Pith. sign in

REVIEW 4 major objections 5 minor 49 references

KEPLA: A Knowledge-Enhanced Deep Learning Framework for Accurate Protein-Ligand Binding Affinity Prediction

T0 review · 4 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read KEPLA claims that an interaction-free model jointly trained on a biochemical knowledge graph and affinity regression outperforms 3D-structure-based state-of-the-art methods on protein–ligand binding benchmarks, with RMSE gains of 5.28% on…

desk verdict A credible in-domain win for knowledge-enhanced affinity prediction, but the cross-domain claims need a leakage check and numeric tables before they can be trusted. read the letter →

arxiv 2506.13196 v6 pith:FD2TOQ5P submitted 2025-06-16 cs.LG

Han Liu , Keyan Ding , Peilin Chen , Yinwei Wei , Liqiang Nie , Dapeng Wu , Shiqi Wang This is my paper · ORCID
classification cs.LG
keywords protein-ligandbindingaffinityknowledgegraphembeddingGeneOntologycrossattentioninteraction-freepredictionPDBbindCSAR-HiQdrugdiscovery
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

The paper sets out to show that protein–ligand binding affinity can be predicted more accurately and more interpretably when an interaction-free deep model is trained not only on sequence and molecular-graph structure but also on a small biochemical knowledge graph. KEPLA builds that graph from Gene Ontology annotations for proteins and from computed ligand properties (molecular descriptors and chemical features), then ties the protein and ligand encoders to it through a joint knowledge-graph embedding objective while a cross-attention module forms the final affinity prediction. On PDBbind core and CSAR-HiQ, the authors report consistent improvements over state-of-the-art baselines, including 3D-structure-based methods, with RMSE reductions of 5.28% and 12.42% relative to the strongest baseline (GIANT), and further gains in cross-domain and cold-start splits. If this holds, sequence-plus-graph models could compete with structure-based scoring while remaining applicable to targets without known complex structures, and the knowledge-graph relations plus attention maps would give chemists grounded explanations instead of a black-box score.

What carries the argument

Two coupled mechanisms carry the argument. First, a joint objective $L = L_{\mathrm{PLA}} + \beta L_{\mathrm{KGE}} + \lambda \|\Theta\|_2^2$ makes the same protein and ligand encoders serve both affinity regression and knowledge-graph embedding: global average-pooled encoder outputs are linearly projected and scored as KG entities, with RotatE $\|\mathbf{h} \circ \mathbf{r} - \mathbf{t}\|$ for the protein–GO subgraph and TransE $\|\mathbf{h} + \mathbf{r} - \mathbf{t}\|$ for the ligand–property subgraph. Second, a cross-attention interaction map $\mathbf{V} = H_p^\top H_d$ between protein fragments and ligand atoms produces softmax-normalized attention weights over both sides, and the weighted fragment and atom sums are concatenated and decoded by an MLP. This design lets the KG objective shape global representations while the attention objective preserves fine-grained local contact information; the ablation study attributes the reported gains to exactly this combination.

What would settle it

Re-run the cross-domain and cold splits with the knowledge graph built only from source-domain (training) entities, or with every triplet removed that involves a test protein or ligand; if the RMSE advantage over DrugBAN collapses, knowledge-graph leakage rather than knowledge-enhanced representations explains the reported generalization.

Watch

Extended reading notes

Core claim

KEPLA's central claim is that explicit biochemical knowledge, injected during encoding rather than bolted on afterwards, is what lets an interaction-free model surpass interaction-based state of the art; the paper presents this as the first general interaction-free framework to integrate such knowledge deeply. The paper constructs two knowledge graphs—protein–GO triplets and ligand–property triplets—from PDBbind refined-set entities, and uses the encoders' pooled global representations as entity embeddings optimized with RotatE for proteins and TransE for ligands. Simultaneously, local fragment-level representations are combined by cross attention into a joint vector decoded by an MLP into the affinity, and the two losses are minimized together. The authors report that KEPLA improves RMSE by 5.28% on the PDBbind core set and 12.42% on CSAR-HiQ against the best baseline (GIANT), outperforms interaction-free baselines in clustering-based and cold splits, and produces attention- and knowledge-based explanations that align with crystallographic binding-site residues for the three examined complexes.

