interface energetics and increase binder designability by 89.5 percent, thereby establishing a model-independent framework for controllable protein generation toward arbitrary non-differentiable objectives.

What carries the argument

Particle-based Feynman-Kac steering, which propagates weighted particle trajectories through the diffusion process according to user-defined reward potentials derived from ProteinMPNN and relaxation.

If this is right

Steered designs exhibit consistently better predicted interface energetics than unguided ones.
Binder designability rises by 89.5 percent under the same sampling budget.
The method supplies a general route for imposing arbitrary non-differentiable objectives on any diffusion-based protein generator.
Steering remains effective without retraining the underlying diffusion model.

Where Pith is reading between the lines

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

The same steering machinery could be applied to other diffusion or flow-matching backbones for proteins or small molecules.
Design campaigns could incorporate experimental feedback directly into the reward function without gradient requirements.
Higher designability may reduce the number of candidates that need to be synthesized and tested in the wet lab.

Load-bearing premise

The guiding potentials derived from ProteinMPNN and structural relaxation accurately reflect the target properties and do not introduce systematic biases that would invalidate the reported gains when tested on independent experimental data.

What would settle it

Laboratory synthesis and binding-affinity measurement of steered versus unsteered binder designs, with the steered set showing measurably higher success rate on the experimental assay.

read the original abstract

Proteins underpin most biological function, and the ability to design them with tailored structures and properties is central to advances in biotechnology. Diffusion-based generative models have emerged as powerful tools for protein design, but steering them toward proteins with specified properties remains challenging. The Feynman-Kac (FK) framework provides a principled way to guide diffusion models using user-defined rewards. In this paper, we enable FK-based steering of RFdiffusion through the development of guiding potentials that leverage ProteinMPNN and structural relaxation to guide the diffusion process towards desired properties. We show that steering can be used to consistently improve predicted interface energetics and increase binder designability by $89.5\%$. Together, these results establish that diffusion-based protein design can be effectively steered toward arbitrary, non-differentiable objectives, providing a model-independent framework for controllable protein generation.

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.

Desk Editor's Note private letter to a colleague

FK steering on RFdiffusion works as a general way to add non-differentiable control but the 89.5% designability lift looks vulnerable to circular evaluation through ProteinMPNN.

read the letter

The main thing here is that they adapt particle-based Feynman-Kac steering to RFdiffusion by building guiding potentials from ProteinMPNN and structural relaxation, then report consistent gains on predicted interface energetics plus an 89.5% lift in binder designability. That combination is the concrete new piece; FK steering itself has been used elsewhere, but the specific integration with these protein tools and the relaxation step for handling non-differentiable rewards is not in the prior work they cite. The method gives a model-agnostic route to steer diffusion generators toward user-specified objectives, which is practically useful when the target property lacks a clean gradient. They show the steering actually moves the samples in the intended direction on the metrics they track, and the framework looks implementable if you already have the base diffusion and ProteinMPNN models running. The soft spot is the evaluation setup. Designability is measured with the same class of ProteinMPNN-derived quantities that supply the guiding potential, so the reported improvement risks being partly an artifact of optimizing the evaluation function itself rather than an independent demonstration of better binders. The abstract does not mention orthogonal predictors, held-out test sets with different scoring, or any experimental validation, which leaves the numbers hard to interpret. Baselines and error bars are also not detailed enough in the summary to rule out selection effects. This paper is mainly for people already working on controllable generative models for proteins who need ideas for adding steering without retraining. A reader focused on diffusion methods or binder design could extract the technical recipe and try it, but anyone outside that niche will probably wait for clearer separation between steering signal and evaluation metric. It deserves peer review so the methods section can be checked for independent validation and to see whether the circularity concern actually materializes once the full protocol is visible.

