pith. sign in

arxiv: 2606.02629 · v1 · pith:AQBXLG7Knew · submitted 2026-05-30 · 🧬 q-bio.QM · cs.AI· cs.LG

Enhancing Protein-Protein Interaction Prediction with Hierarchical Motif-based Multimodal Protein Embedding

Zaifei Yang , Samuel Ping-Man Choi , James Kwok This is my paper

Pith reviewed 2026-06-28 18:11 UTC · model grok-4.3

classification 🧬 q-bio.QM cs.AIcs.LG
keywords protein-protein interactionmultimodal embeddinghierarchical motifPPI predictionmeso-scale motifssequence structure functionmulti-label classification
0
0 comments X

The pith

A hierarchical motif-based multimodal encoder outperforms state-of-the-art models on protein-protein interaction prediction.

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

Existing PPI prediction methods overlook the hierarchical organization of proteins, particularly meso-scale motifs that regulate interactions, and fail to integrate sequence, structure, and function modalities effectively. MMM-PPI builds protein embeddings bottom-up across three scales to address these gaps: micro-scale encoding of three modal residue features, meso-scale aggregation into spatially-informed motif embeddings, and macro-scale integration of motifs into protein embeddings while modeling motif importance and inter-modal correlations. The resulting pre-trained encoder applies directly to large-scale prediction tasks. Experiments on multiple datasets show consistent outperformance over prior multi-label models, with the largest gains appearing in challenging data partitions and limited-data regimes. A reader would care because more reliable PPI maps could clarify cellular mechanisms and support applications such as target identification.

Core claim

MMM-PPI constructs PPI embeddings in a bottom-up multi-modal manner across three scales. At the micro-scale, three modal residue features are encoded. At the meso-scale, a multimodal motif encoder aggregates residues into spatially-informed motif embeddings. At the macro-scale, a multimodal protein encoder integrates motifs into protein embeddings by jointly modeling motif importance and inter-modal correlations. The pre-trained encoder yields superior performance on multiple PPI datasets compared with state-of-the-art multi-label models, especially under challenging data partitions and limited data scenarios.

What carries the argument

The Hierarchical Motif-based Multi-Modal protein Encoder that aggregates multimodal features bottom-up from micro-scale residues through meso-scale motifs to macro-scale proteins while capturing motif importance and cross-modal correlations.

If this is right

  • The pre-trained encoder can be used off-the-shelf for large-scale PPI prediction without retraining.
  • Gains are largest on challenging data partitions that split interactions by protein families or limited training sets.
  • Joint modeling of motif importance and inter-modal correlations improves multi-label interaction predictions.
  • The three-scale construction integrates sequence, structure, and function modalities more effectively than prior single-scale or single-modality approaches.

Where Pith is reading between the lines

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

  • The same hierarchical motif aggregation might transfer to predicting other biomolecular associations such as protein-RNA or protein-small-molecule interactions.
  • Meso-scale motif embeddings could serve as interpretable features for downstream biological network analysis beyond prediction accuracy.
  • Testing the encoder on time-resolved or condition-specific PPI datasets would reveal whether the motif-level representations capture regulatory dynamics.

Load-bearing premise

Meso-scale motifs critically regulate PPIs and bottom-up multimodal aggregation across the three scales will overcome the two stated limitations of prior methods.

What would settle it

Running MMM-PPI on a new PPI benchmark where ablating the meso-scale motif component produces no drop in accuracy relative to a flat multimodal baseline would falsify the claim.

Figures

Figures reproduced from arXiv: 2606.02629 by James Kwok, Samuel Ping-Man Choi, Zaifei Yang.

