Pith. sign in

REVIEW 4 major objections 6 minor 62 references

Steering Protein Family Design through Profile Bayesian Flow

T0 review · 4 major / 6 minor · reviewed 2026-08-08 · deepseek-v4-flash

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

Pith's one-line read ProfileBFN claims that a Bayesian flow network trained on one-hot single sequences can, when conditioned on a family's amino-acid frequency profile, generate diverse and novel sequences that retain family structure and function.

desk verdict Profile-BFN derivation is real and the single-sequence training trick is clever, but the family-generation claims rest on an untested train/sample shift and proxy metrics. read the letter →

arxiv 2502.07671 v2 pith:TEKDU7ZH submitted 2025-02-11 q-bio.BM

Jingjing Gong , Yu Pei , Siyu Long , Yuxuan Song , Zhe Zhang , Wenhao Huang , Ziyao Cao , Shuyi Zhang
show 2 more authors
Hao Zhou Wei-Ying Ma
This is my paper · ORCID
classification q-bio.BM
keywords proteinfamilydesignBayesianflownetworksMSAprofilesgenerativemodelingsequencegenerationenzymefunctionpredictionrepresentationlearningdiscretediffusion
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

ProfileBFN is a generative model that designs new members of a protein family. It extends Bayesian Flow Networks so the object being transmitted is an MSA profile—the position-by-position frequency of each amino acid among homologous sequences—instead of a single amino-acid token. The central claim is that this profile formulation makes a one-hot encoded single sequence a special case, so the model can be trained on ordinary single protein sequences and then conditioned on a family profile at generation time without ever training on large MSA datasets. If that claim is right, protein family design gets the efficiency of single-sequence pretraining plus the evolutionary signal of a family profile, producing diverse and novel sequences that still fold and function like the family. The paper reports that generated sequences recover family-specific contacts, that generated enzymes are classified by CLEAN as having the target function far more often than earlier methods, and that the learned representations transfer well to standard protein prediction tasks.

What carries the argument

The central object is the profile Bayesian flow: given a family profile $\rho$ (a probability vector over amino acids at each position), the model accumulates noisy evidence $y \sim \mathcal{N}(K\beta(t)\rho, \beta(t)C)$ with $C$ the covariance matrix of the simplex, takes the softmax to get an information state $\theta = \mathrm{softmax}(y)$, and a transformer network $f_{\phi}(\theta^{(1)}, \ldots, \theta^{(m)}, t)$ predicts the true profile. The identity that carries the argument is the limiting KL divergence $n D_{\mathrm{KL}}(q(z|\rho)\|p(z)) \to \frac{1}{2}\beta'(t)K\|p_{\phi} - \rho\|^2$, which turns profile-based Bayesian transmission into a squared-error loss on probability vectors. Because a one-hot vector is a degenerate profile, the same loss covers single-sequence training and profile-conditioned generation.

What would settle it

Condition a trained ProfileBFN on a real family profile and on a version of the same profile whose position-wise frequency vectors are randomly permuted among positions; if the non-parametric contact-prediction precision and CLEAN function accuracy do not decline substantially for the permuted input, the model is not reading the profile information and the family-design claim fails.

Watch

Extended reading notes

Core claim

The paper's discovery is that the discrete Bayesian flow objective survives replacing a one-hot token with a whole profile, with only the mean of the accumulated Gaussian evidence changing from $K\beta(t)e_x$ to $K\beta(t)\rho$. The noisy channel becomes $q(z|\rho) = (1-\omega)/K + \omega\rho(z)$, the information state is $\theta = \mathrm{softmax}(y)$ with $y \sim \mathcal{N}(K\beta(t)\rho, \beta(t)C)$, and the limiting KL divergence yields the training loss $L(P) = \sum_i \frac{1}{2}\beta'(t)K\|P_{\phi}^{(i)} - P^{(i)}\|^2$. Since a single sequence is a degenerate profile, one network can be trained on single sequences and later conditioned on multi-sequence profiles. The authors present this unified representation as the reason ProfileBFN can balance structural conservation with diversity and novelty: on CAMEO targets its non-parametric contact-prediction precision can exceed that of the searched MSA itself; on three enzyme families its Accuracy × Uniqueness is far above PoET and EvoDiff; and fine-tuned on prediction tasks it outperforms DPLM at the same parameter scale.

Load-bearing premise

The model is trained on noise added to one-hot amino-acid identities, but at sampling time it is given noise added to fuzzy family profiles; the paper assumes the network still makes accurate predictions under this distribution shift, relying only on an empirical choice of a larger starting noise time t0=0.6.