Referee Report

1 major / 0 minor

Summary. The paper introduces particle-based Feynman-Kac steering to guide RFdiffusion models for controllable protein design. Guiding potentials are constructed from ProteinMPNN and structural relaxation to steer the generative process toward improved interface energetics and higher binder designability, with a reported 89.5% increase in the latter.

Significance. If the quantitative gains are shown to arise from an independent evaluation protocol rather than optimization of the same ProteinMPNN-derived quantities used for steering, the work would provide a principled, model-independent method for incorporating arbitrary non-differentiable rewards into diffusion-based protein generators. This could meaningfully extend controllable design capabilities beyond current gradient-based or classifier-guidance approaches.

major comments (1)

[Abstract] Abstract and Results sections: the headline 89.5% increase in binder designability is obtained by steering with a ProteinMPNN-derived guiding potential. The manuscript must explicitly define the designability metric (e.g., ProteinMPNN recovery rate, sequence likelihood under the same model, or post-relaxation interface energy) and demonstrate that this metric is computed independently of the steering objective; otherwise the reported lift risks being at least partly circular.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their constructive feedback, which highlights an important point about metric clarity and independence. We address the major comment below and will revise the manuscript accordingly.

read point-by-point responses

Referee: [Abstract] Abstract and Results sections: the headline 89.5% increase in binder designability is obtained by steering with a ProteinMPNN-derived guiding potential. The manuscript must explicitly define the designability metric (e.g., ProteinMPNN recovery rate, sequence likelihood under the same model, or post-relaxation interface energy) and demonstrate that this metric is computed independently of the steering objective; otherwise the reported lift risks being at least partly circular.

Authors: We agree that the designability metric must be defined explicitly and that its independence from the steering objective should be demonstrated to eliminate any risk of circularity. In the revised manuscript we will add a precise definition in the Abstract and expand the Results and Methods sections to state that binder designability is quantified as the fraction of generated designs that, after an independent post-generation protocol consisting of structural relaxation with OpenMM followed by Rosetta interface energy scoring, achieve an energy below a fixed threshold. This evaluation pipeline is distinct from the ProteinMPNN-based guiding potential, which is applied only during the particle-based Feynman-Kac steering steps to shape the diffusion trajectory; the final metric uses a separate relaxation engine and scoring function not employed in the steering objective. We will include a schematic of the evaluation workflow and a short ablation confirming that the 89.5% lift persists when designability is measured with an alternative scorer (e.g., AF2 pLDDT on the binder interface). revision: yes

Circularity Check

0 steps flagged

No significant circularity: external ProteinMPNN rewards treated as independent inputs

full rationale

The paper's core contribution is the application of particle-based Feynman-Kac steering to RFdiffusion using guiding potentials derived from ProteinMPNN and structural relaxation. The reported 89.5% increase in binder designability and improvements in predicted interface energetics are presented as empirical outcomes of this steering process. No equation or derivation step in the manuscript reduces these gains to a quantity fitted inside the paper or to a self-referential definition of the guiding potential itself. ProteinMPNN functions as an external, pre-existing tool whose outputs serve as rewards rather than being redefined or fitted within the current work. Evaluation metrics appear to draw on the same class of predictors but are not shown by the paper's own text to be identical to the steering objective by construction, leaving the central claims with independent empirical content. This configuration is consistent with a low circularity score as the method remains self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Based on the abstract alone, the central claim rests on the assumption that ProteinMPNN scores and relaxation energies serve as faithful, unbiased rewards for the target properties. No explicit free parameters or invented entities are described.

axioms (1)

domain assumption Feynman-Kac framework can be applied to steer diffusion trajectories in protein latent space using external reward functions
Invoked to justify the guiding procedure

pith-pipeline@v0.9.0 · 5447 in / 1225 out tokens · 33262 ms · 2026-05-17T23:29:58.101246+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

FK steering ... reweighting ... potentials Gt ... reward function ... ProteinMPNN and structural relaxation ... interface free energy reward
IndisputableMonolith/Foundation/RealityFromDistinction.lean reality_from_one_distinction unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

steering ... improves predicted interface energetics and increase binder designability by 89.5%

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

The tractability landscape of diffusion alignment: regularization, rewards, and computational primitives
cs.LG 2026-05 unverdicted novelty 7.0

The choice of closeness measure in diffusion reward alignment determines the computational primitives and tractable reward classes, with linear exponential tilts sufficing for KL with convex rewards and proximal oracl...