Figure 1
Figure 1. Figure 1: Overview of the MMM-PPI. (a) The protein pair is encoded using the hierarchical encoding process based on the multimodal motif encoder and multimodal protein encoder. (b) For PPI prediction, the pre-trained encoder is used to generate embeddings for the proteins, which are processed through a GNN for classification. (c) Overview of the multimodal motif encoder. (d) Overview of the multimodal protein encode… view at source ↗
read the original abstract

Protein-protein interactions (PPIs) are essential for many biological processes. However, existing PPI prediction approaches suffer from two major limitations: they overlook the hierarchical organization of proteins, particularly meso-scale motifs that critically regulate PPIs, and fail to effectively integrate sequence, structure, and function modalities. To address these limitations, we propose MMM-PPI, a Hierarchical Motif-based Multi-Modal protein Encoder for PPI Prediction that constructs PPI embeddings in a bottom-up multi-modal manner across three scales. At the micro-scale, we encode three modal residue features; at the meso-scale, a novel multimodal motif encoder aggregates residues into spatially-informed motif embeddings; at the macro-scale, a multimodal protein encoder integrates motifs into protein embeddings by jointly modeling motif importance and inter-modal correlations. The pre-trained encoder can be used off-the-shelf for large-scale PPI prediction. Extensive experiments on multiple PPI datasets show that MMM-PPI outperforms state-of-the-art multi-label PPI prediction models, particularly under challenging data partitions and limited data scenarios. Codes are in https://github.com/yzf-code/MMM-PPI.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

0 major / 2 minor

Summary. The manuscript introduces MMM-PPI, a hierarchical motif-based multimodal protein encoder for protein-protein interaction (PPI) prediction. It constructs embeddings bottom-up across three scales: micro-scale encoding of three modal residue features, meso-scale multimodal motif encoder that aggregates residues into spatially-informed motif embeddings, and macro-scale multimodal protein encoder that integrates motifs while modeling importance and inter-modal correlations. The pre-trained encoder is applied off-the-shelf to large-scale PPI prediction and is reported to outperform state-of-the-art multi-label PPI models on multiple datasets, with particular gains under challenging data partitions and limited-data scenarios.

Significance. If the experimental claims hold after full verification, the work could advance PPI prediction by explicitly incorporating meso-scale motifs and joint multimodal aggregation, addressing two stated limitations of prior methods. The open-source code release supports reproducibility and downstream use.

minor comments (2)
  1. The abstract asserts outperformance 'particularly under challenging data partitions and limited data scenarios' but provides no dataset names, partition definitions, baseline models, or quantitative metrics; these details are required in §4 or §5 to evaluate the central claim.
  2. The description of the meso-scale motif encoder mentions 'spatially-informed motif embeddings' without specifying how spatial information is encoded or aggregated from residue features; a concrete description or equation in §3.2 would clarify the novelty.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for the careful summary of MMM-PPI and for noting its potential to advance PPI prediction through explicit meso-scale motif modeling and joint multimodal aggregation. We appreciate the recognition that the open-source code aids reproducibility. Because the report lists no specific major comments, we provide a brief overall response and remain available to address any additional questions or to supply further verification experiments.

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The provided abstract and description outline a three-scale bottom-up multimodal encoder architecture (micro-scale residue features, meso-scale motif aggregation, macro-scale protein integration) without any equations, parameter fits, or derivations shown. No self-citations, uniqueness theorems, or ansatzes are invoked. Outperformance claims rest on external experiments rather than internal reductions to inputs. The derivation chain is an architectural proposal with no load-bearing steps that collapse by construction.

Axiom & Free-Parameter Ledger

1 free parameters · 2 axioms · 0 invented entities

The central claim rests on the domain assumption that hierarchical motifs are critical for PPIs and that standard deep learning training will produce generalizable embeddings; no invented entities or explicit free parameters are named in the abstract.

free parameters (1)
  • model hyperparameters and training settings
    Any neural network encoder requires numerous fitted parameters whose values are not specified in the abstract.
axioms (2)
  • domain assumption Proteins exhibit hierarchical organization in which meso-scale motifs critically regulate protein-protein interactions.
    Explicitly stated in the abstract as a key limitation of existing approaches that the new model addresses.
  • domain assumption Integrating sequence, structure, and function modalities at multiple scales will overcome the stated limitations of prior PPI predictors.
    Core premise underlying the design of the three-scale encoder.

pith-pipeline@v0.9.1-grok · 5728 in / 1331 out tokens · 31048 ms · 2026-06-28T18:11:41.504249+00:00 · methodology

discussion (0)

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

Reference graph

Works this paper leans on

83 extracted references · 3 canonical work pages

  1. [1]

    Microbiological Reviews , volume=

    Protein-protein interactions: methods for detection and analysis , author=. Microbiological Reviews , volume=. 1995 , publisher=