Editorial extensions

If this is right

  • ProfileBFN makes protein family design possible without a large MSA training set: the model trains on 190 million single sequences and consumes a family profile only at generation time.
  • The generated clusters carry family structure: on the CAMEO benchmark, ProfileBFN's non-parametric contact-prediction precision (LR P@L, P@L/2, P@L/5) reaches or exceeds the searched MSA used as reference.
  • Functional enzyme design improves: ProfileBFN-Profile (650M) achieves 95.19%, 98.98%, and 42.67% Accuracy × Uniqueness on the P40925, Q7X7H9, and Q15165 families, compared with at most 88.69% for EvoDiff and 33.3% for PoET.
  • The same model can augment AlphaFold2's MSA for orphan proteins: adding ProfileBFN-generated virtual sequences raises TM-score, LDDT, and pLDDT above MSAGPT's augmentation.
  • The representation learned in pretraining transfers to downstream protein tasks, beating DPLM at equal model size across thermostability, metal-ion binding, localization, EC/GO annotation, and PPI prediction.

Reading between the lines

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

  • If the one-hot-to-profile generalization is as robust as the paper suggests, the same construction applies to any categorical generative modeling problem representable by frequency profiles, such as antibody repertoires or multi-specific binder design.
  • A directly testable extension is to make the initial sampling time t0 a function of the conditioning profile's entropy; because the paper fixes t0=0.6 for all profiles, higher-entropy families may be underexplored and lower-entropy families overexplored.
  • The non-parametric contact results suggest an iterative pipeline the paper only demonstrates once: feeding ProfileBFN-generated sequences back into profile construction could yield progressively deeper virtual MSAs and further improve structure prediction for orphan proteins.
  • Since CLEAN is itself a learned classifier, the enzyme-function results should be validated with wet-lab assays on a small set of generated candidates before the functional claim is taken as fully established.
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 / 6 minor

Summary. The paper proposes ProfileBFN, a generative model that extends Bayesian Flow Networks to protein family profiles. The central idea is to treat an MSA profile as a categorical distribution and derive a continuous-time Bayesian flow and a corresponding squared-error loss for profile inputs (Theorems 3.1 and 3.2, Appendix A.2). Because a single protein sequence is a one-hot profile, the model can be trained on single sequences and then conditioned on family profiles at generation time (Section 3.3). The paper reports improved diversity and novelty, higher CCMPred contact precision on generated clusters, better CLEAN-based enzyme functional classification, competitive representation learning results, and higher sampling efficiency than the DPLM baseline.

Significance. If the central claim holds, ProfileBFN is a significant contribution: it offers family-conditioned protein generation while avoiding the cost of constructing and training on large MSA profile datasets. The formal derivation of the profile Bayesian flow is a genuine extension of BFN, and the broad evaluation across generation, representation learning, and antibody tasks is commendable. The paper also makes a thoughtful methodological point about the fragility of parameterized structure prediction metrics and proposes a non-parametric alternative with a hacking experiment to motivate it. The significance is, however, conditional on the validity of the train/sample distribution shift discussed below, which is the main risk to the headline claim.

major comments (4)
  1. [Section 3.3, Algorithm 2, Eq. (10)] The sampling procedure uses the profile-conditioned flow with y ~ N(Kβ(t)ρ, β(t)C), whereas the network was trained exclusively on one-hot evidence y ~ N(Kβ(t)e_x, β(t)C). These two input distributions differ not only in the mean but in form: the profile-conditioned evidence is a mixture of Gaussians over the one-hot tokens, not a single Gaussian with mean Kβ(t)ρ. The manuscript does not analyze the resulting posterior mismatch, and the only mitigation is the heuristic t0=0.6. This is load-bearing for the claim that family conditioning is operational. Please add a systematic validation: measure the distance between the conditioning profile and the profile of the generated ensemble, ablate t0, and compare against a sampling scheme that explicitly draws one-hot tokens from the profile before adding noise.
  2. [Section 4.1, Table 1, Appendix D.2.2] The main structural evidence is the CCMPred long-range contact precision computed on generated clusters. Since CCMPred is itself a fitted model, the paper should include a negative control showing that the high LR P@L values are family-specific, e.g., generating from an unrelated family profile or from a shuffled profile and showing that precision drops to baseline. The comparison with the 'Searched MSA' row is also confounded by differences in cluster size, sequence diversity, and alignment depth; please report the effective number of sequences and the diversity of the MSAs used for CCMPred, and show that a random subset of the searched MSA of comparable size and diversity yields similar or lower precision.
  3. [Section 4.1, Table 2, Appendix E.1.5] The functional claim is based entirely on CLEAN's EC-number classification of the generated enzymes. The metric Accuracy × Uniqueness is useful, but the paper should report the sequence identity of the generated enzymes to the reference family and to CLEAN's training set, and provide CLEAN confidence scores for the classifications. The Q15165 accuracy of 42.67% shows that functional fidelity is not reliable for all families and should be discussed explicitly. Without these controls, the risk remains that CLEAN is responding to sequence similarity to the training distribution rather than to genuinely retained enzymatic function.
  4. [Appendix A.2, Eq. (9)] Equation (9) defines the training loss for profile targets, but training is performed exclusively with one-hot targets. The paper asserts that a one-hot is a 'degenerate profile' and that the network trained this way will generalize to arbitrary profile inputs at generation time, but this generalization is not established either theoretically or empirically. Provide a fixed-point analysis of the iterative sampling in Algorithm 2, or at least an empirical demonstration that the profile of the generated ensemble converges to the conditioning profile, not merely to a smoothed or biased version of it.
minor comments (6)
  1. [Section 2.1] The sentence 'it have been widely used...' should read 'they have been widely used...'.
  2. [Algorithm 1] The loss expression has an unbalanced parenthesis: L(P) = sum_i 1/2 β1 K ||(P^(i)_φ − P^(i)||^2. The closing parenthesis after P^(i) is missing.
  3. [Table 3] The caption does not explain the symbols †, ♡, and ⋄ used to mark quoted and reproduced baseline numbers; a full legend should be added.
  4. [Appendix E.1.5] The text refers to 'Table 5' when presenting the detailed enzyme results, but the table is labeled Table 6; correct the cross-reference.
  5. [Section 3.3] The decoding step a(i) = arg max_k(P^(i)_{φ;(N+1)})_k should specify how ties are broken (e.g., random sampling among tied amino acids).
  6. [Section 4.2, Figure 4] The 'entropy sum' used to quantify the sampling process is not defined; please provide the precise formula and how it is normalized across positions.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: ProfileBFN's loss is derived, not fitted, and its family-level claims are evaluated with external predictors.