Reference graph

Works this paper leans on

23 extracted references · 23 canonical work pages · cited by 1 Pith paper

[1]

L.et al.De novo design of protein structure and function with RFdiffusion.Nature620, 1089–1100, DOI: 10.1038/s41586-023-06415-8 (2023)

Joseph L. Watson, David Juergens, Nathaniel R. Bennett, Brian L. Trippe, Jason Yim, Helen E. Eisenach, Woody Ahern, Andrew J. Borst, Robert J. Ragotte, Lukas F. Milles, Basile I. M. Wicky, Nikita Hanikel, Samuel J. Pellock, Alexis Courbet, William Sheffler, Jue Wang, Preetham Venkatesh, Isaac Sappington, Susana Vázquez Torres, Anna Lauko, Valentin De Bort...

work page doi:10.1038/s41586-023-06415-8 2023
[2]

Trippe, Valentin De Bortoli, Emile Mathieu, Arnaud Doucet, Regina Barzilay, and Tommi Jaakkola

Jason Yim, Brian L. Trippe, Valentin De Bortoli, Emile Mathieu, Arnaud Doucet, Regina Barzilay, and Tommi Jaakkola. SE(3) diffusion model with application to protein backbone generation

work page
[3]

Se (3) diffusion model with application to protein backbone generation,

doi: 10.48550/ARXIV.2302.02277. URL https://arxiv.org/abs/2302.02277

work page doi:10.48550/arxiv.2302.02277
[4]

Wu, Kevin K

Kevin E. Wu, Kevin K. Yang, Rianne van den Berg, Sarah Alamdari, James Y. Zou, Alex X. Lu, and Ava P. Amini. Protein structure generation via folding diffusion.Nature Communications, 15 (1), February 2024. ISSN 2041-1723. doi: 10.1038/s41467-024-45051-2. URL http://dx.doi.org/10. 1038/s41467-024-45051-2. 13

work page doi:10.1038/s41467-024-45051-2 2024
[5]

Trippe, Jason Yim, Doug Tischer, David Baker, Tamara Broderick, Regina Barzilay, and Tommi Jaakkola

Brian L. Trippe, Jason Yim, Doug Tischer, David Baker, Tamara Broderick, Regina Barzilay, and Tommi Jaakkola. Diffusion probabilistic modeling of protein backbones in 3D for the motif- scaffolding problem, 2022. URL https://arxiv.org/abs/2206.04119

work page arXiv 2022
[6]

Mackessy, Stefanie K

Susana Vázquez Torres, Melisa Benard Valle, Stephen P. Mackessy, Stefanie K. Menzies, Nicholas R. Casewell, Shirin Ahmadi, Nick J. Burlet, Edin Muratspahić, Isaac Sappington, Max D. Overath, Esperanza Rivera-de Torre, Jann Ledergerber, Andreas H. Laustsen, Kim Boddum, Asim K. Bera, Alex Kang, Evans Brackenbrough, Iara A. Cardoso, Edouard P. Crittenden, Re...

work page doi:10.1038/s41586-024-08393-x 2025
[7]

Carr, Justin Decarreau, Rebecca Skotheim, Ryan D

Sanela Rankovic, Kenneth D. Carr, Justin Decarreau, Rebecca Skotheim, Ryan D. Kibler, Sebastian Ols, Sangmin Lee, Jung-Ho Chun, Marti R. Tooley, Justas Dauparas, Helen E. Eisenach, Matthias Glögl, Connor Weidle, Andrew J. Borst, David Baker, and Neil P. King. Computational design of bifaceted protein nanomaterials.Nature Materials, 24(10):1635–1643, July ...