    1995

  2. [2]

    Briefings in bioinformatics , volume=

    A survey on computational models for predicting protein--protein interactions , author=. Briefings in bioinformatics , volume=. 2021 , publisher=

    2021

  3. [3]

    Future Medicinal Chemistry , volume=

    Targeting protein--protein interactions for therapeutic intervention: A challenge for the future , author=. Future Medicinal Chemistry , volume=. 2009 , publisher=

    2009

  4. [4]

    Genome Medicine , volume=

    Protein-protein interaction networks: Probing disease mechanisms using model systems , author=. Genome Medicine , volume=. 2013 , publisher=

    2013

  5. [5]

    Journal of Medicinal Chemistry , volume=

    Drug design targeting protein--protein interactions (PPIs) using multiple ligand simultaneous docking (MLSD) and drug repositioning: discovery of raloxifene and bazedoxifene as novel inhibitors of IL-6/GP130 interface , author=. Journal of Medicinal Chemistry , volume=

  6. [6]

    Current Pharmaceutical Design , volume=

    Computational drug design targeting protein-protein interactions , author=. Current Pharmaceutical Design , volume=. 2012 , publisher=

    2012

  7. [7]

    Cellular and Molecular Life Sciences , volume=

    A context-dependent and disordered ubiquitin-binding motif , author=. Cellular and Molecular Life Sciences , volume=. 2022 , publisher=

    2022

  8. [8]

    Bioinformatics , volume=

    Identification of context-dependent motifs by contrasting ChIP binding data , author=. Bioinformatics , volume=. 2010 , publisher=

    2010

  9. [9]

    International Symposium on BioInformatics and BioEngineering , pages=

    Supervised statistical and machine learning approaches to inferring pairwise and module-based protein interaction networks , author=. International Symposium on BioInformatics and BioEngineering , pages=

  10. [10]

    Bioinformatics , volume=

    Prediction of protein--protein interactions using random decision forest framework , author=. Bioinformatics , volume=. 2005 , publisher=

    2005

  11. [11]

    Nucleic Acids Research , volume=

    Using support vector machine combined with auto covariance to predict protein--protein interactions from protein sequences , author=. Nucleic Acids Research , volume=. 2008 , publisher=

    2008

  12. [12]

    Heterogeneous data integration by tree-augmented naive

    Lin, Xiaotong and Chen, Xue-wen , journal=. Heterogeneous data integration by tree-augmented naive. 2013 , publisher=

    2013

  13. [13]

    International Conference on Advanced Intelligent Computing Theories and Applications , pages=

    Detection of protein-protein interactions from amino acid sequences using a rotation forest model with a novel PR-LPQ descriptor , author=. International Conference on Advanced Intelligent Computing Theories and Applications , pages=

  14. [14]

    Journal of Biosciences , volume=

    Machine-learning techniques for the prediction of protein--protein interactions , author=. Journal of Biosciences , volume=. 2019 , publisher=

    2019

  15. [15]

    Nucleic Acids Research , volume=

    Struct2Net: a web service to predict protein--protein interactions using a structure-based approach , author=. Nucleic Acids Research , volume=. 2010 , publisher=

    2010

  16. [16]

    Nature , volume=

    Structure-based prediction of protein--protein interactions on a genome-wide scale , author=. Nature , volume=. 2012 , publisher=

    2012

  17. [17]

    Nature Communications , volume=

    Improved prediction of protein-protein interactions using AlphaFold2 , author=. Nature Communications , volume=. 2022 , publisher=

    2022

  18. [18]

    Nature , volume=

    Highly accurate protein structure prediction with AlphaFold , author=. Nature , volume=

  19. [19]

    IEEE Transactions on Neural Networks and Learning Systems , volume=

    A comprehensive survey on graph neural networks , author=. IEEE Transactions on Neural Networks and Learning Systems , volume=. 2020 , publisher=

    2020

  20. [20]

    Bioinformatics , volume=

    FIMO: Scanning for occurrences of a given motif , author=. Bioinformatics , volume=. 2011 , publisher=

    2011

  21. [21]