full rationale

The central derivation is self-contained. Eq. (9) is obtained from Theorem 3.2, whose proof (Appendix A.2) computes the n->infty KL divergence n D_KL(q(z|rho)||p(z)) for the noisy channel q(z|rho)=(1-omega)/K + omega rho(z), yielding 1/2 beta'(t) K ||p_phi - rho||^2. For a one-hot rho this reduces to the original BFN loss, so training on single sequences is literally the degenerate-profile case rather than a fitted proxy. Generating from a family profile is an inference-time conditioning choice (Algorithm 2), with t0=0.6 chosen as a heuristic; the train/sample distribution shift is a real generalization risk but is not an equivalence-by-construction and is not hidden by the paper. Evaluation uses external tools (CCMPred for co-evolutionary contacts, CLEAN for EC-number classification, ESMFold/AF2 for structure), none of which is used as a training objective or fitted to the generated outputs. There is no load-bearing self-citation: the cited prior work (Graves et al., DPLM, PoET, EvoDiff) provides methods or baselines, and the only author-group citations are to external standard results. The paper's own caveats about parameterized structure metrics (Appendix D.2.2) and about SaProt's edge on structure-related tasks are limitations, not circularity. Hence no circular step reaches the threshold required by the rubric.

Assumptions & free parameters 4 free parameters · 3 assumptions · 0 invented entities

No invented physical entities. Free parameters are limited to hyperparameters chosen by hand. The main assumptions are domain assumptions about the sufficiency of profiles and transfer from one-hot to profile conditioning.

free parameters (4)
  • beta(1) = Not explicitly reported; set via beta(1)*K = constant following Graves et al. (2023).
    Controls the noise schedule; chosen by hand based on prior BFN work (Section D.1).
  • initial time t0 = 0.3 for single-sequence sampling, 0.6 for profile sampling
    Hand-selected to balance information loss vs. exploration; critical to generation results (Section 3.3).
  • beta(t) schedule = Linear schedule selected empirically from linear/square/exponential on an 8M model
    Design choice that affects generation quality; no ablation reported (Section D.1).
  • t-sampling mix = 90% per-position independent t, 10% shared t across positions
    Training strategy chosen manually; no ablation reported (Section 3.2).
assumptions (3)
  • standard math The discrete Bayesian Flow Network theory from Graves et al. (2023) applies as an underlying framework.
    The paper extends BFN; the original results (e.g., VLB decomposition, flow) are taken as background, cited in Section 2.2.
  • domain assumption A per-position MSA profile (marginal frequency vector) is a sufficient conditioning signal for generating structurally and functionally conserved family members.
    The method inputs only the profile, discarding pairwise coevolution information; the paper relies on empirical results to support this, but gives no theoretical justification.
  • domain assumption A network trained on one-hot evidence generalizes to profile evidence at sampling time.
    Training uses single sequences (degenerate profiles), while generation uses profile-conditioned noise; no analysis of this distribution shift is provided (Section 3.2).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Steering Protein Family Design through Profile Bayesian Flow." pith.science (2026). https://pith.science/paper/TEKDU7ZH