work page doi:10.1038/s41563-025-02295-7 2025
[8]

Leveille, Andi Liu, Asim K

StephenA.Rettie, DavidJuergens, VictorAdebomi, YensiFloresBueso, QinqinZhao, AlexandriaN. Leveille, Andi Liu, Asim K. Bera, Joana A. Wilms, Alina Üffing, Alex Kang, Evans Brackenbrough, Mila Lamb, Stacey R. Gerben, Analisa Murray, Paul M. Levine, Maika Schneider, Vibha Vasireddy, Sergey Ovchinnikov, Oliver H. Weiergräber, Dieter Willbold, Joshua A. Kritze...

work page doi:10.1038/s41589-025-01929-w 2025
[9]

Refining Generative Process with Discriminator Guidance in Score-based Diffusion Models, 2022

Dongjun Kim, Yeongmin Kim, Se Jung Kwon, Wanmo Kang, and Il-Chul Moon. Refining Generative Process with Discriminator Guidance in Score-based Diffusion Models, 2022. URL https://arxiv.org/abs/2211.17091

work page arXiv 2022
[10]

Improving Sample Quality of Diffusion Models Using Self-Attention Guidance, 2022

Susung Hong, Gyuseong Lee, Wooseok Jang, and Seungryong Kim. Improving Sample Quality of Diffusion Models Using Self-Attention Guidance, 2022. URL https://arxiv.org/abs/2210.00939

work page arXiv 2022
[11]

Sidney Lyayuga Lisanza, Jacob Merle Gershon, Samuel W. K. Tipps, Jeremiah Nelson Sims, Lucas Arnoldt, Samuel J. Hendel, Miriam K. Simma, Ge Liu, Muna Yase, Hongwei Wu, Claire D. Tharp, Xinting Li, Alex Kang, Evans Brackenbrough, Asim K. Bera, Stacey Gerben, Bruce J. Wittmann, Andrew C. McShan, and David Baker. Multistate and functional protein design usin...

work page doi:10.1038/s41587-024-02395-w 2024
[12]

A general framework for inference-time scaling and steering of diffusion models

Raghav Singhal, Zachary Horvitz, Ryan Teehan, Mengye Ren, Zhou Yu, Kathleen McKeown, and Rajesh Ranganath. A General Framework for Inference-time Scaling and Steering of Diffusion Models, 2025. URL https://arxiv.org/abs/2501.06848

work page arXiv 2025
[13]

Feynman-Kac Correctors in Diffusion: Annealing, Guidance, and Product of Experts, 2025

Marta Skreta, Tara Akhound-Sadegh, Viktor Ohanesian, Roberto Bondesan, Alán Aspuru-Guzik, Arnaud Doucet, Rob Brekelmans, Alexander Tong, and Kirill Neklyudov. Feynman-Kac Correctors in Diffusion: Annealing, Guidance, and Product of Experts, 2025. URL https://arxiv.org/abs/25 03.02819

work page 2025
[14]

Springer New York, 2004

Pierre Del Moral.Feynman-Kac Formulae. Springer New York, 2004. ISBN 9781468493931. doi: 10.1007/978-1-4684-9393-1. URL http://dx.doi.org/10.1007/978-1-4684-9393-1

work page doi:10.1007/978-1-4684-9393-1 2004
[15]

doi: 10.1007/978-3-030-47845-2

Nicolas Chopin and Omiros Papaspiliopoulos.An Introduction to Sequential Monte Carlo. Springer International Publishing, 2020. ISBN 9783030478452. doi: 10.1007/978-3-030-47845-2. URL http://dx.doi.org/10.1007/978-3-030-47845-2. 14

work page doi:10.1007/978-3-030-47845-2 2020
[16]