Load-bearing premise

The load-bearing premise is that no test protein or ligand appears in the knowledge-graph triplets used during training; if a test entity is in the graph, the reported cross-domain and cold-start gains could come from reading entity identity off the knowledge graph rather than from generalizing.

Editorial extensions

If this is right

  • Affinity prediction can be done from a protein sequence and a ligand's 2D graph with accuracy competitive with 3D-structure-based models, so targets without resolved complex structures remain addressable.
  • Gene Ontology terms and simple ligand descriptors such as hydrogen-bond donor and acceptor counts, aromatic rings, and rotatable bonds carry predictive signal beyond ESM and GCN features, making external annotations a cheap and reusable accuracy source.
  • The joint training recipe—shared encoders, a knowledge-graph embedding objective, and a cross-attention decoder—can be transferred to other molecular interaction tasks, such as drug–drug or protein–protein interaction prediction, by swapping in the appropriate knowledge graph.
  • Knowledge-graph triplet scores and cross-attention maps give a route to explain predictions at the level of GO functions and binding-site residues or atoms, which is useful for prioritizing which interactions to verify experimentally.

Reading between the lines

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

  • If the leak-free condition holds, the cold-start gains imply that functional annotations, rather than protein identity, carry much of the predictive signal; a direct test is to compare affinity accuracy for test proteins that do versus do not have GO terms in the training knowledge graph.
  • The mechanism does not depend on Gene Ontology specifically; replacing it with pathway or disease annotations, or adding them, would test whether the benefit comes from functional semantics or from any relational side information.
  • The attention-map interpretability could become a quantitative benchmark by measuring agreement between top-attention protein fragments and crystallographic contact residues across the whole core set, not only the three showcased complexes.
  • Because the knowledge graph contains roughly 60,000 triplets, a scaled-up version with more GO evidence codes or ligand property types would test whether the reported gains grow or saturate with knowledge coverage.
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 / 5 minor

Summary. The paper proposes KEPLA, an interaction-free deep learning framework for protein-ligand binding affinity prediction. KEPLA encodes protein sequences with a pretrained ESM model and ligand molecular graphs with a GCN, and jointly optimizes a knowledge-graph embedding loss over Gene Ontology and ligand-property triplets together with a cross-attention-based affinity regression loss. Experiments are reported on PDBbind and CSAR-HiQ under random, clustering-based cross-domain, and cold splits, with ablations and interpretability case studies. The central claim is that KEPLA consistently outperforms state-of-the-art baselines, including interaction-based methods, with RMSE improvements of 5.28% on PDBbind core and 12.42% on CSAR-HiQ.

Significance. If the results hold, KEPLA would be a useful contribution to interaction-free affinity prediction: it combines an established protein language model with a structured knowledge graph in a multi-task objective, provides code and a constructed KG dataset, and reports five-run statistics with ablations. The interpretability analyses are also a strength, although they are post-hoc rather than independent validation. The main value of the paper depends on whether the cross-domain and cold-start evaluations are leak-free and whether the comparison to baselines is controlled for the large pretrained encoder; these points are currently not established.

