Pith. sign in

REVIEW 2 major objections 5 minor 185 references

Geometric deep learning assists protein engineering. Opportunities and Challenges

T0 review · 2 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read Geometric deep learning is emerging as the computational backbone of protein engineering.

desk verdict A useful but unvalidated perspective that surveys geometric deep learning for protein engineering well, yet proposes an XAI-driven pipeline whose own Section 5.3 undermines the key premise. read the letter →

arxiv 2506.16091 v1 pith:FLLKQZ4P submitted 2025-06-19 q-bio.QM

Julián García-Vinuesa , Jorge Rojas , Nicole Soto-García , Nicolás Martínez , Diego Alvarez-Saravia , Roberto Uribe-Paredes , Mehdi D. Davari , Carlos Conca
show 2 more authors
Juan A. Asenjo David Medina-Ortiz
This is my paper
classification q-bio.QM
keywords geometricdeeplearningproteinengineeringgraphneuralnetworksexplainableAIstabilitypredictionprotein–proteininteractionsdenovodesignstructure-based
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This perspective argues that geometric deep learning (GDL)—machine learning that works directly on protein structures encoded as graphs or surfaces—is becoming the computational backbone of protein engineering. The authors claim that GDL captures the spatial, topological, and physicochemical features that sequence-only models miss, improving prediction of stability, solubility, function, and molecular interactions while enabling de novo design. The paper's central proposal is a unified pipeline that couples GDL predictors with explainable-AI attribution and structure-based validation, converting black-box predictions into prioritized mutational candidates and reducing the number of experimental screening rounds. Because this is a perspective, the argument is programmatic: it assembles evidence from more than ninety published GDL models rather than demonstrating a new experimental result.

What carries the argument

The central object is the protein structure graph, in which nodes are residues or atoms and edges encode covalent bonds, hydrogen bonds, salt bridges, or spatial proximity below a distance cutoff, enriched with sequence and physicochemical features. The learning machinery is message-passing graph neural networks, particularly graph convolutional and graph attention architectures, whose inductive biases let a model aggregate local structural context. The paper's integrative machinery is a rule-based prioritization layer that combines gradient-based attribution, perturbation- and subgraph-based explanations, counterfactual examples, and attention weights with uncertainty-aware probabilistic predictions and structure-based validation filters, converting model outputs into ranked mutational candidates.

What would settle it

A controlled experiment on a standardized mutational library would settle the claim: use the same GDL fitness model to propose mutations under two regimes—ranked purely by predicted fitness versus ranked by explainability-derived hotspot scores—and count how many experimentally tested variants are needed to reach a fixed improvement in stability or activity. If explainability-guided prioritization does not beat black-box ranking, the pipeline's central promise of fewer experimental rounds is not supported.

Watch

Extended reading notes

Core claim

The paper's central claim is that GDL is not merely an incremental improvement over sequence-based deep learning but a change in how proteins can be engineered: by learning on graphs, surfaces, or volumetric grids of molecular structure, GDL models can represent allosteric couplings, binding interfaces, and conformational features that linear sequences hide. The survey of more than ninety models finds that graph-based representations dominate, with graph convolutional and graph attention networks the most common architectures and protein–protein interaction prediction the most studied task. The proposed innovation is a closed design loop: train a GDL fitness model on mutational libraries or high-throughput assays; extract hypotheses from attention weights, attribution maps, and counterfactual graph perturbations; aggregate those signals into probabilistic hotspot scores; and filter candidates through physics-based relaxation and structural validation before wet-lab testing. The authors argue that this loop can support transparent, autonomous protein design while making model predictions biologically interpretable.

Load-bearing premise

The whole argument depends on the assumption that the residues singled out by attention weights, attribution scores, and counterfactual perturbations are the residues that causally determine protein fitness; the paper itself cautions that these signals can be unstable across training runs and can miss true causal determinants.

Editorial extensions

If this is right

  • GDL-based fitness models can prioritize mutations by predicted stability or activity, so experimental screens can be smaller and directed-evolution cycles shorter.
  • Predicted structures can substitute for experimental structures in the pipeline, extending geometric modeling to proteins that lack crystal or cryo-EM data.
  • Attention, attribution, and counterfactual explanations can nominate specific residues as mutational hotspots, turning black-box predictions into testable mechanistic hypotheses.
  • Hybrid generative GDL models can produce novel sequences that fit a designed backbone, making de novo protein design a practical computational task.
  • Homology-aware benchmarking and uncertainty quantification become necessary conditions for trusting GDL predictions across protein families.

Reading between the lines

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

  • The decisive test the paper leaves implicit is comparative: run the same GDL fitness model twice on the same mutational library, once with XAI-based hotspot selection and once with plain black-box ranking, and count how many wet-lab experiments are needed to reach a fixed improvement; the pipeline's added value is that number.
  • Because the authors themselves note that attention weights vary across training runs and that attribution can miss causal determinants, users should require explanation consistency across random seeds before treating any nominated residue as a design target.
  • GDL's comparative advantage should appear first in tasks where geometry matters most—allosteric regulation, conformational flexibility, cryptic pockets, and interface contacts—so benchmarks that isolate those regimes are the most informative way to test the claim.
  • As high-throughput assays and laboratory automation mature, the same loop could close automatically, but structural prediction error and explanation instability will likely keep a human validator in the loop for at least the next generation of tools.
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

2 major / 5 minor

Summary. This perspective surveys geometric deep learning (GDL) for protein engineering, covering structural representations, graph construction, GNN architectures, and applications in functional annotation, physicochemical property prediction, intermolecular interactions, and de novo design. It then proposes a unified pipeline that integrates GDL with explainable AI (XAI) and structure-based validation to support transparent, autonomous protein design. The paper does not introduce new models or experiments; it is a literature synthesis with a schematic framework and a census of over 90 published models (Figure 3).

Significance. If taken as a perspective, the paper offers a broad and current literature overview and a plausible roadmap for combining geometric modeling with interpretability. Its strongest value is the synthesis of many recent methods and the explicit acknowledgement of key limitations, particularly the instability of attention-based and attribution-based explanations in Section 5.3. The proposed pipeline is timely but remains at the conceptual stage; the census in Figure 3, if made reproducible, could be a useful community resource. The paper is honest about open challenges and does not overclaim empirical results, which is appropriate for its format.

major comments (2)
  1. [Section 6 vs. Section 5.3] The central proposal of Section 6 is that outputs from attribution, attention, and counterfactual layers can be aggregated into a rule-based module that assigns 'probabilistic relevance scores' and prioritizes mutational hotspots for experimental validation. This premise is directly contradicted by the manuscript's own Section 5.3, which states that attention weights 'can be unreliable, varying across training runs and random seeds' and that attribution methods 'often fail to capture the true causal determinants of molecular function and can produce unstable outputs.' The paper provides no experiment, case study, or simulation that tests whether these XAI outputs actually identify residues that are causal for fitness or whether the pipeline improves over black-box ranking. Figure 4 is explicitly a schematic. Since the pipeline is the paper's main constructive contribution, the claim that it enables 'transparent, autonomous design' is currently unsupported at the load-bearing point. I recommend either adding a proof-of-concept demonstration on a public fitness landscape (e.g., ProteinGym) or repositioning the pipeline as a research agenda with the causal relevance of XAI outputs listed as an open challenge.
  2. [Figure 3 and Section 4 introduction] The text states that Figure 3 presents 'a systematic analysis of more than 90 GDL-based models reported in the literature,' but no search strategy, inclusion/exclusion criteria, model list, or data/code availability are provided. The reader cannot verify the reported counts or proportions (e.g., 73% graph-based representations, GCN/GAT dominance, the ranking of protein–protein interaction as the most common task). Because Section 4 organizes its task-oriented overview around this census and the claims are quantitative, this lack of methodological transparency undermines a central piece of evidence. I recommend adding a supplementary table listing all included models with their representations, architectures, and tasks, plus a methods paragraph describing the literature search and curation process.
minor comments (5)
  1. [Section 1] There is a grammar error in the sentence 'GDL can captures the complex geometry of protein structures' (the modal 'can' should be followed by the bare infinitive 'capture').
  2. [Section 3.4.3] The text reads 'The performance of these models are typically evaluated' (subject–verb agreement); it should be 'is typically evaluated.'
  3. [Section 3.4.3] 'Frechet distance' should be written as 'Fréchet distance' with the accent.
  4. [Section 5] The word 'oportunities' in the opening sentence should be 'opportunities.'
  5. [Section 2] The paragraph on transfer learning asserts that curated benchmarks such as TAPE, SABDab, and ProteinGym 'have enhanced comparability and reproducibility,' but the paper does not discuss how these benchmarks specifically relate to geometric deep learning, as opposed to sequence-based models; a brief clarification would help.