@misc{pith2026250207671,
  author       = {Pith},
  title        = {Pith review of: Steering Protein Family Design through Profile Bayesian Flow},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TEKDU7ZH}},
  note         = {Machine review of arXiv:2502.07671}
}
read the original abstract

Protein family design emerges as a promising alternative by combining the advantages of de novo protein design and mutation-based directed evolution.In this paper, we propose ProfileBFN, the Profile Bayesian Flow Networks, for specifically generative modeling of protein families. ProfileBFN extends the discrete Bayesian Flow Network from an MSA profile perspective, which can be trained on single protein sequences by regarding it as a degenerate profile, thereby achieving efficient protein family design by avoiding large-scale MSA data construction and training. Empirical results show that ProfileBFN has a profound understanding of proteins. When generating diverse and novel family proteins, it can accurately capture the structural characteristics of the family. The enzyme produced by this method is more likely than the previous approach to have the corresponding function, offering better odds of generating diverse proteins with the desired functionality.

Figures

Figures reproduced from arXiv: 2502.07671 by the authors.

Figure 1
Figure 1. Example of contact map obtained using ProfileBFN (left) and Searched MSA (right). [PITH_FULL_IMAGE:figures/full_fig_p007_1.png] view at source ↗
Figure 2
Figure 2. Three structurally conserved but sequence-novel lysozymes are generated by ProfileBFN. [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. Sampling efficiency comparison. Pro￾fileBFN has a higher sampling efficiency com￾pared to its competitors. Sampling Process Reflects Protein Conserva￾tion The sampling process of ProfileBFN is essentially a transition from a high entropy state to a low entropy state. In this paragraph, we explore the relationship between this pro￾cess and the conservation of different protein sites. Specifically, we sum the entropy … view at source ↗
Figures from the paper (9 more)
Figure 4
Figure 4. Figure 4: ProfileBFN’s sampling process implies the conservation of proteins. [PITH_FULL_IMAGE:figures/full_fig_p010_4.png]
Figure 5
Figure 5. Figure 5: Sequence novelty and predicted structural conservation of phage lysozymes generated by [PITH_FULL_IMAGE:figures/full_fig_p023_5.png]
Figure 6
Figure 6. Figure 6: Hacking ESMFold’s pLDDT by BLOSUM62 Matrix [PITH_FULL_IMAGE:figures/full_fig_p024_6.png]
Figure 7
Figure 7. Figure 7: Trivial cases of PoET generated repeated sequence with high pLDDT after ESMFold. [PITH_FULL_IMAGE:figures/full_fig_p025_7.png]
Figure 8
Figure 8. Figure 8: Visualization of improved structure prediction sample compared with AlphaFold2 and [PITH_FULL_IMAGE:figures/full_fig_p029_8.png]
Figure 9
Figure 9. Figure 9: ProfileBFN-profile Generation Result of Contact Prediction of Protein 8YXK with differ [PITH_FULL_IMAGE:figures/full_fig_p031_9.png]
Figure 10
Figure 10. Figure 10: Samples of sequences conditioned on enzyme P40925 family by model ProfileBFN, [PITH_FULL_IMAGE:figures/full_fig_p031_10.png]
Figure 11
Figure 11. Figure 11: Samples of sequences conditioned on enzyme Q7X7H9 family by model ProfileBFN, [PITH_FULL_IMAGE:figures/full_fig_p032_11.png]
Figure 12
Figure 12. Figure 12: Samples of sequences conditioned on enzyme Q15165 family by model ProfileBFN, [PITH_FULL_IMAGE:figures/full_fig_p033_12.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

62 extracted references · 47 canonical work pages

  1. [1]

    Accurate structure prediction of biomolecular interactions with alphafold 3

    Josh Abramson, Jonas Adler, Jack Dunger, Richard Evans, Tim Green, Alexander Pritzel, Olaf Ronneberger, Lindsay Willmore, Andrew J Ballard, Joshua Bambrick, et al. Accurate structure prediction of biomolecular interactions with alphafold 3. Nature, pp.\ 1--3, 2024

    work page 2024

  2. [2]

    Rosettaantibodydesign (rabd): A general framework for computational antibody design

    Jared Adolf-Bryfogle, Oleks Kalyuzhniy, Michael Kubitz, Brian D Weitzner, Xiaozhen Hu, Yumiko Adachi, William R Schief, and Roland L Dunbrack Jr. Rosettaantibodydesign (rabd): A general framework for computational antibody design. PLoS computational biology, 14 0 (4): 0 e1006112, 2018

    work page 2018

  3. [3]

    Protein generation with evolutionary diffusion: sequence is all you need

    Sarah Alamdari, Nitya Thakkar, Rianne van den Berg, Alex X Lu, Nicolo Fusi, Ava P Amini, and Kevin K Yang. Protein generation with evolutionary diffusion: sequence is all you need. BioRxiv, pp.\ 2023--09, 2023

    work page 2023

  4. [4]