major comments (4)
  1. [Section IV-A, Section III-D, Section III-E, Section V-B] The knowledge graph is built from 'all proteins and ligands in the refined set of PDBbind' (Section IV-A), while the cross-domain clustering split and cold split in Section V-B are defined on the same PDBbind data. Section III-D defines the KG loss as an average over 'the set of all KG triplets' S (Eq. 8), but Section III-E says that for each mini-batch 'we select the associated KG triplets.' The paper never states that triplets whose head is a test-fold protein or ligand are excluded, nor that the KG is rebuilt per fold. If test entities appear as heads in triplets scored during training, their global representations receive entity-specific supervisory signal from GO/LP tails, which would violate the claim that test proteins and ligands are unseen in the cold setting and would inflate cross-domain performance. Please specify the exact triplet sampling procedure and verify experimentally that no test-entity triplets contribute to training, for example by reporting the intersection between KG head entities and test entity sets for each split.
  2. [Section IV-D and Table II] KEPLA uses ESM2_t36_3B, a 3-billion-parameter pretrained protein language model, as its protein encoder, while the interaction-free baselines in Table II (DeepDTA, GraphDTA, DrugBAN) use much smaller CNN/GCN encoders trained from scratch. The reported improvements of 5.28% on PDBbind core and 12.42% on CSAR-HiQ over GIANT could therefore be due to the capacity of the pretrained backbone rather than the proposed KG and cross-attention modules. The paper should include a controlled comparison where a baseline or an ablated KEPLA variant uses the same ESM backbone, or at least report results with a smaller protein encoder, to isolate the contribution of the proposed components.
  3. [Section V-B and Figure 2] Cross-domain and cold-split results are presented only as box plots without a numeric table, making the central cross-domain claims unverifiable. The text states improvements of 12.25% in RMSE and 8.37% in R over DrugBAN on the PDBbind dataset, but these numbers cannot be checked from the figure. Please provide a table analogous to Table II with means and standard deviations for all metrics under both the clustering-based and cold pair splits, along with the number of test pairs and, ideally, statistical significance tests across the five runs.
  4. [Section V-B] The cold pair split description is ambiguous about how pairs are formed after selecting 70% of proteins and ligands. If 70% of protein entities and 70% of ligand entities are selected independently, it is unclear which pairs constitute the training set and whether pairs with one entity in the selected set and the other outside are discarded. This ambiguity affects the validity of the statement that neither proteins nor ligands in the test set are observed during training. Please clarify the exact pair-generation procedure for the cold split.
minor comments (5)
  1. [Table II] The KEPLA row in Table II appears to repeat each value three times (e.g., '1.202 (0.017)1.202 (0.017)1.202 (0.017)'), which makes the table difficult to read and the reported numbers ambiguous; please fix the formatting.
  2. [Section V-E] The knowledge-level interpretability is generated by applying the same KG score function used during training, so it is not an independent validation of the retrieved knowledge. This is acceptable as a post-hoc explanation, but the paper should describe it as such rather than as confirmatory evidence.
  3. [Abstract and Section I] The claim of being the 'first general framework' for deep integration of biochemical knowledge into PLA prediction is not consistent with the cited works [6] and [32], which also integrate knowledge with sequence and structural features; please temper or qualify the novelty claim.
  4. [Section IV-A] The phrase 'nanoscale affinity' in Section V-D should be 'nanomolar affinity'; the current wording is imprecise.
  5. [References] Reference [35] (RotatE) is missing its publication year; please add the complete citation information.

Circularity Check

1 steps flagged · score 2.0 of 10

Knowledge-level interpretability recovers the KG triplets used as training targets; the affinity prediction itself is not circular.

  1. other [Section V-E, Knowledge-Level Interpretability with Knowledge Graph, Figure 5]
    "Specifically, we identify entities most closely related to the target protein or ligand by applying the KG scoring function to all associated triplets. Lower scores indicate stronger semantic relevance."