Circularity Check

0 steps flagged · score 2.0 of 10

No circular derivation; the perspective assembles existing GDL/XAI methods, and its self-citations to Medina-Ortiz et al. 2024b are not load-bearing.

full rationale

This is a perspective/review, not a derivation. It explicitly states that rather than introducing a new model architecture, it aims to consolidate existing strategies into a cohesive pipeline, and no equation is introduced, no parameter is fitted, and no prediction is compared with a target. The Section 6 pipeline is a schematic proposal (Figure 4), and its components (GDL training, permutation importance, counterfactual perturbations, attention weights, subgraph discovery, rule-based aggregation, physics-based filtering) are described in the text rather than imported as derived results. The repeated citations to Medina-Ortiz et al. 2024b, an overlapping-author interpretability review, support motivational claims about transparency and interpretability, but the central claim that GDL is emerging as a foundational paradigm and that a GDL+XAI+validation pipeline could support transparent design does not reduce to those citations: the pipeline description is self-contained, and external citations (Nandan et al. 2025, Hao et al. 2021, Montavon et al. 2018, etc.) carry the methodological content. The most serious weakness is an internal-premise tension, not circularity: Section 5.3 says attention weights can be unreliable and attribution methods often fail to capture true causal determinants, while Section 6 aggregates these same outputs into probabilistic relevance scores for hotspot prioritization. That is an unsupported causal assumption about XAI outputs, but the paper never defines the relevance scores in terms of measured fitness or fits them to the design target, so the proposal is not circular by construction. Given the minor, non-load-bearing self-citations, the score is 2.

Assumptions & free parameters 0 free parameters · 2 assumptions · 0 invented entities

This perspective makes no fitted parameters and posits no new entities. Its proposed pipeline rests on two domain assumptions: that graph representations based on predicted structures preserve task-relevant protein information, and that explainability outputs can identify causal determinants. Both are acknowledged in the text as open challenges.

assumptions (2)
  • domain assumption Graph representations built from predicted structures preserve enough spatial, topological, and physicochemical information for predictive and generative protein engineering tasks.
    Sections 3.1 to 3.4 build the pipeline on structures from X-ray, NMR, cryo-EM, AlphaFold, RoseTTAFold, or ESMFold. Section 4.2 concedes that mutant structure accuracy is variable and may constrain applicability.
  • domain assumption Post hoc explanation and attention outputs identify residues that are causal for protein fitness, not merely statistically correlated.
    Sections 5.3 and 6 place XAI at the center of the proposed pipeline. The text itself notes attention weights can be unreliable across training runs and seeds, and attribution methods may fail to capture true causal determinants.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Geometric deep learning assists protein engineering. Opportunities and Challenges." pith.science (2026). https://pith.science/paper/FLLKQZ4P

@misc{pith2026250616091,
  author       = {Pith},
  title        = {Pith review of: Geometric deep learning assists protein engineering. Opportunities and Challenges},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FLLKQZ4P}},
  note         = {Machine review of arXiv:2506.16091}
}
read the original abstract

Protein engineering is experiencing a paradigmatic shift through the integration of geometric deep learning into computational design workflows. While traditional strategies, such as rational design and directed evolution, have enabled relevant advances, they remain limited by the complexity of sequence space and the cost of experimental validation. Geometric deep learning addresses these limitations by operating on non-Euclidean domains, capturing spatial, topological, and physicochemical features essential to protein function. This perspective outlines the current applications of GDL across stability prediction, functional annotation, molecular interaction modeling, and de novo protein design. We highlight recent methodological advances in model generalization, interpretability, and robustness, particularly under data-scarce conditions. A unified framework is proposed that integrates GDL with explainable AI and structure-based validation to support transparent, autonomous design. As GDL converges with generative modeling and high-throughput experimentation, it is emerging as a central technology in next-generation protein engineering and synthetic biology.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

185 extracted references · 67 canonical work pages

  1. [1]

    Abdin, O., Nim, S., Wen, H., and Kim, P. M. (2022). Pepnn: a deep attention model for the identification of peptide binding sites. Communications biology , 5(1):503

    2022

  2. [2]

    J., Bambrick, J., et al

    Abramson, J., Adler, J., Dunger, J., Evans, R., Green, T., Pritzel, A., Ronneberger, O., Willmore, L., Ballard, A. J., Bambrick, J., et al. (2024). Accurate structure prediction of biomolecular interactions with alphafold 3. Nature , 630(8016):493--500

    2024

  3. [3]

    and Das, R

    Ahmetoglu, H. and Das, R. (2022). A comprehensive review on detection of cyber-attacks: Data sets, methods, challenges, and future research directions. Internet of Things , 20:100615

    2022

  4. [4]

    Akansha, S. (2025). Over-squashing in graph neural networks: A comprehensive survey. Neurocomputing , page 130389

    2025

  5. [5]

    Albu, A.-I., Bocicor, M.-I., and Czibula, G. (2023). Mm-stackens: A new deep multimodal stacked generalization approach for protein--protein interaction prediction. Computers in Biology and Medicine , 153:106526

    2023

  6. [6]

    Atz, K., Grisoni, F., and Schneider, G. (2021a). Geometric deep learning on molecular representations. arXiv preprint arXiv:2107.12375

    arXiv 2021

  7. [7]

    Atz, K., Grisoni, F., and Schneider, G. (2021b). Geometric deep learning on molecular representations. Nature Machine Intelligence 2021 3:12 , 3:1023--1032

    2021

  8. [8]

    Audagnotto, M., Czechtizky, W., De Maria, L., K \"a ck, H., Papoian, G., Tornberg, L., Tyrchan, C., and Ulander, J. (2022). Machine learning/molecular dynamic protein structure prediction approach to investigate the protein conformational ensemble. Scientific Reports , 12(1):10018

    2022

Show all 185 references
  1. [9]

    Avery, C., Patterson, J., Grear, T., Frater, T., and Jacobs, D. J. (2022). Protein function analysis through machine learning. Biomolecules , 12(9):1246

    2022

  2. [10]

    R., Wang, J., Cong, Q., Kinch, L

    Baek, M., DiMaio, F., Anishchenko, I., Dauparas, J., Ovchinnikov, S., Lee, G. R., Wang, J., Cong, Q., Kinch, L. N., Schaeffer, R. D., et al. (2021). Accurate prediction of protein structures and interactions using a three-track neural network. Science , 373(6557):871--876

    2021

  3. [11]

    Bajaj, S., Son, H., Liu, J., Guan, H., and Serafini, M. (2024). Graph neural network training systems: A performance comparison of full-graph and mini-batch. arXiv preprint arXiv:2406.00552

    2024 arXiv

  4. [12]

    Balaji, V., Kozyrakis, C., Chechik, G., and Maron, I. H. (2025). Efficient gnn training through structure-aware randomized mini-batching. arXiv preprint arXiv:2504.18082

    2025 arXiv

  5. [13]

    S., Elvati, P., Kozarekar, S., VanEpps, J

    Baranwal, M., Magner, A., Saldinger, J., Turali-Emre, E. S., Elvati, P., Kozarekar, S., VanEpps, J. S., Kotov, N. A., Violi, A., and Hero, A. O. (2022). Struct2graph: a graph attention network for structure based predictions of protein--protein interactions. BMC bioinformatics...