    Nucleic Acids Research , volume=

    MEME SUITE: Tools for motif discovery and searching , author=. Nucleic Acids Research , volume=. 2009 , publisher=

    2009

  22. [22]

    Proceedings of the National Academy of Sciences , volume=

    Statistical significance for genomewide studies , author=. Proceedings of the National Academy of Sciences , volume=. 2003 , publisher=

    2003

  23. [23]

    Learning sequence, structure, and function representations of proteins with language models , author=

  24. [24]

    International Conference on Knowledge Discovery & Data Mining , pages=

    Heterogeneous graph neural network , author=. International Conference on Knowledge Discovery & Data Mining , pages=

  25. [25]

    Nature Reviews Methods Primers , volume=

    Graph neural networks , author=. Nature Reviews Methods Primers , volume=. 2024 , publisher=

    2024

  26. [26]

    International Conference on Learning Representations , year=

    How Powerful are Graph Neural Networks? , author=. International Conference on Learning Representations , year=

  27. [27]

    Journal of Molecular Biology , volume=

    Recognition of spatial motifs in protein structures , author=. Journal of Molecular Biology , volume=. 1999 , publisher=

    1999

  28. [28]

    arXiv preprint arXiv:2501.10282 , year=

    Computational protein science in the era of large language models (LLMs) , author=. arXiv preprint arXiv:2501.10282 , year=

    work page arXiv

  29. [29]

    Briefings in Functional Genomics and Proteomics , volume=

    A review on models and algorithms for motif discovery in protein--protein interaction networks , author=. Briefings in Functional Genomics and Proteomics , volume=. 2008 , publisher=

    2008

  30. [30]

    Current Opinion in Structural Biology , volume=

    Affinity and specificity of motif-based protein--protein interactions , author=. Current Opinion in Structural Biology , volume=. 2019 , publisher=

    2019

  31. [31]

    Cell Communication and Signaling , volume=

    Motif mediated protein-protein interactions as drug targets , author=. Cell Communication and Signaling , volume=. 2016 , publisher=

    2016

  32. [32]

    Bioessays , volume=

    The tetratricopeptide repeat: A structural motif mediating protein-protein interactions , author=. Bioessays , volume=. 1999 , publisher=

    1999

  33. [33]

    Genes & Development , volume=

    The POZ domain: A conserved protein-protein interaction motif , author=. Genes & Development , volume=. 1994 , publisher=

    1994

  34. [34]

    Proceedings of the National Academy of Sciences , volume=

    Biological structure and function emerge from scaling unsupervised learning to 250 million protein sequences , author=. Proceedings of the National Academy of Sciences , volume=. 2021 , publisher=

    2021

  35. [35]

    Science , volume=

    Evolutionary-scale prediction of atomic-level protein structure with a language model , author=. Science , volume=. 2023 , publisher=

    2023

  36. [36]

    Bioinformatics , volume=

    UniRef: comprehensive and non-redundant UniProt reference clusters , author=. Bioinformatics , volume=. 2007 , publisher=

    2007

  37. [37]

    Nucleic Acids Research , volume=

    CATH: Increased structural coverage of functional space , author=. Nucleic Acids Research , volume=. 2021 , publisher=

    2021

  38. [38]

    IEEE Transactions on Pattern Analysis and Machine Intelligence , volume=

    Prottrans: Toward understanding the language of life through self-supervised learning , author=. IEEE Transactions on Pattern Analysis and Machine Intelligence , volume=

  39. [39]

    The Eleventh International Conference on Learning Representations , year=

    Protein Representation Learning by Geometric Structure Pretraining , author=. The Eleventh International Conference on Learning Representations , year=

  40. [40]

    Bioinformatics , volume=

    Structure-aware protein self-supervised learning , author=. Bioinformatics , volume=. 2023 , publisher=

    2023

  41. [41]

    NAR Genomics and Bioinformatics , volume =

    Heinzinger, Michael and Weissenow, Konstantin and Sanchez, Joaquin Gomez and Henkel, Adrian and Mirdita, Milot and Steinegger, Martin and Rost, Burkhard , title =. NAR Genomics and Bioinformatics , volume =. 2024 , month =. doi:10.1093/nargab/lqae150 , url =

    work page doi:10.1093/nargab/lqae150 2024

  42. [42]