    The scoring function used here is the same F minimized in Eq. (8) of Section III-D, and the head embeddings are the encoder outputs trained on the triplet set S. For any entity present in S, the triplets involving that entity were training targets, so ranking those triplets by F will return the entity's own GO/LP annotations at the top by construction. The knowledge-level interpretability therefore reads out the KG supervision used to train the encoder, rather than providing an independent explanation of the affinity prediction.

full rationale

The central affinity-prediction claim is not circular. The ground-truth affinity y_{p,d} used in Eq. (16) is never an input to the KG triplet construction; the KG is built from GO annotations and RDKit-computed ligand properties (Section IV-A), which are independent of the binding-affinity labels. The KG objective in Eq. (8) is an auxiliary regularizer that aligns encoder outputs with these external annotations, and the affinity decoder still learns a nontrivial mapping from attention-pooled representations to the regression target. The reported performance gains are empirical and externally falsifiable on held-out splits. The one self-referential step is the knowledge-level interpretability in Section V-E, which uses the same score function minimized during training to retrieve top GO/LP entities; for entities already in the KG, this is a tautological readout of training targets rather than an independent validation. Separately, the manuscript leaves an ambiguity: the KG is described as covering all proteins and ligands in the refined set (Section IV-A), while cross-domain and cold splits are defined on PDBbind data (Section V-B), without stating whether test-entity triplets are excluded from S during training. If they are not excluded, the unseen-entity claim would be violated and cross-domain results could be inflated by transductive KG supervision. That is a data-contamination and correctness risk, not a circularity of the derivation chain; it does not make the affinity prediction equivalent to the KG inputs.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The model contributes no new physical entities. Its central claims rest on hyperparameters, domain assumptions about GO and ligand-property informativeness, and the unstated split-leakage assumption.

free parameters (5)
  • beta (KG loss trade-off) = 0.1
    Chosen by hand to balance the KG embedding and PLA losses in Eq. 17; no sensitivity analysis is reported.
  • Average pooling window size s = 9
    Controls the granularity of protein local fragments; chosen by hand in Section IV-D.
  • Clustering thresholds gamma_protein and gamma_ligand = 0.001 / 0.5
    Used to construct disjoint source and target domains in the cross-domain split; the values are choices that directly affect the difficulty of the evaluation.
  • L2 regularization coefficient lambda = not reported
    Appears in the total loss Eq. 17 but no value is given, which limits exact reproduction.
  • Top-1000 GO terms retained = 1000
    The KG keeps only the 1,000 most frequent GO terms, a truncation that defines the KG entities in Section IV-A.
assumptions (4)
  • domain assumption ESM embeddings encode protein function and binding-relevant semantics.
    The protein encoder relies on pretrained ESM2_t36_3B representations to support the KG alignment objective (Section III-B).
  • domain assumption GO annotations and RDKit ligand property descriptors are accurate and complete enough to define useful KG triplets.
    KG construction in Section IV-A retrieves GO terms via UniProt and ligand properties via RDKit; missing or noisy annotations are not modeled.
  • domain assumption Cross attention weights correspond to physical binding interactions.
    The interpretability analysis in Section V-D treats attention weights as indicators of binding-site residues and atoms, but no quantitative validation is provided.
  • ad hoc to paper Test protein and ligand entities are absent from KG triplets during training in cross-domain and cold splits.
    The paper does not state that the KG is rebuilt per fold or that test entities are removed; this unstated assumption is required for the cold-start generalization claims.

how reviews work

0 comments
Cite this review

Pith. "Pith review of KEPLA: A Knowledge-Enhanced Deep Learning Framework for Accurate Protein-Ligand Binding Affinity Prediction." pith.science (2026). https://pith.science/paper/FD2TOQ5P

@misc{pith2026250613196,
  author       = {Pith},
  title        = {Pith review of: KEPLA: A Knowledge-Enhanced Deep Learning Framework for Accurate Protein-Ligand Binding Affinity Prediction},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FD2TOQ5P}},
  note         = {Machine review of arXiv:2506.13196}
}
read the original abstract

Accurate prediction of protein-ligand binding affinity is critical for drug discovery. While recent deep learning approaches have demonstrated promising results, they often rely solely on structural features of proteins and ligands, overlooking their valuable biochemical knowledge associated with binding affinity. To address this limitation, we propose KEPLA, a novel deep learning framework that explicitly integrates prior knowledge from Gene Ontology and ligand properties to enhance prediction performance. KEPLA takes protein sequences and ligand molecular graphs as input and optimizes two complementary objectives: (1) aligning global representations with knowledge graph relations to capture domain-specific biochemical insights, and (2) leveraging cross attention between local representations to construct fine-grained joint embeddings for prediction. Experiments on two benchmark datasets across both in-domain and cross-domain scenarios demonstrate that KEPLA consistently outperforms state-of-the-art baselines. Furthermore, interpretability analyses based on knowledge graph relations and cross attention maps provide valuable insights into the underlying predictive mechanisms.