    2022

  6. [14]

    A., Olivares-Gil, A., Rodr \' guez, J

    Barbero-Aparicio, J. A., Olivares-Gil, A., Rodr \' guez, J. J., Garc \' a-Osorio, C., and D \' ez-Pastor, J. F. (2024a). Addressing data scarcity in protein fitness landscape analysis: A study on semi-supervised and deep transfer learning techniques. Information Fusion , 102:102035

    2024

  7. [15]

    A., Olivares-Gil, A., Rodríguez, J

    Barbero-Aparicio, J. A., Olivares-Gil, A., Rodríguez, J. J., García-Osorio, C., and Díez-Pastor, J. F. (2024b). Addressing data scarcity in protein fitness landscape analysis: A study on semi-supervised and deep transfer learning techniques. Information Fusion , 102:102035

    2024

  8. [16]

    W., Hamrick, J

    Battaglia, P. W., Hamrick, J. B., Bapst, V., Sanchez-Gonzalez, A., Zambaldi, V., Malinowski, M., Tacchetti, A., Raposo, D., Santoro, A., Faulkner, R., et al. (2018). Relational inductive biases, deep learning, and graph networks. arXiv preprint arXiv:1806.01261

    2018 arXiv

  9. [17]

    and Wang, B

    Behnam, A. and Wang, B. (2024). Graph neural network causal explanation via neural causal models. In European Conference on Computer Vision , pages 410--427. Springer

    2024

  10. [18]

    M., Lima, A

    Bertoline, L. M., Lima, A. N., Krieger, J. E., and Teixeira, S. K. (2023). Before and after alphafold2: An overview of protein structure prediction. Frontiers in bioinformatics , 3:1120370

    2023

  11. [19]

    Bischl, B., Binder, M., Lang, M., Pielok, T., Richter, J., Coors, S., Thomas, J., Ullmann, T., Becker, M., Boulesteix, A.-L., et al. (2023). Hyperparameter optimization: Foundations, algorithms, best practices, and open challenges. Wiley Interdisciplinary Reviews: Data Mining ...

    2023

  12. [20]

    C., Esvelt, K

    Biswas, S., Khimulya, G., Alley, E. C., Esvelt, K. M., and Church, G. M. (2021). Low-n protein engineering with data-efficient deep learning. Nature methods , 18(4):389--396

    2021

  13. [21]

    M., Bruna, J., Cohen, T., and Veličković, P

    Bronstein, M. M., Bruna, J., Cohen, T., and Veličković, P. (2021). Geometric deep learning: Grids, groups, graphs, geodesics, and gauges. CoRR

    2021

  14. [22]

    Cao, C., Wang, H., Yang, J.-R., Chen, Q., Guo, Y.-M., and Chen, J.-Z. (2024). Mcpnet: Development of an interpretable deep learning model based on multiple conformations of the compound for predicting developmental toxicity. Computers in Biology and Medicine , 171:108037

    2024

  15. [23]

    Cao, W., Yan, Z., He, Z., and He, Z. (2020). A comprehensive survey on geometric deep learning. IEEE Access , 8:35929--35949

    2020

  16. [24]

    Cao, W., Zheng, C., Yan, Z., and Xie, W. (2022). Geometric deep learning: progress, applications and challenges. Science China. Information Sciences , 65(2):126101

    2022

  17. [25]

    Chen, N., Yu, J., Zhe, L., Wang, F., Li, X., and Wong, K.-C. (2024). Tp-lmmsg: a peptide prediction graph neural network incorporating flexible amino acid property representation. Briefings in Bioinformatics , 25(4):bbae308

    2024

  18. [26]

    R., He, X., and Fushimi, R

    Chen, Y.-Y., Kunz, M. R., He, X., and Fushimi, R. (2022). Recent progress toward catalyst properties, performance, and prediction with data-driven methods. Current Opinion in Chemical Engineering , 37:100843

    2022

  19. [27]

    Cheng, P., Mao, C., Tang, J., Yang, S., Cheng, Y., Wang, W., Gu, Q., Han, W., Chen, H., Li, S., et al. (2024). Zero-shot prediction of mutation effects with multimodal deep representation learning guides protein engineering. Cell Research , 34(9):630--647

    2024

  20. [28]

    and Bailey-Kellogg, C

    Dai, B. and Bailey-Kellogg, C. (2021). Protein interaction interface region prediction by geometric deep learning. Bioinformatics , 37(17):2580--2588

    2021

  21. [29]

    Dai, W. (2025). A survey of deep learning methods in protein bioinformatics and its impact on protein design

    2025

  22. [30]

    J., Milles, L

    Dauparas, J., Anishchenko, I., Bennett, N., Bai, H., Ragotte, R. J., Milles, L. F., Wicky, B. I. M., Courbet, A., de Haas, R. J., Bethel, N., Leung, P. J. Y., Huddy, T. F., Pellock, S., Tischer, D., Chan, F., Koepnick, B., Nguyen, H., Kang, A., Sankaran, B., Bera, A. K., King,...

    2022

  23. [31]

    L., Radojewski, P., and McKinley, R

    Diaz, I., Scherer, F., Berli, Y., Wiest, R., Hammer, H., Hoepner, R., Betancourt, A. L., Radojewski, P., and McKinley, R. (2025). Learning from spatially inhomogenous data: resolution-adaptive convolutions for multiple sclerosis lesion segmentation. arXiv preprint arXiv:2503.21829

    2025 arXiv

  24. [32]

    Ding, H., Li, X., Han, P., Tian, X., Jing, F., Wang, S., Song, T., Fu, H., and Kang, N. (2024). Meg-ppis: a fast protein--protein interaction site prediction method based on multi-scale graph information and equivariant graph neural network. Bioinformatics , 40(5):btae269

    2024

  25. [33]

    O., Dirks, R

    Dror, R. O., Dirks, R. M., Grossman, J., Xu, H., and Shaw, D. E. (2012). Biomolecular simulation: a computational microscope for molecular biology. Annual review of biophysics , 41(1):429--452

    2012

  26. [34]

    and Malliaros, F

    Duval, A. and Malliaros, F. D. (2021). Graphsvx: Shapley value explanations for graph neural networks. In Machine Learning and Knowledge Discovery in Databases. Research Track: European Conference, ECML PKDD 2021, Bilbao, Spain, September 13--17, 2021, Proceedings, Part II 21 ...

    2021

  27. [35]

    P., Joshi, C

    Dwivedi, V. P., Joshi, C. K., Luu, A. T., Laurent, T., Bengio, Y., and Bresson, X. (2023). Benchmarking graph neural networks. Journal of Machine Learning Research , 24(43):1--48

    2023

  28. [36]

    L., and Birol, I

    Ebrahimikondori, H., Sutherland, D., Yanai, A., Richter, A., Salehi, A., Li, C., Coombe, L., Kotkoff, M., Warren, R. L., and Birol, I. (2024). Structure-aware deep learning model for peptide toxicity prediction. Protein Science , 33(7):e5076

    2024

  29. [37]

    Fang, Y., Jiang, Y., Wei, L., Ma, Q., Ren, Z., Yuan, Q., and Wei, D.-Q. (2023). Deepprosite: structure-aware protein binding site prediction using esmfold and pretrained language model. Bioinformatics , 39(12):btad718

    2023

  30. [38]

    M., and Correia, B

    Gainza, P., Sverrisson, F., Monti, F., Rodolà, E., Boscaini, D., Bronstein, M. M., and Correia, B. E. (2020). Deciphering interaction fingerprints from protein molecular surfaces using geometric deep learning. Nature Methods , 17:184--192