Trippe, Christian A

Luhuan Wu, Brian L. Trippe, Christian A. Naesseth, David M. Blei, and John P. Cunningham. Practical and Asymptotically Exact Conditional Sampling in Diffusion Models. 2023. doi: 10.48550/ARXIV.2306.17775. URL https://arxiv.org/abs/2306.17775

work page doi:10.48550/arxiv.2306.17775 2023
[17]

Dauparas, I

J. Dauparas, I. Anishchenko, N. Bennett, H. Bai, R. J. Ragotte, L. F. Milles, B. I. M. Wicky, A. Courbet, R. J. de Haas, N. Bethel, P. J. Y. Leung, T. F. Huddy, S. Pellock, D. Tischer, F. Chan, B. Koepnick, H. Nguyen, A. Kang, B. Sankaran, A. K. Bera, N. P. King, and D. Baker. Robust deep learning–based protein sequence design using ProteinMPNN.Science, 3...

work page doi:10.1126/science.add2187 2022
[18]

Alford, Andrew Leaver-Fay, Jeliazko R

Rebecca F. Alford, Andrew Leaver-Fay, Jeliazko R. Jeliazkov, Matthew J. O’Meara, Frank P. DiMaio, Hahnbeom Park, Maxim V. Shapovalov, P. Douglas Renfrew, Vikram K. Mulligan, Kalli Kappel, Jason W. Labonte, Michael S. Pacella, Richard Bonneau, Philip Bradley, Roland L. Dunbrack, Rhiju Das, David Baker, Brian Kuhlman, Tanja Kortemme, and Jeffrey J. Gray. Th...

work page doi:10.1021/acs.jctc.7b0 2017
[19]

URL http://dx.doi.org/10.1021/acs.jctc.7b00125

work page doi:10.1021/acs.jctc.7b00125
[20]

Boltz- 2: Towards accurate and efficient binding affinity prediction.bioRxiv, 2025

Saro Passaro, Gabriele Corso, Jeremy Wohlwend, Mateo Reveiz, Stephan Thaler, Vignesh Ram Somnath, Noah Getz, Tally Portnoi, Julien Roy, Hannes Stark, David Kwabi-Addo, Dominique Beaini, Tommi Jaakkola, and Regina Barzilay. Boltz-2: Towards Accurate and Efficient Binding Affinity Prediction. June 2025. doi: 10.1101/2025.06.14.659707. URL http://dx.doi.org/...

work page doi:10.1101/2025.06.14.659707 2025
[21]

Bond, Artur Schmidtchen, and Johan Malmström

Erik Hartman, Firdaus Samsudin, Peter J. Bond, Artur Schmidtchen, and Johan Malmström. Mining the endogenous peptidome for peptide binders with deep learning-driven optimization and molecular simulations. January 2025. doi: 10.1101/2025.01.20.633551. URL http://dx.doi.org/10. 1101/2025.01.20.633551

work page doi:10.1101/2025.01.20.633551 2025
[22]

Abigail S. L. Sudol, John Butler, Dylan P. Ivory, Ivo Tews, and Max Crispin. Extensive sub- strate recognition by the streptococcal antibody-degrading enzymes IdeS and EndoS.Nature Communications, 13(1), December 2022. ISSN 2041-1723. doi: 10.1038/s41467-022-35340-z. URL http://dx.doi.org/10.1038/s41467-022-35340-z

work page doi:10.1038/s41467-022-35340-z 2022
[23]

Kagawa, Maurice R

Todd F. Kagawa, Maurice R. O’Connell, Pania Mouat, Max Paoli, Paul W. O’Toole, and Jakki C. Cooney. Model for Substrate Interactions in C5a Peptidase from Streptococcus pyogenes: A 1.9 Å Crystal Structure of the Active Form of ScpA.Journal of Molecular Biology, 386(3):754–772, February 2009. ISSN 0022-2836. doi: 10.1016/j.jmb.2008.12.074. URL http://dx.do...