Figures

Figures reproduced from arXiv: 2506.13196 by the authors.

Figure 1
Figure 1. Overview of the KEPLA framework. (a) Existing PLA prediction strategy. (b) KEPLA strategy, which incorporates a biochemical KG to enhance protein and ligand encoding. (c) KEPLA framework with two encoders: an ESM-based encoder for protein sequences and a GCN-based encoder for ligand SMILES, producing local and global representations. Global representations are used for KG embedding, while local representations are f… view at source ↗
Figure 2
Figure 2. Cross-domain performance comparison on the PDBbind dataset (statistics over five independent runs). (a) Performance comparison of interaction￾free methods under clustering-based pair split. (b) Performance comparison of interaction-free methods under cold pair split. The box plots display the median as the center line and the mean as a square marker. The minimum and lower percentile indicate the worst and second-wor… view at source ↗
Figure 3
Figure 3. Ablation study in terms of RMSE and R on the PDBbind and CSAR-HiQ datasets under random split (statistics over five independent runs). (a) Performance of different variants of the KG module. (b) Performance of different variants of the cross attention module. The vertical bars represent the mean values, while the black lines denote standard deviations (error bars). Dots indicate the performance scores from each indi… view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Visualization of ligands and binding pockets for structural-level interpretability study. (a) Interpretability of co-crystallized ligands. The upper section of each panel displays the 2D structures of ligands, with atoms highlighted in orange to indicate those predicte…
Figure 5
Figure 5. Figure 5: Biochemical knowledge of proteins and ligands for knowledge-level interpretability study. (a) Interpretability of protein biological knowledge. The upper part of each panel displays the GO entities most closely associated with the protein, ranked in ascending order acc…
Figure 6
Figure 6. Figure 6: Performance improvements on PDBbind benchmark when trained on the general set. (a) RMSE improvements on PDBbind core set. (b) MAE improvements on PDBbind core set. Additionally, expanding our knowledge graph with more com￾prehensive domain-specific and cross-disciplina…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

49 extracted references · 46 canonical work pages

  1. [1]

    On the frustration to predict binding affinities from protein–ligand structures with deep neural networks,

    M. V olkov, J.-A. Turk, N. Drizard, N. Martin, B. Hoffmann, Y . Gaston- Mathé, and D. Rognan, “On the frustration to predict binding affinities from protein–ligand structures with deep neural networks,”Journal of Medicinal Chemistry, vol. 65, no. 11, pp. 7946–7958, 2022

    work page 2022

  2. [2]

    Computational approaches stream- lining drug discovery,

    A. V . Sadybekov and V . Katritch, “Computational approaches stream- lining drug discovery,”Nature, vol. 616, no. 7958, pp. 673–685, 2023

    2023

  3. [3]

    A network integration approach for drug-target interaction prediction and computational drug repositioning from heterogeneous information,

    Y . Luo, X. Zhao, J. Zhou, J. Yang, Y . Zhang, W. Kuang, J. Peng, L. Chen, and J. Zeng, “A network integration approach for drug-target interaction prediction and computational drug repositioning from heterogeneous information,”Nature Communications, vol. 8, no. 1, p. 573, 2017

    work page 2017

  4. [4]

    Deep learning in drug design: protein-ligand binding affinity prediction,

    M. A. Rezaei, Y . Li, D. Wu, X. Li, and C. Li, “Deep learning in drug design: protein-ligand binding affinity prediction,”IEEE/ACM Transactions on Computational Biology and Bioinformatics, vol. 19, no. 1, pp. 407–417, 2020

    work page 2020

  5. [5]