    2020

  31. [39]

    Gao, Y., Shi, R., Yu, G., Huang, Y., and Yang, Y. (2025). Zerpi: A graph neural network model for zero-shot prediction of rna-protein interactions. Methods

    2025

  32. [40]

    Gao, Z., Tan, C., and Li, S. Z. (2022). Alphadesign: A graph protein design method and benchmark on alphafolddb

    2022

  33. [41]

    Ge, F., Zhou, J., Zhang, M., and Yu, D.-J. (2025). Mfp-mfl: Leveraging graph attention and multi-feature integration for superior multifunctional bioactive peptide prediction. International Journal of Molecular Sciences , 26(3):1317

    2025

  34. [42]

    D., Kosciolek, T., Leman, J

    Gligorijevi \'c , V., Renfrew, P. D., Kosciolek, T., Leman, J. K., Berenberg, D., Vatanen, T., Chandler, C., Taylor, B. C., Fisk, I. M., Vlamakis, H., et al. (2021). Structure-based protein function prediction using graph convolutional networks. Nature communications , 12(1):3168

    2021

  35. [43]

    Gong, H., Zhang, Y., Dong, C., Wang, Y., Chen, G., Liang, B., Li, H., Liu, L., Xu, J., and Li, G. (2023). Unbiased curriculum learning enhanced global-local graph neural network for protein thermodynamic stability prediction. Bioinformatics , 39

    2023

  36. [44]

    G., Hotzel, I., Lafrance-Vanasse, J., Rajpal, A., Cho, K., and Wilson, A

    Gruver, N., Stanton, S., Frey, N., Rudner, T. G., Hotzel, I., Lafrance-Vanasse, J., Rajpal, A., Cho, K., and Wilson, A. G. (2023). Protein design with guided discrete diffusion. Advances in neural information processing systems , 36:12489--12517

    2023

  37. [45]

    Guntuboina, C., Das, A., Mollaei, P., Kim, S., and Barati Farimani, A. (2023). Peptidebert: A language model based on transformers for peptide property prediction. The Journal of Physical Chemistry Letters , 14(46):10427--10434

    2023

  38. [46]

    Hao, Y., Dong, L., Wei, F., and Xu, K. (2021). Self-attention attribution: Interpreting information interactions inside transformer. In Proceedings of the AAAI Conference on Artificial Intelligence , volume 35, pages 12963--12971

    2021

  39. [47]

    G., Gharabli, H., Acevedo-Rocha, C

    Harding-Larsen, D., Funk, J., Madsen, N. G., Gharabli, H., Acevedo-Rocha, C. G., Mazurenko, S., and Welner, D. H. (2024). Protein representations: Encoding biological information for machine learning in biocatalysis. Biotechnology Advances , 77:108459

    2024

  40. [48]

    Hawkins-Hooker, A., Depardieu, F., Baur, S., Couairon, G., Chen, A., and Bikard, D. (2021). Generating functional protein variants with variational autoencoders. PLoS computational biology , 17(2):e1008736

    2021

  41. [49]

    He, Y., Zheng, Z., Soga, P., Zhu, Y., Dong, Y., and Li, J. (2024). Explaining graph neural networks with large language models: A counterfactual perspective on molecule graphs. In Findings of the Association for Computational Linguistics: EMNLP 2024 , pages 7079--7096

    2024

  42. [50]

    Hou, Z., Xu, Z., Yan, C., Luo, H., and Luo, J. (2025). Cpi-ggs: A deep learning model for predicting compound-protein interaction based on graphs and sequences. Computational Biology and Chemistry , 115:108326

    2025

  43. [51]

    Hu, W., Liu, B., Gomes, J., Zitnik, M., Liang, P., Pande, V., and Leskovec, J. (2019). Strategies for pre-training graph neural networks. arXiv preprint arXiv:1905.12265

    2019 arXiv

  44. [52]

    Hu, Y., Yang, H., Li, M., Zhong, Z., Zhou, Y., Bai, F., and Wang, Q. (2024). Exploring protein conformational changes using a large-scale biophysical sampling augmented deep learning strategy. Advanced Science , 11(44):2400884

    2024

  45. [53]

    Huang, Q., Yamada, M., Tian, Y., Singh, D., and Chang, Y. (2022). Graphlime: Local interpretable model explanations for graph neural networks. IEEE Transactions on Knowledge and Data Engineering , 35(7):6968--6972

    2022

  46. [54]

    Imre, A., Balogh, B., and M \'a ndity, I. (2025). Graphcpp: The new state-of-the-art method for cell-penetrating peptide prediction via graph neural networks. British Journal of Pharmacology , 182(3):495--509

    2025

  47. [55]

    Ingraham, J., Garg, V., Barzilay, R., and Jaakkola, T. (2019). Generative models for graph-based protein design. In Wallach, H., Larochelle, H., Beygelzimer, A., d Alch\' e -Buc, F., Fox, E., and Garnett, R., editors, Advances in Neural Information Processing Systems , volume ...

    2019

  48. [56]

    B., Baranov, M., Costello, Z., Barber, K

    Ingraham, J. B., Baranov, M., Costello, Z., Barber, K. W., Wang, W., Ismail, A., Frappier, V., Lord, D. M., Ng-Thow-Hing, C., Vlack, E. R. V., Tie, S., Xue, V., Cowles, S. C., Leung, A., Rodrigues, J. V., Morales-Perez, C. L., Ayoub, A. M., Green, R., Puentes, K., Oplinger, F....

    2023

  49. [57]

    Isert, C., Atz, K., and Schneider, G. (2023). Structure-based drug design with geometric deep learning. Current Opinion in Structural Biology , 79:102548

    2023

  50. [58]

    Jamasb, A., Vi \ n as Torn \'e , R., Ma, E., Du, Y., Harris, C., Huang, K., Hall, D., Lio, P., and Blundell, T. (2022). Graphein-a python library for geometric deep learning and network analysis on biomolecular structures and interaction networks. Advances in Neural Informatio...

    2022

  51. [59]

    Jha, K., Saha, S., and Singh, H. (2022). Prediction of protein--protein interaction using graph neural networks. Scientific Reports , 12(1):8360

    2022

  52. [60]

    Jia, M., Yao, Y., and Liu, Y. (2025). Review on graph neural networks for process soft sensor development, fault diagnosis, and process monitoring. Industrial & Engineering Chemistry Research

    2025

  53. [61]

    Jiang, J., Zhang, C., Ke, L., Hayes, N., Zhu, Y., Qiu, H., Zhang, B., Zhou, T., and Wei, G.-W. (2025). A review of machine learning methods for imbalanced data challenges in chemistry. Chemical Science , 16:7637--7658

    2025

  54. [62]

    Jiao, P., Wang, B., Wang, X., Liu, B., Wang, Y., and Li, J. (2023). Struct2go: protein function prediction based on graph pooling algorithm and alphafold2 structure information. Bioinformatics , 39(10):btad637

    2023

  55. [63]

    Jim \'e nez, J., Skalic, M., Martinez-Rosell, G., and De Fabritiis, G. (2018). K deep: protein--ligand absolute binding affinity prediction via 3d-convolutional neural networks. Journal of chemical information and modeling , 58(2):287--296

    2018

  56. [64]

    Jin, W., Wohlwend, J., Barzilay, R., and Jaakkola, T. (2022). Iterative refinement graph neural network for antibody sequence-structure co-design

    2022

  57. [65]

    Jin, Y., Lu, J., Shi, R., and Yang, Y. (2021). Embeddti: Enhancing the molecular representations via sequence embedding and graph convolutional network for the prediction of drug-target interaction. Biomolecules , 11(12):1783

    2021

  58. [66]

    Jing, B., Eismann, S., Suriana, P., Townshend, R. J. L., and Dror, R. (2021). Learning from protein structure with geometric vector perceptrons

    2021

  59. [67]

    Jumper, J., Evans, R., Pritzel, A., Green, T., Figurnov, M., Ronneberger, O., Tunyasuvunakool, K., Bates, R., Žídek, A., Potapenko, A., Bridgland, A., Meyer, C., Kohl, S. A. A., Ballard, A. J., Cowie, A., Romera-Paredes, B., Nikolov, S., Jain, R., Adler, J., Back, T., Petersen...