work page doi:10.1016/j.jmb.2008.12.074 2009

[1] [1]

L.et al.De novo design of protein structure and function with RFdiffusion.Nature620, 1089–1100, DOI: 10.1038/s41586-023-06415-8 (2023)

Joseph L. Watson, David Juergens, Nathaniel R. Bennett, Brian L. Trippe, Jason Yim, Helen E. Eisenach, Woody Ahern, Andrew J. Borst, Robert J. Ragotte, Lukas F. Milles, Basile I. M. Wicky, Nikita Hanikel, Samuel J. Pellock, Alexis Courbet, William Sheffler, Jue Wang, Preetham Venkatesh, Isaac Sappington, Susana Vázquez Torres, Anna Lauko, Valentin De Bort...

work page doi:10.1038/s41586-023-06415-8 2023

[2] [2]

Trippe, Valentin De Bortoli, Emile Mathieu, Arnaud Doucet, Regina Barzilay, and Tommi Jaakkola

Jason Yim, Brian L. Trippe, Valentin De Bortoli, Emile Mathieu, Arnaud Doucet, Regina Barzilay, and Tommi Jaakkola. SE(3) diffusion model with application to protein backbone generation

work page

[3] [3]

Se (3) diffusion model with application to protein backbone generation,

doi: 10.48550/ARXIV.2302.02277. URL https://arxiv.org/abs/2302.02277

work page doi:10.48550/arxiv.2302.02277

[4] [4]

Wu, Kevin K

Kevin E. Wu, Kevin K. Yang, Rianne van den Berg, Sarah Alamdari, James Y. Zou, Alex X. Lu, and Ava P. Amini. Protein structure generation via folding diffusion.Nature Communications, 15 (1), February 2024. ISSN 2041-1723. doi: 10.1038/s41467-024-45051-2. URL http://dx.doi.org/10. 1038/s41467-024-45051-2. 13

work page doi:10.1038/s41467-024-45051-2 2024

[5] [5]

Trippe, Jason Yim, Doug Tischer, David Baker, Tamara Broderick, Regina Barzilay, and Tommi Jaakkola

Brian L. Trippe, Jason Yim, Doug Tischer, David Baker, Tamara Broderick, Regina Barzilay, and Tommi Jaakkola. Diffusion probabilistic modeling of protein backbones in 3D for the motif- scaffolding problem, 2022. URL https://arxiv.org/abs/2206.04119

work page arXiv 2022

[6] [6]

Mackessy, Stefanie K

Susana Vázquez Torres, Melisa Benard Valle, Stephen P. Mackessy, Stefanie K. Menzies, Nicholas R. Casewell, Shirin Ahmadi, Nick J. Burlet, Edin Muratspahić, Isaac Sappington, Max D. Overath, Esperanza Rivera-de Torre, Jann Ledergerber, Andreas H. Laustsen, Kim Boddum, Asim K. Bera, Alex Kang, Evans Brackenbrough, Iara A. Cardoso, Edouard P. Crittenden, Re...

work page doi:10.1038/s41586-024-08393-x 2025

[7] [7]

Carr, Justin Decarreau, Rebecca Skotheim, Ryan D

Sanela Rankovic, Kenneth D. Carr, Justin Decarreau, Rebecca Skotheim, Ryan D. Kibler, Sebastian Ols, Sangmin Lee, Jung-Ho Chun, Marti R. Tooley, Justas Dauparas, Helen E. Eisenach, Matthias Glögl, Connor Weidle, Andrew J. Borst, David Baker, and Neil P. King. Computational design of bifaceted protein nanomaterials.Nature Materials, 24(10):1635–1643, July ...

work page doi:10.1038/s41563-025-02295-7 2025

[8] [8]