    International Conference on Machine Learning , pages=

    Structure-informed language models are protein designers , author=. International Conference on Machine Learning , pages=

  43. [43]

    The Twelfth International Conference on Learning Representations , year=

    SaProt: Protein Language Modeling with Structure-aware Vocabulary , author=. The Twelfth International Conference on Learning Representations , year=

  44. [44]

    International Conference on Learning Representations , year=

    OntoProtein: Protein Pretraining With Gene Ontology Embedding , author=. International Conference on Learning Representations , year=

  45. [45]

    International Conference on Learning Representations , year=

    Protein representation learning via knowledge enhanced primary structure reasoning , author=. International Conference on Learning Representations , year=

  46. [46]

    Xu, Minghao and Yuan, Xinyu and Miret, Santiago and Tang, Jian , booktitle=. Prot

  47. [47]

    Advanced Science , volume=

    A multimodal protein representation framework for quantifying transferability across biochemical downstream tasks , author=. Advanced Science , volume=. 2023 , publisher=

    2023

  48. [48]

    Science , pages=

    Simulating 500 million years of evolution with a language model , author=. Science , pages=. 2025 , publisher=

    2025

  49. [49]

    Journal of Chemical Theory and Computation , volume=

    The application of ligand-mapping molecular dynamics simulations to the rational design of peptidic modulators of protein--protein interactions , author=. Journal of Chemical Theory and Computation , volume=. 2015 , publisher=

    2015

  50. [50]

    Wiley Interdisciplinary Reviews: Computational Molecular Science , volume=

    Computational close up on protein--protein interactions: How to unravel the invisible using molecular dynamics simulations? , author=. Wiley Interdisciplinary Reviews: Computational Molecular Science , volume=. 2015 , publisher=

    2015

  51. [51]

    Current Drug Targets , volume=

    Recent advances in protein-protein docking , author=. Current Drug Targets , volume=. 2016 , publisher=

    2016

  52. [52]

    Biophysical Reviews , pages=

    Protein--protein interaction prediction methods: from docking-based to AI-based approaches , author=. Biophysical Reviews , pages=. 2022 , publisher=

    2022

  53. [53]

    Scientific Reports , volume=

    Predicting protein-protein interactions from matrix-based protein sequence using convolution neural network and feature-selective rotation forest , author=. Scientific Reports , volume=. 2019 , publisher=

    2019

  54. [54]

    Journal of Theoretical Biology , volume=

    Protein--protein interaction inference based on semantic similarity of gene ontology terms , author=. Journal of Theoretical Biology , volume=. 2016 , publisher=

    2016

  55. [55]

    Nucleic Acids Research , volume=

    Prediction of yeast protein--protein interaction network: insights from the Gene Ontology and annotations , author=. Nucleic Acids Research , volume=. 2006 , publisher=

    2006

  56. [56]

    IEEE/ACM Transactions on Computational Biology and Bioinformatics , volume=

    Analyzing effect of multi-modality in predicting protein-protein interactions , author=. IEEE/ACM Transactions on Computational Biology and Bioinformatics , volume=. 2022 , publisher=

    2022

  57. [57]

    2021 International Joint Conference on Neural Networks , pages=

    Prediction of protein-protein interactions using deep multi-modal representations , author=. 2021 International Joint Conference on Neural Networks , pages=

    2021

  58. [58]

    Neurocomputing , volume=

    Sequence-based prediction of protein-protein interaction sites by simplified long short-term memory network , author=. Neurocomputing , volume=. 2019 , publisher=

    2019

  59. [59]

    Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence (IJCAI-21) , pages =

    Learning Unknown from Correlations: Graph Neural Network for Inter-novel-protein Interaction Prediction , author =. Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence (IJCAI-21) , pages =. 2021 , doi =

    2021

  60. [60]

    2023 , isbn =

    Zhao, Ziyuan and Qian, Peisheng and Yang, Xulei and Zeng, Zeng and Guan, Cuntai and Tam, Wai Leong and Li, Xiaoli , title =. 2023 , isbn =. doi:10.24963/ijcai.2023/554 , booktitle =

    work page doi:10.24963/ijcai.2023/554 2023

  61. [61]