    2021

  60. [68]

    Kakkad, J., Jannu, J., Sharma, K., Aggarwal, C., and Medya, S. (2023). A survey on explainability of graph neural networks. arXiv preprint arXiv:2306.01958

    2023 arXiv

  61. [69]

    Khakzad, H., Igashov, I., Schneuing, A., Goverde, C., Bronstein, M., and Correia, B. (2023). A new age in protein design empowered by deep learning. Cell Systems , 14:925--939

    2023

  62. [70]

    Kipf, T. N. and Welling, M. (2016). Variational graph auto-encoders. arXiv preprint arXiv:1611.07308

    2016 arXiv

  63. [71]

    E., and Kolinski, A

    Kmiecik, S., Gront, D., Kolinski, M., Wieteska, L., Dawid, A. E., and Kolinski, A. (2016). Coarse-grained protein models and their applications. Chemical Reviews , 116(14):7898--7936. PMID: 27333362

    2016

  64. [72]

    F., Abriata, L

    Krapp, L. F., Abriata, L. A., Rodriguez, F. C., and Peraro, M. D. (2023). Pesto: parameter-free geometric deep learning for accurate prediction of protein binding interfaces. Nature Communications , 14:2175

    2023

  65. [73]

    F., Meireles, F

    Krapp, L. F., Meireles, F. A., Abriata, L. A., Devillard, J., Vacle, S., Marcaida, M. J., and Peraro, M. D. (2024). Context-aware geometric deep learning for protein sequence design. Nature Communications , 15:6273

    2024

  66. [74]

    R., Morey-Burrows, F

    Krishna, R., Wang, J., Ahern, W., Sturmfels, P., Venkatesh, P., Kalvet, I., Lee, G. R., Morey-Burrows, F. S., Anishchenko, I., Humphreys, I. R., McHugh, R., Vafeados, D., Li, X., Sutherland, G. A., Hitchcock, A., Hunter, C. N., Kang, A., Brackenbrough, E., Bera, A. K., Baek, M...

    2024

  67. [75]

    W., Qiu, T., and Batista, V

    Kyro, G. W., Qiu, T., and Batista, V. S. (2025). A model-centric review of deep learning for protein design. arXiv preprint arXiv:2502.19173

    2025 arXiv

  68. [76]

    Lavecchia, A. (2024). Advancing drug discovery with deep attention neural networks. Drug Discovery Today , page 104067

    2024

  69. [77]

    Lei, Y., Wang, X., Fang, M., Li, H., Li, X., and Zeng, J. (2024). Pepgb: facilitating peptide drug discovery via graph neural networks. arXiv preprint arXiv:2401.14665

    2024 arXiv

  70. [78]

    and Ming, D

    Li, B. and Ming, D. (2024). Gatsol, an enhanced predictor of protein solubility through the synergy of 3d structure graph and large language modeling. BMC Bioinformatics , 25:204

    2024

  71. [79]

    Li, G., Yao, S., and Fan, L. (2024). Prostage: Predicting effects of mutations on protein stability by using protein embeddings and graph convolutional networks. Journal of Chemical Information and Modeling , 64:340--347

    2024

  72. [80]

    Li, K., Quan, L., Jiang, Y., Wu, H., Wu, J., Li, Y., Zhou, Y., Wu, T., and Lyu, Q. (2023a). Simultaneous prediction of interaction sites on the protein and peptide sides of complexes through multilayer graph convolutional networks. Journal of Chemical Information and Modeling ...

    2023

  73. [81]

    and Liu, Z.-P

    Li, P. and Liu, Z.-P. (2023). Geobind: segmentation of nucleic acid binding interface on protein surface with geometric deep learning. Nucleic Acids Research , 51(10):e60--e60

    2023

  74. [82]

    Li, S., Liu, F., Jiao, L., Chen, P., Liu, X., and Li, L. (2022). Mfnet: A novel gnn-based multi-level feature network with superpixel priors. IEEE Transactions on Image Processing , 31:7306--7321

    2022

  75. [83]

    F., Medina-Ortiz, D., Davari, M

    Li, W.-r., Cadet, X. F., Medina-Ortiz, D., Davari, M. D., Sowdhamini, R., Damour, C., Li, Y., Miranville, A., and Cadet, F. (2025). From thermodynamics to protein design: Diffusion models for biomolecule generation towards autonomous protein engineering. arXiv preprint arXiv:2...

    2025 arXiv

  76. [84]

    Li, Z., Jiang, C., and Li, J. (2023b). Deepgatgo: A hierarchical pretraining-based graph-attention model for automatic protein function prediction

    2023

  77. [85]

    C., Tresadern, G., and Bonnet, P

    Libouban, P.-Y., Aci-S \`e che, S., G \'o mez-Tamayo, J. C., Tresadern, G., and Bonnet, P. (2023). The impact of data on structure-based binding affinity predictions using deep neural networks. International Journal of Molecular Sciences , 24(22):16120

    2023

  78. [86]

    Lin, E., Lin, C.-H., and Lane, H.-Y. (2022). De novo peptide and protein design using generative adversarial networks: an update. Journal of Chemical Information and Modeling , 62(4):761--774

    2022

  79. [87]

    and AlQuraishi, M

    Lin, Y. and AlQuraishi, M. (2023). Generating novel, designable, and diverse protein structures by equivariantly diffusing oriented residue clouds

    2023

  80. [88]

    Lin, Z., Akin, H., Rao, R., Hie, B., Zhu, Z., Lu, W., Smetanin, N., Verkuil, R., Kabeli, O., Shmueli, Y., et al. (2023). Evolutionary-scale prediction of atomic-level protein structure with a language model. Science , 379(6637):1123--1130

    2023

  81. [89]

    Liu, J., Guo, Z., You, H., Zhang, C., and Lai, L. (2024a). All-atom protein sequence design based on geometric deep learning. Angewandte Chemie , 136(50):e202411461

    2024

  82. [90]

    Liu, J., Guo, Z., You, H., Zhang, C., and Lai, L. (2024b). All‐atom protein sequence design based on geometric deep learning. Angewandte Chemie , 136

    2024

  83. [91]

    Liu, M., Li, C., Chen, R., Cao, D., and Zeng, X. (2024c). Geometric deep learning for drug discovery. Expert Systems with Applications , 240:122498

    2024

  84. [92]

    Liu, Q., Xun, G., and Feng, Y. (2019). The state-of-the-art strategies of protein engineering for enzyme stabilization. Biotechnology advances , 37(4):530--537

    2019

  85. [93]

    Lopardo, G., Precioso, F., and Garreau, D. (2024). Attention meets post-hoc interpretability: A mathematical perspective. arXiv preprint arXiv:2402.03485

    2024

  86. [94]

    Z., Lozano, A., Chenthamarakshan, V., Das, P., and Tang, J

    Lu, J., Chen, X., Lu, S. Z., Lozano, A., Chenthamarakshan, V., Das, P., and Tang, J. (2025). Aligning protein conformation ensemble generation with physical feedback. arXiv preprint arXiv:2505.24203

    2025 arXiv

  87. [95]

    A., Tolomei, G., De Rijke, M., and Silvestri, F

    Lucic, A., Ter Hoeve, M. A., Tolomei, G., De Rijke, M., and Silvestri, F. (2022). Cf-gnnexplainer: Counterfactual explanations for graph neural networks. In International Conference on Artificial Intelligence and Statistics , pages 4499--4511. PMLR

    2022

  88. [96]