    Exploring the predictive capabilities of alphafold using adversarial protein sequences

    Ismail R Alkhouri, Sumit Jha, Andre Beckus, George Atia, Susmit Jha, Rickard Ewetz, and Alvaro Velasquez. Exploring the predictive capabilities of alphafold using adversarial protein sequences. IEEE Transactions on Artificial Intelligence, 2024

    work page 2024

  5. [5]

    Design by directed evolution

    Frances H Arnold. Design by directed evolution. Accounts of chemical research, 31 0 (3): 0 125--131, 1998

    work page 1998

  6. [6]

    Synthetic biology

    Steven A Benner and A Michael Sismour. Synthetic biology. Nature reviews genetics, 6 0 (7): 0 533--543, 2005

    work page 2005

  7. [7]

    In the light of directed evolution: pathways of adaptive protein evolution

    Jesse D Bloom and Frances H Arnold. In the light of directed evolution: pathways of adaptive protein evolution. Proceedings of the National Academy of Sciences, 106 0 (supplement\_1): 0 9995--10000, 2009

    work page 2009

  8. [8]

    MSAGPT: Neural Prompting Protein Structure Prediction via MSA Generative Pre-Training

    Bo Chen, Zhilei Bei, Xingyi Cheng, Pan Li, Jie Tang, and Le Song. Msagpt: Neural prompting protein structure prediction via msa generative pre-training. arXiv preprint arXiv:2406.05347, 2024

    work page Pith review arXiv 2024

Show all 62 references
  1. [9]

    Uniprot: a hub for protein information

    UniProt Consortium. Uniprot: a hub for protein information. Nucleic acids research, 43 0 (D1): 0 D204--D212, 2015

    2015

  2. [10]

    De novo protein design: fully automated sequence selection

    Bassil I Dahiyat and Stephen L Mayo. De novo protein design: fully automated sequence selection. Science, 278 0 (5335): 0 82--87, 1997

    1997

  3. [11]

    Flip: Benchmark tasks in fitness landscape inference for proteins

    Christian Dallago, Jody Mou, Kadina E Johnston, Bruce J Wittmann, Nicholas Bhattacharya, Samuel Goldman, Ali Madani, and Kevin K Yang. Flip: Benchmark tasks in fitness landscape inference for proteins. bioRxiv, pp.\ 2021--11, 2021

    2021

  4. [12]

    Multiple sequence alignment

    Robert C Edgar and Serafim Batzoglou. Multiple sequence alignment. Current opinion in structural biology, 16 0 (3): 0 368--373, 2006

    2006

  5. [13]

    Bayesian flow networks

    Alex Graves, Rupesh Kumar Srivastava, Timothy Atkinson, and Faustino Gomez. Bayesian flow networks. arXiv preprint arXiv:2308.07037, 2023

    2023 arXiv

  6. [14]

    Protein sequence analysis

    M Michael Gromiha. Protein sequence analysis. Protein bioinformatics: from sequence to function. Elsevier Inc., New Delhi, India, pp.\ 29--62, 2010

    2010

  7. [15]

    Simulating 500 million years of evolution with a language model

    Tomas Hayes, Roshan Rao, Halil Akin, Nicholas J Sofroniew, Deniz Oktay, Zeming Lin, Robert Verkuil, Vincent Q Tran, Jonathan Deaton, Marius Wiggert, et al. Simulating 500 million years of evolution with a language model. bioRxiv, pp.\ 2024--07, 2024

    2024

  8. [16]

    Amino acid substitution matrices from protein blocks

    Steven Henikoff and Jorja G Henikoff. Amino acid substitution matrices from protein blocks. Proceedings of the National Academy of Sciences, 89 0 (22): 0 10915--10919, 1992

    1992

  9. [17]

    Precision medicine, ai, and the future of personalized health care

    Kevin B Johnson, Wei-Qi Wei, Dilhan Weeraratne, Mark E Frisse, Karl Misulis, Kyu Rhee, Juan Zhao, and Jane L Snowdon. Precision medicine, ai, and the future of personalized health care. Clinical and translational science, 14 0 (1): 0 86--93, 2021

    2021

  10. [18]

    Highly accurate protein structure prediction with alphafold

    John Jumper, Richard Evans, Alexander Pritzel, Tim Green, Michael Figurnov, Olaf Ronneberger, Kathryn Tunyasuvunakool, Russ Bates, Augustin Z \' dek, Anna Potapenko, et al. Highly accurate protein structure prediction with alphafold. nature, 596 0 (7873): 0 583--589, 2021

    2021

  11. [19]

    A comprehensive evolutionary classification of proteins encoded in complete eukaryotic genomes

    Eugene V Koonin, Natalie D Fedorova, John D Jackson, Aviva R Jacobs, Dmitri M Krylov, Kira S Makarova, Raja Mazumder, Sergei L Mekhedov, Anastasia N Nikolskaya, B Sridhar Rao, et al. A comprehensive evolutionary classification of proteins encoded in complete eukaryotic genomes...