    Nature Communications , volume=

    Hierarchical graph learning for protein--protein interaction , author=. Nature Communications , volume=. 2023 , publisher=

    2023

  62. [62]

    Li , booktitle=

    Lirong Wu and Yijun Tian and Yufei Huang and Siyuan Li and Haitao Lin and Nitesh V Chawla and Stan Z. Li , booktitle=. 2024 , url=

    2024

  63. [63]

    Neural Information Processing Systems , year=

    Hierarchical question-image co-attention for visual question answering , author=. Neural Information Processing Systems , year=

  64. [64]

    IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Deep modular co-attention networks for visual question answering , author=. IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  65. [65]

    IEEE Conference on Computer Vision and Pattern Recognition , pages=

    Visual question answering with memory-augmented networks , author=. IEEE Conference on Computer Vision and Pattern Recognition , pages=

  66. [66]

    Artificial Intelligence in Medicine , volume=

    DDI-GCN: drug-drug interaction prediction via explainable graph convolutional networks , author=. Artificial Intelligence in Medicine , volume=. 2023 , publisher=

    2023

  67. [67]

    Nucleic Acids Research , volume=

    STRING v11: Protein--protein association networks with increased coverage, supporting functional discovery in genome-wide experimental datasets , author=. Nucleic Acids Research , volume=. 2019 , publisher=

    2019

  68. [68]

    Bioinformatics , volume=

    Multifaceted protein--protein interaction prediction based on Siamese residual RCNN , author=. Bioinformatics , volume=. 2019 , publisher=

    2019

  69. [69]

    Bioinformatics , volume=

    Predicting protein--protein interactions through sequence-based deep learning , author=. Bioinformatics , volume=. 2018 , publisher=

    2018

  70. [70]

    Molecules , volume=

    Deep neural network based predictions of protein interactions using primary sequences , author=. Molecules , volume=. 2018 , publisher=

    2018

  71. [71]

    Current Opinion in Structural Biology , volume=

    The present and the future of motif-mediated protein--protein interactions , author=. Current Opinion in Structural Biology , volume=. 2018 , publisher=

    2018

  72. [72]

    Cell Communication and Signaling , volume=

    High-throughput methods for identification of protein-protein interactions involving short linear motifs , author=. Cell Communication and Signaling , volume=. 2015 , publisher=

    2015

  73. [73]

    BMC bioinformatics , volume=

    DCMF-PPI: a protein-protein interaction predictor based on dynamic condition and multi-feature fusion , author=. BMC bioinformatics , volume=. 2025 , publisher=

    2025

  74. [74]

    BMC biology , volume=

    Prediction of protein--protein interaction based on interaction-specific learning and hierarchical information , author=. BMC biology , volume=. 2025 , publisher=

    2025

  75. [75]

    Journal of Chemical Information and Modeling , volume=

    MicroEnvPPI: Microenvironment-Aware Optimization Enables Generalizable Protein--Protein Interaction Prediction , author=. Journal of Chemical Information and Modeling , volume=. 2025 , publisher=

    2025

  76. [76]

    IEEE Transactions on Neural Networks and Learning Systems , year=

    Accurate Protein--Protein Interaction Prediction: Based on Multiview Heterogeneous Graph Autoencoders and Random Masking , author=. IEEE Transactions on Neural Networks and Learning Systems , year=

  77. [77]

    Nucleic acids research , volume=

    AlphaFold Protein Structure Database: massively expanding the structural coverage of protein-sequence space with high-accuracy models , author=. Nucleic acids research , volume=. 2022 , publisher=

    2022

  78. [78]

    ICLR 2024 Workshop on Generative and Experimental Perspectives for Biomolecular Design , year=

    Revealing data leakage in protein interaction benchmarks , author=. ICLR 2024 Workshop on Generative and Experimental Perspectives for Biomolecular Design , year=

    2024

  79. [79]

    Nature , volume=

    Highly accurate protein structure prediction for the human proteome , author=. Nature , volume=. 2021 , publisher=

    2021

  80. [80]

    Nature methods , volume=

    Flaws in evaluation schemes for pair-input computational predictions , author=. Nature methods , volume=. 2012 , publisher=

    2012

Showing first 80 references.