    Luo, D., Cheng, W., Xu, D., Yu, W., Zong, B., Chen, H., and Zhang, X. (2020). Parameterized explainer for graph neural network. Advances in neural information processing systems , 33:19620--19631

    2020

  89. [97]

    Luo, J., Zhao, K., Chen, J., Yang, C., Qu, F., Liu, Y., Jin, X., Yan, K., Zhang, Y., and Liu, B. (2024). imfp-lg: Identify novel multi-functional peptides using protein language models and graph-based deep learning. Genomics, Proteomics & Bioinformatics , 22(6):qzae084

    2024

  90. [98]

    Luo, X., Wang, L., Hu, P., and Hu, L. (2023). Predicting protein-protein interactions using sequence and network information via variational graph autoencoder. IEEE/ACM Transactions on Computational Biology and Bioinformatics , 20(5):3182--3194

    2023

  91. [99]

    and Tarakanova, A

    Madani, M. and Tarakanova, A. (2024). Predicting protein thermal stability changes upon single and multi-point mutations via restricted attention subgraph neural network. Journal of the Mechanics and Physics of Solids , 184

    2024

  92. [100]

    and Bayzid, M

    Mahbub, S. and Bayzid, M. S. (2022). Egret: edge aggregated graph attention networks and transfer learning improve protein--protein interaction site prediction. Briefings in Bioinformatics , 23(2):bbab578

    2022

  93. [101]

    R., and Baker, D

    Mansoor, S., Baek, M., Park, H., Lee, G. R., and Baker, D. (2024). Protein ensemble generation through variational autoencoder latent space sampling. Journal of Chemical Theory and Computation , 20(7):2689--2695

    2024

  94. [102]

    Medina-Ortiz, D., Contreras, S., Fern \'a ndez, D., Soto-Garc \' a, N., Moya, I., Cabas-Mora, G., and Olivera-Nappa, \'A . (2024a). Protein language models and machine learning facilitate the identification of antimicrobial peptides. International Journal of Molecular Sciences...

    2024

  95. [103]

    Medina-Ortiz, D., Contreras, S., Quiroz, C., and Olivera-Nappa, \'A . (2020). Development of supervised learning predictive models for highly non-linear biological, biomedical, and general datasets. Frontiers in molecular biosciences , 7:13

    2020

  96. [104]

    Medina-Ortiz, D., Khalifeh, A., Anvari-Kazemabad, H., and Davari, M. D. (2024b). Interpretable and explainable predictive machine learning models for data-driven protein engineering. Biotechnology Advances , page 108495

    2024

  97. [105]

    Meng, L., Wei, L., and Wu, R. (2025a). Mvgnn-ppis: A novel multi-view graph neural network for protein-protein interaction sites prediction based on alphafold3-predicted structures and transfer learning. International Journal of Biological Macromolecules , page 140096

    2025

  98. [106]

    Meng, Y., Zhang, Z., Zhou, C., Tang, X., Hu, X., Tian, G., Yang, J., and Yao, Y. (2025b). Protein structure prediction via deep learning: an in-depth review. Frontiers in Pharmacology , 16:1498662

    2025

  99. [107]

    Min, Y., Wenkel, F., and Wolf, G. (2020). Scattering gcn: Overcoming oversmoothness in graph convolutional networks. Advances in neural information processing systems , 33:14498--14508

    2020

  100. [108]

    Montavon, G., Samek, W., and M \"u ller, K.-R. (2018). Methods for interpreting and understanding deep neural networks. Digital signal processing , 73:1--15

    2018

  101. [109]

    Moon, S., Hwang, S.-Y., Lim, J., and Kim, W. Y. (2024). Pignet2: a versatile deep learning-based protein--ligand interaction prediction model for binding affinity scoring and virtual screening. Digital Discovery , 3(2):287--299

    2024

  102. [110]

    Morehead, A., Giri, N., Liu, J., Neupane, P., and Cheng, J. (2025). Deep learning for protein-ligand docking: Are we there yet? ArXiv , pages arXiv--2405

    2025

  103. [111]

    T., Starck, S., Dima, A., Wunderlich, S., Bintsi, K.-M., Zaripova, K., Braren, R., Rueckert, D., Kazi, A., and Kaissis, G

    M \"u ller, T. T., Starck, S., Dima, A., Wunderlich, S., Bintsi, K.-M., Zaripova, K., Braren, R., Rueckert, D., Kazi, A., and Kaissis, G. (2024). A survey on graph construction for geometric deep learning in medicine: Methods and recommendations. Transactions on Machine Learni...

    2024

  104. [112]

    and Andonie, R

    Musat, B. and Andonie, R. (2022). Information bottleneck in deep learning-a semiotic approach. International Journal of Computers Communications & Control , 17(1)

    2022

  105. [113]

    P., Ortner, C., Cs \'a nyi, G., and Ceriotti, M

    Musil, F., Grisafi, A., Bart \'o k, A. P., Ortner, C., Cs \'a nyi, G., and Ceriotti, M. (2021). Physics-inspired structural representations for molecules and materials. Chemical Reviews , 121(16):9759--9815

    2021

  106. [114]

    Musil, M., Jezik, A., Horackova, J., Borko, S., Kabourek, P., Damborsky, J., and Bednar, D. (2024). Fireprot 2.0: web-based platform for the fully automated design of thermostable proteins. Briefings in Bioinformatics , 25(1):bbad425

    2024

  107. [115]

    Nandan, M., Mitra, S., and De, D. (2025). Graphxai: a survey of graph neural networks (gnns) for explainable ai (xai). Neural Computing and Applications , pages 1--52

    2025

  108. [116]

    P., Nguyen, T., Le, T

    Nguyen, T., Le, H., Quinn, T. P., Nguyen, T., Le, T. D., and Venkatesh, S. (2021a). Graphdta: predicting drug--target binding affinity with graph neural networks. Bioinformatics , 37(8):1140--1147

    2021

  109. [117]

    M., Nguyen, T., Le, T

    Nguyen, T. M., Nguyen, T., Le, T. M., and Tran, T. (2021b). Gefa: early fusion approach in drug-target affinity prediction. IEEE/ACM transactions on computational biology and bioinformatics , 19(2):718--728

    2021

  110. [118]

    N., Marks, D., and Gal, Y

    Notin, P., Dias, M., Frazer, J., Marchena-Hurtado, J., Gomez, A. N., Marks, D., and Gal, Y. (2022). Tranception: protein fitness prediction with autoregressive transformers and inference-time retrieval. In International Conference on Machine Learning , pages 16990--17017. PMLR

    2022

  111. [119]

    Notin, P., Kollasch, A., Ritter, D., Van Niekerk, L., Paul, S., Spinner, H., Rollins, N., Shaw, A., Orenbuch, R., Weitzman, R., et al. (2023). Proteingym: Large-scale benchmarks for protein fitness prediction and design. Advances in Neural Information Processing Systems , 36:6...