    2004

  12. [20]

    Precision medicine

    Michael R Kosorok and Eric B Laber. Precision medicine. Annual review of statistics and its application, 6 0 (1): 0 263--286, 2019

    2019

  13. [21]

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

    Zeming Lin, Halil Akin, Roshan Rao, Brian Hie, Zhongkai Zhu, Wenting Lu, Nikita Smetanin, Allan dos Santos Costa, Maryam Fazel-Zarandi, Tom Sercu, Sal Candido, et al. Language models of protein sequences at the scale of evolution enable accurate structure prediction. bioRxiv, 2022

    2022

  14. [22]

    Evolutionary-scale prediction of atomic-level protein structure with a language model

    Zeming Lin, Halil Akin, Roshan Rao, Brian Hie, Zhongkai Zhu, Wenting Lu, Nikita Smetanin, Robert Verkuil, Ori Kabeli, Yaniv Shmueli, et al. Evolutionary-scale prediction of atomic-level protein structure with a language model. Science, 379 0 (6637): 0 1123--1130, 2023

    2023

  15. [23]

    Msa-cuda: multiple sequence alignment on graphics processing units with cuda

    Yongchao Liu, Bertil Schmidt, and Douglas L Maskell. Msa-cuda: multiple sequence alignment on graphics processing units with cuda. In 2009 20th IEEE International Conference on Application-specific Systems, Architectures and Processors, pp.\ 121--128. IEEE, 2009

    2009

  16. [24]

    Decoupled weight decay regularization

    I Loshchilov. Decoupled weight decay regularization. arXiv preprint arXiv:1711.05101, 2017

    2017 arXiv

  17. [25]

    Antigen-specific antibody design and optimization with diffusion-based generative models for protein structures

    Shitong Luo, Yufeng Su, Xingang Peng, Sheng Wang, Jian Peng, and Jianzhu Ma. Antigen-specific antibody design and optimization with diffusion-based generative models for protein structures. Advances in Neural Information Processing Systems, 35: 0 9754--9767, 2022

    2022

  18. [26]

    A unified analysis of evolutionary and population constraint in protein domains highlights structural features and pathogenic sites

    Stuart A MacGowan, F \'a bio Madeira, Thiago Britto-Borges, and Geoffrey J Barton. A unified analysis of evolutionary and population constraint in protein domains highlights structural features and pathogenic sites. Communications Biology, 7 0 (1): 0 447, 2024

    2024

  19. [27]

    Progen: Language modeling for protein generation

    Ali Madani, Bryan McCann, Nikhil Naik, Nitish Shirish Keskar, Namrata Anand, Raphael R Eguchi, Po-Ssu Huang, and Richard Socher. Progen: Language modeling for protein generation. arXiv preprint arXiv:2004.03497, 2020

    2004 arXiv

  20. [28]

    Large language models generate functional protein sequences across diverse families

    Ali Madani, Ben Krause, Eric R Greene, Subu Subramanian, Benjamin P Mohr, James M Holton, Jose Luis Olmos, Caiming Xiong, Zachary Z Sun, Richard Socher, et al. Large language models generate functional protein sequences across diverse families. Nature Biotechnology, 41 0 (8): ...

    2023

  21. [29]

    Language models enable zero-shot prediction of the effects of mutations on protein function

    Joshua Meier, Roshan Rao, Robert Verkuil, Jason Liu, Tom Sercu, and Alexander Rives. Language models enable zero-shot prediction of the effects of mutations on protein function. bioRxiv, 2021. doi:10.1101/2021.07.09.450648. URL https://www.biorxiv.org/content/10.1101/2021.07.0...

    2021 doi

  22. [30]

    Improved structure-related prediction for insufficient homologous proteins using msa enhancement and pre-trained language model

    Qiaozhen Meng, Fei Guo, and Jijun Tang. Improved structure-related prediction for insufficient homologous proteins using msa enhancement and pre-trained language model. Briefings in Bioinformatics, 24 0 (4): 0 bbad217, 2023

    2023

  23. [31]

    Uniclust databases of clustered and deeply annotated protein sequences and alignments

    Milot Mirdita, Lars Von Den Driesch, Clovis Galiez, Maria J Martin, Johannes S \"o ding, and Martin Steinegger. Uniclust databases of clustered and deeply annotated protein sequences and alignments. Nucleic acids research, 45 0 (D1): 0 D170--D176, 2017

    2017

  24. [32]

    A heuristic approach to high-speed multiple sequence alignment for phylogenetic tree construction

    Akash Nag and Sunil Karforma. A heuristic approach to high-speed multiple sequence alignment for phylogenetic tree construction. IPASJ International Journal of Computer Science, 4 0 (4): 0 10--15, 2016