Leveille, Andi Liu, Asim K

StephenA.Rettie, DavidJuergens, VictorAdebomi, YensiFloresBueso, QinqinZhao, AlexandriaN. Leveille, Andi Liu, Asim K. Bera, Joana A. Wilms, Alina Üffing, Alex Kang, Evans Brackenbrough, Mila Lamb, Stacey R. Gerben, Analisa Murray, Paul M. Levine, Maika Schneider, Vibha Vasireddy, Sergey Ovchinnikov, Oliver H. Weiergräber, Dieter Willbold, Joshua A. Kritze...

work page doi:10.1038/s41589-025-01929-w 2025

[9] [9]

Refining Generative Process with Discriminator Guidance in Score-based Diffusion Models, 2022

Dongjun Kim, Yeongmin Kim, Se Jung Kwon, Wanmo Kang, and Il-Chul Moon. Refining Generative Process with Discriminator Guidance in Score-based Diffusion Models, 2022. URL https://arxiv.org/abs/2211.17091

work page arXiv 2022

[10] [10]

Improving Sample Quality of Diffusion Models Using Self-Attention Guidance, 2022

Susung Hong, Gyuseong Lee, Wooseok Jang, and Seungryong Kim. Improving Sample Quality of Diffusion Models Using Self-Attention Guidance, 2022. URL https://arxiv.org/abs/2210.00939

work page arXiv 2022

[11] [11]

Sidney Lyayuga Lisanza, Jacob Merle Gershon, Samuel W. K. Tipps, Jeremiah Nelson Sims, Lucas Arnoldt, Samuel J. Hendel, Miriam K. Simma, Ge Liu, Muna Yase, Hongwei Wu, Claire D. Tharp, Xinting Li, Alex Kang, Evans Brackenbrough, Asim K. Bera, Stacey Gerben, Bruce J. Wittmann, Andrew C. McShan, and David Baker. Multistate and functional protein design usin...

work page doi:10.1038/s41587-024-02395-w 2024

[12] [12]

A general framework for inference-time scaling and steering of diffusion models

Raghav Singhal, Zachary Horvitz, Ryan Teehan, Mengye Ren, Zhou Yu, Kathleen McKeown, and Rajesh Ranganath. A General Framework for Inference-time Scaling and Steering of Diffusion Models, 2025. URL https://arxiv.org/abs/2501.06848

work page arXiv 2025

[13] [13]

Feynman-Kac Correctors in Diffusion: Annealing, Guidance, and Product of Experts, 2025

Marta Skreta, Tara Akhound-Sadegh, Viktor Ohanesian, Roberto Bondesan, Alán Aspuru-Guzik, Arnaud Doucet, Rob Brekelmans, Alexander Tong, and Kirill Neklyudov. Feynman-Kac Correctors in Diffusion: Annealing, Guidance, and Product of Experts, 2025. URL https://arxiv.org/abs/25 03.02819

work page 2025

[14] [14]

Springer New York, 2004

Pierre Del Moral.Feynman-Kac Formulae. Springer New York, 2004. ISBN 9781468493931. doi: 10.1007/978-1-4684-9393-1. URL http://dx.doi.org/10.1007/978-1-4684-9393-1

work page doi:10.1007/978-1-4684-9393-1 2004

[15] [15]

doi: 10.1007/978-3-030-47845-2

Nicolas Chopin and Omiros Papaspiliopoulos.An Introduction to Sequential Monte Carlo. Springer International Publishing, 2020. ISBN 9783030478452. doi: 10.1007/978-3-030-47845-2. URL http://dx.doi.org/10.1007/978-3-030-47845-2. 14

work page doi:10.1007/978-3-030-47845-2 2020

[16] [16]