    2023

  112. [120]

    Panagiotaki, E., De Martini, D., and Kunze, L. (2023). Semantic interpretation and validation of graph attention-based explanations for gnn models. In 2023 21st International Conference on Advanced Robotics (ICAR) , pages 375--380. IEEE

    2023

  113. [121]

    Petrini, L. (2023). Breaking the curse of dimensionality in deep neural networks by learning invariant representations. arXiv preprint arXiv:2310.16154

    2023 arXiv

  114. [122]

    M., Gulati, K., Poluri, K

    Poluri, K. M., Gulati, K., Poluri, K. M., and Gulati, K. (2017). Biotechnological and biomedical applications of protein engineering methods. Protein Engineering Techniques: Gateways to Synthetic Protein Universe , pages 103--134

    2017

  115. [123]

    Ranjan, E., Sanyal, S., and Talukdar, P. (2020). Asap: Adaptive structure aware pooling for learning hierarchical graph representations. In Proceedings of the AAAI conference on artificial intelligence , volume 34, pages 5470--5477

    2020

  116. [124]

    Rao, B., Zhang, L., and Zhang, G. (2020). Acp-gcn: the identification of anticancer peptides based on graph convolution networks. IEEE Access , 8:176005--176011

    2020

  117. [125]

    Rao, R., Bhattacharya, N., Thomas, N., Duan, Y., Chen, P., Canny, J., Abbeel, P., and Song, Y. (2019). Evaluating protein transfer learning with tape. Advances in neural information processing systems , 32

    2019

  118. [126]

    H., Tarafder, S., and Bhattacharya, D

    Roche, R., Moussad, B., Shuvo, M. H., Tarafder, S., and Bhattacharya, D. (2024). Equipnas: improved protein--nucleic acid binding site prediction using protein-language-model-informed equivariant deep graph neural networks. Nucleic Acids Research , 52(5):e27--e27

    2024

  119. [127]

    C., Cifuentes, J., Mu \ n oz-Camargo, C., Reyes, L

    Ruiz Puentes, P., Henao, M. C., Cifuentes, J., Mu \ n oz-Camargo, C., Reyes, L. H., Cruz, J. C., and Arbel \'a ez, P. (2022). Rational discovery of antimicrobial peptides by means of artificial intelligence. Membranes , 12(7):708

    2022

  120. [128]

    Sabei, A., Hognon, C., Martin, J., and Frezza, E. (2024). Dynamics of protein--rna interfaces using all-atom molecular dynamics simulations. The Journal of Physical Chemistry B , 128(20):4865--4886

    2024

  121. [129]

    Saranti, A., Pfeifer, B., Gollob, C., Stampfer, K., and Holzinger, A. (2024). From 3d point-cloud data to explainable geometric deep learning: State-of-the-art and future challenges. Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery , 14(6):e1554

    2024

  122. [130]

    Sarumi, O. A. and Heider, D. (2024). Large language models and their applications in bioinformatics. Computational and Structural Biotechnology Journal

    2024

  123. [131]

    G., Hoogeboom, E., and Welling, M

    Satorras, V. G., Hoogeboom, E., and Welling, M. (2021). E (n) equivariant graph neural networks. In International conference on machine learning , pages 9323--9332. PMLR

    2021

  124. [132]

    u tt, K. T., M \

    Schnake, T., Eberle, O., Lederer, J., Nakajima, S., Sch \"u tt, K. T., M \"u ller, K.-R., and Montavon, G. (2021). Higher-order explanations of graph neural networks via relevant walks. IEEE transactions on pattern analysis and machine intelligence , 44(11):7581--7596

    2021

  125. [133]

    I., and Deane, C

    Schneider, C., Raybould, M. I., and Deane, C. M. (2022). Sabdab in the age of biotherapeutics: updates including sabdab-nano, the nanobody structure tracker. Nucleic acids research , 50(D1):D1368--D1372

    2022

  126. [134]

    Shao, Y., Li, H., Gu, X., Yin, H., Li, Y., Miao, X., Zhang, W., Cui, B., and Chen, L. (2024). Distributed graph neural network training: A survey. ACM Computing Surveys , 56:39

    2024

  127. [135]

    M., and Tastan Bishop,

    Sheik Amamuddy, O., Veldman, W., Manyumwa, C., Khairallah, A., Agajanian, S., Oluyemi, O., Verkhivker, G. M., and Tastan Bishop, . (2020). Integrated computational approaches and tools for allosteric drug discovery. International journal of molecular sciences , 21(3):847

    2020

  128. [136]

    Shi, C., Wang, C., Lu, J., Zhong, B., and Tang, J. (2023). Protein sequence and structure co-design with equivariant translation

    2023

  129. [137]

    Shlens, J. (2014). Notes on kullback-leibler divergence and likelihood. arXiv preprint arXiv:1404.2000

    2014 arXiv

  130. [138]

    Shrikumar, A., Greenside, P., and Kundaje, A. (2017). Learning important features through propagating activation differences. In International conference on machine learning , pages 3145--3153. PMlR

    2017

  131. [139]

    and Tishby, N

    Shwartz-Ziv, R. and Tishby, N. (2017). Opening the black box of deep neural networks via information. arXiv preprint arXiv:1703.00810

    2017 arXiv

  132. [140]

    and Zacharias, M

    Siebenmorgen, T. and Zacharias, M. (2020). Computational prediction of protein--protein binding affinities. Wiley Interdisciplinary Reviews: Computational Molecular Science , 10(3):e1448

    2020

  133. [141]

    Song, Y., Yuan, Q., Zhao, H., and Yang, Y. (2023a). Accurately identifying nucleic-acid-binding sites through geometric graph learning on language model predicted structures. Briefings in Bioinformatics , 24(6):bbad360

    2023

  134. [142]

    Song, Y., Zhou, C., Wang, X., and Lin, Z. (2023b). Ordered gnn: Ordering message passing to deal with heterophily and over-smoothing. arXiv preprint arXiv:2302.01524

    2023 arXiv

  135. [143]

    Srivastava, N., Hinton, G., Krizhevsky, A., Sutskever, I., and Salakhutdinov, R. (2014). Dropout: a simple way to prevent neural networks from overfitting. The journal of machine learning research , 15(1):1929--1958

    2014

  136. [144]

    St \"a rk, H., Ganea, O., Pattanaik, L., Barzilay, R., and Jaakkola, T. (2022). Equibind: Geometric deep learning for drug binding structure prediction. In International conference on machine learning , pages 20503--20521. PMLR

    2022

  137. [145]

    Strokach, A., Becerra, D., Corbi-Verge, C., Perez-Riba, A., and Kim, P. M. (2020). Fast and flexible protein design using deep graph neural networks. Cell Systems , 11(4):402--411.e4

    2020

  138. [146]

    Sun, R., Wu, L., Lin, H., Huang, Y., and Li, S. Z. (2024). Enhancing protein predictive models via proteins data augmentation: A benchmark and new directions

    2024

  139. [147]

    Sun, T.-J., Bu, H.-L., Yan, X., Sun, Z.-H., Zha, M.-S., and Dong, G.-F. (2022). Labampsgcn: A framework for identifying lactic acid bacteria antimicrobial peptides based on graph convolutional neural network. Frontiers in Genetics , 13:1062576

    2022

  140. [148]

    E., and Bronstein, M

    Sverrisson, F., Feydy, J., Correia, B. E., and Bronstein, M. M. (2020). Fast end-to-end learning on protein surfaces. bioRxiv

    2020

  141. [149]

    Tarafder, S., Roche, R., and Bhattacharya, D. (2024). The landscape of rna 3d structure modeling with transformer networks. Biology Methods and Protocols , 9(1):bpae047

    2024

  142. [150]

    Teague, S. J. (2003). Implications of protein flexibility for drug discovery. Nature reviews Drug discovery , 2(7):527--541

    2003

  143. [151]

    Thuillier, K., Baroukh, C., Bockmayr, A., Cottret, L., Paulevé, L., and Siegel, A. (2022). Merrin: Metabolic regulation rule inference from time series data. Bioinformatics , 38(Supplement\_2):ii127--ii133

    2022

  144. [152]

    Varadi, M., Tsenkov, M., and Velankar, S. (2025). Challenges in bridging the gap between protein structure prediction and functional interpretation. Proteins: Structure, Function, and Bioinformatics , 93(1):400--410

    2025

  145. [153]

    Veli c kovi \'c , P., Cucurull, G., Casanova, A., Romero, A., Lio, P., and Bengio, Y. (2017). Graph attention networks. 6th International Conference on Learning Representations, ICLR 2018 - Conference Track Proceedings

    2017

  146. [154]

    Verma, Y., Heinonen, M., and Garg, V. (2023). Abode: Ab initio antibody design using conjoined odes

    2023

  147. [155]

    S., Isalan, M., Heap, J

    Vidal, L. S., Isalan, M., Heap, J. T., and Ledesma-Amaro, R. (2023). A primer to directed evolution: current methodologies and future directions. RSC Chemical Biology , 4(4):271--291