    2016

  25. [33]

    Progen2: exploring the boundaries of protein language models

    Erik Nijkamp, Jeffrey A Ruffolo, Eli N Weinstein, Nikhil Naik, and Ali Madani. Progen2: exploring the boundaries of protein language models. Cell systems, 14 0 (11): 0 968--978, 2023

    2023

  26. [34]

    Ablang: an antibody language model for completing antibody sequences

    Tobias H Olsen, Iain H Moal, and Charlotte M Deane. Ablang: an antibody language model for completing antibody sequences. Bioinformatics Advances, 2 0 (1): 0 vbac046, 2022

    2022

  27. [35]

    Methods for the directed evolution of proteins

    Michael S Packer and David R Liu. Methods for the directed evolution of proteins. Nature Reviews Genetics, 16 0 (7): 0 379--394, 2015

    2015

  28. [36]

    Msa transformer

    Roshan M Rao, Jason Liu, Robert Verkuil, Joshua Meier, John Canny, Pieter Abbeel, Tom Sercu, and Alexander Rives. Msa transformer. In International Conference on Machine Learning, pp.\ 8844--8856. PMLR, 2021

    2021

  29. [37]

    Hhblits: lightning-fast iterative protein sequence searching by hmm-hmm alignment

    Michael Remmert, Andreas Biegert, Andreas Hauser, and Johannes S \"o ding. Hhblits: lightning-fast iterative protein sequence searching by hmm-hmm alignment. Nature methods, 9 0 (2): 0 173--175, 2012

    2012

  30. [38]

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

    Alexander Rives, Joshua Meier, Tom Sercu, Siddharth Goyal, Zeming Lin, Jason Liu, Demi Guo, Myle Ott, C Lawrence Zitnick, Jerry Ma, et al. Biological structure and function emerge from scaling unsupervised learning to 250 million protein sequences. Proceedings of the National ...

    2021

  31. [39]

    Continuous automated model evaluation (cameo)—perspectives on the future of fully automated evaluation of structure prediction methods

    Xavier Robin, Juergen Haas, Rafal Gumienny, Anna Smolinski, Gerardo Tauriello, and Torsten Schwede. Continuous automated model evaluation (cameo)—perspectives on the future of fully automated evaluation of structure prediction methods. Proteins: Structure, Function, and Bioinf...

    1977

  32. [40]

    Deciphering antibody affinity maturation with language models and weakly supervised learning

    Jeffrey A Ruffolo, Jeffrey J Gray, and Jeremias Sulam. Deciphering antibody affinity maturation with language models and weakly supervised learning. arXiv preprint arXiv:2112.07782, 2021

    2021 arXiv

  33. [41]

    Ccmpred—fast and precise prediction of protein residue--residue contacts from correlated mutations

    Stefan Seemayer, Markus Gruber, and Johannes S \"o ding. Ccmpred—fast and precise prediction of protein residue--residue contacts from correlated mutations. Bioinformatics, 30 0 (21): 0 3128--3130, 2014

    2014

  34. [42]

    Protmamba: a homology-aware but alignment-free protein state space model

    Damiano Sgarbossa, Cyril Malbranke, and Anne-Florence Bitbol. Protmamba: a homology-aware but alignment-free protein state space model. bioRxiv, pp.\ 2024--05, 2024

    2024

  35. [43]

    Generative enzyme design guided by functionally important sites and small-molecule substrates

    Zhenqiao Song, Yunlong Zhao, Wenxian Shi, Wengong Jin, Yang Yang, and Lei Li. Generative enzyme design guided by functionally important sites and small-molecule substrates. arXiv preprint arXiv:2405.08205, 2024

    2024

  36. [44]

    Mmseqs2 enables sensitive protein sequence searching for the analysis of massive data sets

    Martin Steinegger and Johannes S \"o ding. Mmseqs2 enables sensitive protein sequence searching for the analysis of massive data sets. Nature biotechnology, 35 0 (11): 0 1026--1028, 2017

    2017

  37. [45]

    Saprot: Protein language modeling with structure-aware vocabulary

    Jin Su, Chenchen Han, Yuyang Zhou, Junjie Shan, Xibin Zhou, and Fajie Yuan. Saprot: Protein language modeling with structure-aware vocabulary. bioRxiv, pp.\ 2023--10, 2023

    2023

  38. [46]

    Uniref: comprehensive and non-redundant uniprot reference clusters

    Baris E Suzek, Hongzhan Huang, Peter McGarvey, Raja Mazumder, and Cathy H Wu. Uniref: comprehensive and non-redundant uniprot reference clusters. Bioinformatics, 23 0 (10): 0 1282--1288, 2007

    2007

  39. [47]