Trippe, Christian A

Luhuan Wu, Brian L. Trippe, Christian A. Naesseth, David M. Blei, and John P. Cunningham. Practical and Asymptotically Exact Conditional Sampling in Diffusion Models. 2023. doi: 10.48550/ARXIV.2306.17775. URL https://arxiv.org/abs/2306.17775

work page doi:10.48550/arxiv.2306.17775 2023

[17] [17]

Dauparas, I

J. Dauparas, I. Anishchenko, N. Bennett, H. Bai, R. J. Ragotte, L. F. Milles, B. I. M. Wicky, A. Courbet, R. J. de Haas, N. Bethel, P. J. Y. Leung, T. F. Huddy, S. Pellock, D. Tischer, F. Chan, B. Koepnick, H. Nguyen, A. Kang, B. Sankaran, A. K. Bera, N. P. King, and D. Baker. Robust deep learning–based protein sequence design using ProteinMPNN.Science, 3...

work page doi:10.1126/science.add2187 2022

[18] [18]

Alford, Andrew Leaver-Fay, Jeliazko R

Rebecca F. Alford, Andrew Leaver-Fay, Jeliazko R. Jeliazkov, Matthew J. O’Meara, Frank P. DiMaio, Hahnbeom Park, Maxim V. Shapovalov, P. Douglas Renfrew, Vikram K. Mulligan, Kalli Kappel, Jason W. Labonte, Michael S. Pacella, Richard Bonneau, Philip Bradley, Roland L. Dunbrack, Rhiju Das, David Baker, Brian Kuhlman, Tanja Kortemme, and Jeffrey J. Gray. Th...

work page doi:10.1021/acs.jctc.7b0 2017

[19] [19]

URL http://dx.doi.org/10.1021/acs.jctc.7b00125

work page doi:10.1021/acs.jctc.7b00125

[20] [20]

Boltz- 2: Towards accurate and efficient binding affinity prediction.bioRxiv, 2025

Saro Passaro, Gabriele Corso, Jeremy Wohlwend, Mateo Reveiz, Stephan Thaler, Vignesh Ram Somnath, Noah Getz, Tally Portnoi, Julien Roy, Hannes Stark, David Kwabi-Addo, Dominique Beaini, Tommi Jaakkola, and Regina Barzilay. Boltz-2: Towards Accurate and Efficient Binding Affinity Prediction. June 2025. doi: 10.1101/2025.06.14.659707. URL http://dx.doi.org/...

work page doi:10.1101/2025.06.14.659707 2025

[21] [21]

Bond, Artur Schmidtchen, and Johan Malmström

Erik Hartman, Firdaus Samsudin, Peter J. Bond, Artur Schmidtchen, and Johan Malmström. Mining the endogenous peptidome for peptide binders with deep learning-driven optimization and molecular simulations. January 2025. doi: 10.1101/2025.01.20.633551. URL http://dx.doi.org/10. 1101/2025.01.20.633551

work page doi:10.1101/2025.01.20.633551 2025

[22] [22]

Abigail S. L. Sudol, John Butler, Dylan P. Ivory, Ivo Tews, and Max Crispin. Extensive sub- strate recognition by the streptococcal antibody-degrading enzymes IdeS and EndoS.Nature Communications, 13(1), December 2022. ISSN 2041-1723. doi: 10.1038/s41467-022-35340-z. URL http://dx.doi.org/10.1038/s41467-022-35340-z

work page doi:10.1038/s41467-022-35340-z 2022

[23] [23]

Kagawa, Maurice R

Todd F. Kagawa, Maurice R. O’Connell, Pania Mouat, Max Paoli, Paul W. O’Toole, and Jakki C. Cooney. Model for Substrate Interactions in C5a Peptidase from Streptococcus pyogenes: A 1.9 Å Crystal Structure of the Active Form of ScpA.Journal of Molecular Biology, 386(3):754–772, February 2009. ISSN 0022-2836. doi: 10.1016/j.jmb.2008.12.074. URL http://dx.do...

work page doi:10.1016/j.jmb.2008.12.074 2009