    2023

  148. [156]

    Vinchurkar, T., Abdelmaqsoud, K., and Kitchin, J. R. (2025). Uncertainty quantification in graph neural networks with shallow ensembles. arXiv preprint arXiv:2504.12627

    2025 arXiv

  149. [157]

    and Thai, M

    Vu, M. and Thai, M. T. (2020). Pgm-explainer: Probabilistic graphical model explanations for graph neural networks. Advances in neural information processing systems , 33:12225--12235

    2020

  150. [158]

    Wang, J., Chen, S., Yuan, Q., Chen, J., Li, D., Wang, L., and Yang, Y. (2024a). Predicting the effects of mutations on protein solubility using graph convolution network and protein language model representation. Journal of Computational Chemistry , 45:436--445

    2024

  151. [159]

    Wang, S., Tang, H., Shan, P., Wu, Z., and Zuo, L. (2023). Pros-gnn: Predicting effects of mutations on protein stability using graph neural networks. Computational Biology and Chemistry , 107:107952

    2023

  152. [160]

    Wang, S., Tang, H., Zhao, Y., and Zuo, L. (2022). Bayestab: Predicting effects of mutations on protein stability with uncertainty quantification. Protein Science , 31(11):e4467

    2022

  153. [161]

    Wang, Y., Hou, W., Sheng, N., Zhao, Z., Liu, J., Huang, L., and Wang, J. (2024b). Graph pooling in graph neural networks: Methods and their applications in omics studies. Artificial Intelligence Review , 57(11):294

    2024

  154. [162]

    A., Golovach, N., Alford, R., Rangwala, H., and Clark, P

    Wang, Z., Brand, R., Adolf-Bryfogle, J., Grewal, J., Qi, Y., Combs, S. A., Golovach, N., Alford, R., Rangwala, H., and Clark, P. M. (2024c). Eggnet, a generalizable geometric deep learning framework for protein complex pose scoring. ACS omega , 9(7):7471--7479

    2024

  155. [163]

    L., Juergens, D., Bennett, N

    Watson, J. L., Juergens, D., Bennett, N. R., Trippe, B. L., Yim, J., Eisenach, H. E., Ahern, W., Borst, A. J., Ragotte, R. J., Milles, L. F., Wicky, B. I. M., Hanikel, N., Pellock, S. J., Courbet, A., Sheffler, W., Wang, J., Venkatesh, P., Sappington, I., Torres, S. V., Lauko,...

    2023

  156. [164]

    Wei, L., Ye, X., Xue, Y., Sakurai, T., and Wei, L. (2021). Atse: a peptide toxicity predictor by exploiting structural and evolutionary information based on graph neural network and attention mechanism. Briefings in Bioinformatics , 22(5):bbab041

    2021

  157. [165]

    Wu, F., Tao, Y., Radev, D., and Xu, J. (2023a). When geometric deep learning meets pretrained protein language models. bioRxiv , pages 2023--01

    2023

  158. [166]

    Wu, F., Wu, L., Radev, D., Xu, J., and Li, S. Z. (2023b). Integration of pre-trained protein language models into geometric deep learning networks. Communications Biology , 6(1):876

    2023

  159. [167]

    and Zhou, J.-t

    Wu, Y. and Zhou, J.-t. (2025). A hierarchical and interlamination graph self-attention mechanism-based knowledge graph reasoning architecture. Information Sciences , 686:121345

    2025

  160. [168]

    and Ku, W

    Xia, T. and Ku, W. S. (2021). Geometric graph representation learning on protein structure prediction. In Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining , pages 1873--1883. Association for Computing Machinery

    2021

  161. [169]

    Xie, J., Zhang, Y., Wang, Z., Jin, X., Lu, X., Ge, S., and Min, X. (2025). Ppi-graphomer: enhanced protein-protein affinity prediction using pretrained and graph transformer models. BMC bioinformatics , 26(1):116

    2025

  162. [170]

    Yan, K., Lv, H., Guo, Y., Peng, W., and Liu, B. (2023). samppred-gat: prediction of antimicrobial peptide by graph attention network and predicted peptide structure. Bioinformatics , 39(1):btac715

    2023

  163. [171]

    K., Wu, Z., and Arnold, F

    Yang, K. K., Wu, Z., and Arnold, F. H. (2019). Machine-learning-guided directed evolution for protein engineering. Nature methods , 16(8):687--694

    2019

  164. [172]

    L., Bortoli, V

    Yim, J., Trippe, B. L., Bortoli, V. D., Mathieu, E., Doucet, A., Barzilay, R., and Jaakkola, T. (2023). Se(3) diffusion model with application to protein backbone generation

    2023

  165. [173]

    You, R., Yao, S., Mamitsuka, H., and Zhu, S. (2021). Deepgraphgo: graph neural network for large-scale, multispecies protein function prediction. Bioinformatics , 37(Supplement\_1):i262--i271

    2021

  166. [174]

    Yu, J., Xu, T., and He, R. (2021). Towards the explanation of graph neural networks in digital pathology with information flows. arXiv preprint arXiv:2112.09895

    2021 arXiv

  167. [175]

    Yuan, H., Tang, J., Hu, X., and Ji, S. (2020). Xgnn: Towards model-level explanations of graph neural networks. In Proceedings of the 26th ACM SIGKDD international conference on knowledge discovery & data mining , pages 430--438

    2020

  168. [176]

    Yuan, H., Yu, H., Wang, J., Li, K., and Ji, S. (2021). On explainability of graph neural networks via subgraph explorations. In International conference on machine learning , pages 12241--12252. PMLR

    2021

  169. [177]

    Yuan, Q., Chen, S., Rao, J., Zheng, S., Zhao, H., and Yang, Y. (2022). Alphafold2-aware protein--dna binding site prediction using graph transformer. Briefings in bioinformatics , 23(2):bbab564

    2022

  170. [178]

    Yuan, Q., Tian, C., Song, Y., Ou, P., Zhu, M., Zhao, H., and Yang, Y. (2024). Gpsfun: geometry-aware protein sequence function predictions with language models. Nucleic Acids Research , 52:W248--W255

    2024

  171. [179]

    Zhang, C., Sun, Y., and Hu, P. (2025). An interpretable deep geometric learning model to predict the effects of mutations on protein--protein interactions using large-scale protein language model. Journal of Cheminformatics , 17(1):35

    2025

  172. [180]

    Zhang, Z., Xu, M., Jamasb, A., Chenthamarakshan, V., Lozano, A., Das, P., and Tang, J. (2023). Protein representation learning by geometric structure pretraining

    2023

  173. [181]

    Zhao, C., Liu, T., and Wang, Z. (2022). Panda2: protein function prediction using graph neural networks. NAR Genomics and Bioinformatics , 4(1):lqac004

    2022

  174. [182]

    Zheng, M., Sun, G., Li, X., and Fan, Y. (2024). Egpdi: identifying protein--dna binding sites based on multi-view graph embedding fusion. Briefings in Bioinformatics , 25(4):bbae330

    2024

  175. [183]

    Zhou, Y., Jiang, Y., and Yang, Y. (2023). Agat-ppis: a novel protein--protein interaction site predictor based on augmented graph attention network with initial residual and identity mapping. Briefings in bioinformatics , 24(3):bbad122

    2023

  176. [184]

    H., and Ascher, D

    Zhou, Y., Myung, Y., Rodrigues, C. H., and Ascher, D. B. (2024). Ddmut-ppi: predicting effects of mutations on protein--protein interactions using graph-based deep learning. Nucleic Acids Research , 52(W1):W207--W214

    2024

  177. [185]

    Zhu, W., Ding, X., Shen, H.-B., and Pan, X. (2025). Identifying rna-small molecule binding sites using geometric deep learning with language models. Journal of Molecular Biology , page 169010

    2025

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.