    Poet: A generative model of protein families as sequences-of-sequences

    Timothy Truong Jr and Tristan Bepler. Poet: A generative model of protein families as sequences-of-sequences. Advances in Neural Information Processing Systems, 36: 0 77379--77415, 2023

    2023

  40. [48]

    Attention is all you need

    A Vaswani. Attention is all you need. Advances in Neural Information Processing Systems, 2017

    2017

  41. [49]

    Diffusion language models are versatile protein learners

    Xinyou Wang, Zaixiang Zheng, Fei Ye, Dongyu Xue, Shujian Huang, and Quanquan Gu. Diffusion language models are versatile protein learners. arXiv preprint arXiv:2402.18567, 2024

    2024 arXiv

  42. [50]

    Directed evolution: methodologies and applications

    Yajie Wang, Pu Xue, Mingfeng Cao, Tianhao Yu, Stephan T Lane, and Huimin Zhao. Directed evolution: methodologies and applications. Chemical reviews, 121 0 (20): 0 12384--12444, 2021

    2021

  43. [51]

    De novo design of protein structure and function with rfdiffusion

    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, et al. De novo design of protein structure and function with rfdiffusion. Nature, 620 0 (7976): 0 1089--1100, 2023

    2023

  44. [52]

    High-resolution de novo structure prediction from primary sequence

    Ruidong Wu, Fan Ding, Rui Wang, Rui Shen, Xiwen Zhang, Shitong Luo, Chenpeng Su, Zuofan Wu, Qi Xie, Bonnie Berger, et al. High-resolution de novo structure prediction from primary sequence. BioRxiv, pp.\ 2022--07, 2022

    2022

  45. [53]

    Masked inverse folding with sequence transfer for protein representation learning

    Kevin K Yang, Niccol \`o Zanichelli, and Hugh Yeh. Masked inverse folding with sequence transfer for protein representation learning. Protein Engineering, Design and Selection, 36: 0 gzad015, 2023

    2023

  46. [54]

    Xlnet: Generalized autoregressive pretraining for language understanding

    Zhilin Yang. Xlnet: Generalized autoregressive pretraining for language understanding. arXiv preprint arXiv:1906.08237, 2019

    1906 arXiv

  47. [55]

    Enzyme function prediction using contrastive learning

    Tianhao Yu, Haiyang Cui, Jianan Canal Li, Yunan Luo, Guangde Jiang, and Huimin Zhao. Enzyme function prediction using contrastive learning. Science, 379 0 (6639): 0 1358--1363, 2023

    2023

  48. [56]

    Spearman rank correlation

    Jerrold H Zar. Spearman rank correlation. Encyclopedia of biostatistics, 7, 2005

    2005

  49. [57]

    Unsupervisedly prompting alphafold2 for accurate few-shot protein structure prediction

    Jun Zhang, Sirui Liu, Mengyun Chen, Haotian Chu, Min Wang, Zidong Wang, Jialiang Yu, Ningxi Ni, Fan Yu, Dechin Chen, et al. Unsupervisedly prompting alphafold2 for accurate few-shot protein structure prediction. Journal of Chemical Theory and Computation, 19 0 (22): 0 8460--84...

    2023

  50. [58]

    Enhancing the protein tertiary structure prediction by multiple sequence alignment generation

    Le Zhang, Jiayang Chen, Tao Shen, Yu Li, and Siqi Sun. Enhancing the protein tertiary structure prediction by multiple sequence alignment generation. arXiv preprint arXiv:2306.01824, 2023 b

    2023 arXiv

  51. [59]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

  52. [60]

    @esa (Ref

    \@ifxundefined[1] #1\@undefined \@firstoftwo \@secondoftwo \@ifnum[1] #1 \@firstoftwo \@secondoftwo \@ifx[1] #1 \@firstoftwo \@secondoftwo [2] @ #1 \@temptokena #2 #1 @ \@temptokena \@ifclassloaded agu2001 natbib The agu2001 class already includes natbib coding, so you should ...

  53. [61]

    \@lbibitem[] @bibitem@first@sw\@secondoftwo \@lbibitem[#1]#2 \@extra@b@citeb \@ifundefined br@#2\@extra@b@citeb \@namedef br@#2 \@nameuse br@#2\@extra@b@citeb \@ifundefined b@#2\@extra@b@citeb @num @parse #2 @tmp #1 NAT@b@open@#2 NAT@b@shut@#2 \@ifnum @merge>\@ne @bibitem@firs...

  54. [62]

    - Placeholder or duplicate abstracts will be removed

    @open @close @open @close and [1] URL: #1 \@ifundefined chapter * \@mkboth \@ifxundefined @sectionbib * \@mkboth * \@mkboth\@gobbletwo \@ifclassloaded amsart * \@ifclassloaded amsbook * \@ifxundefined @heading @heading NAT@ctr thebibliography [1] @ \@biblabel @NAT@ctr \@bibset...

    2025

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.