    Dyscore: a boosting scoring method with dynamic properties for identifying true binders and nonbinders in structure-based drug discovery,

    Y . Li, D. Zhou, G. Zheng, X. Li, D. Wu, and Y . Yuan, “Dyscore: a boosting scoring method with dynamic properties for identifying true binders and nonbinders in structure-based drug discovery,”Journal of Chemical Information and Modeling, vol. 62, no. 22, pp. 5550–5567, 2022

    work page 2022

  6. [6]

    Knowledge-enhanced and structure- enhanced representation learning for protein-ligand binding affinity prediction,

    M. Li, Y . Cao, X. Liu, and H. Ji, “Knowledge-enhanced and structure- enhanced representation learning for protein-ligand binding affinity prediction,”Pattern Recognition, vol. 166, p. 111701, 2025

    work page 2025

  7. [7]

    Interformer: an interaction-aware model for protein-ligand docking and affinity prediction,

    H. Lai, L. Wang, R. Qian, J. Huang, P. Zhou, G. Ye, F. Wu, F. Wu, X. Zeng, and W. Liu, “Interformer: an interaction-aware model for protein-ligand docking and affinity prediction,”Nature Communications, vol. 15, no. 1, p. 10223, 2024. 14

    work page 2024

  8. [8]

    Smiles, a chemical language and information system. 1. introduction to methodology and encoding rules,

    D. Weininger, “Smiles, a chemical language and information system. 1. introduction to methodology and encoding rules,”Journal of Chemical Information and Computer Sciences, vol. 28, no. 1, pp. 31–36, 1988

    work page 1988

Show all 49 references
  1. [9]

    Deepdta: deep drug–target binding affinity prediction,

    H. Öztürk, A. Özgür, and E. Ozkirimli, “Deepdta: deep drug–target binding affinity prediction,”Bioinformatics, vol. 34, no. 17, pp. i821– i829, 2018

    2018

  2. [10]

    Deepconv-dti: Prediction of drug-target interactions via deep learning with convolution on protein sequences,

    I. Lee, J. Keum, and H. Nam, “Deepconv-dti: Prediction of drug-target interactions via deep learning with convolution on protein sequences,” PLoS Computational Biology, vol. 15, no. 6, pp. 1–21, 2019

    2019

  3. [11]

    Drug–target affinity prediction using graph neural network and contact maps,

    M. Jiang, Z. Li, S. Zhang, S. Wang, X. Wang, Q. Yuan, and Z. Wei, “Drug–target affinity prediction using graph neural network and contact maps,”RSC advances, vol. 10, no. 35, pp. 20 701–20 712, 2020

    2020

  4. [12]

    Mgraphdta: deep multiscale graph neural network for explainable drug–target binding affinity prediction,

    Z. Yang, W. Zhong, L. Zhao, and C. Y .-C. Chen, “Mgraphdta: deep multiscale graph neural network for explainable drug–target binding affinity prediction,”Chemical Science, vol. 13, no. 3, pp. 816–833, 2022

    2022

  5. [13]

    Attention is all you need,

    A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, Ł. Kaiser, and I. Polosukhin, “Attention is all you need,”International Conference on Neural Information Processing Systems, pp. 1–11, 2017

    2017

  6. [14]

    Transformercpi: improving compound– protein interaction prediction by sequence-based deep learning with self- attention mechanism and label reversal experiments,

    L. Chen, X. Tan, D. Wang, F. Zhong, X. Liu, T. Yang, X. Luo, K. Chen, H. Jiang, and M. Zheng, “Transformercpi: improving compound– protein interaction prediction by sequence-based deep learning with self- attention mechanism and label reversal experiments,”Bioinformatics, vol....

    2020

  7. [15]

    Moltrans: molecular interaction transformer for drug–target interaction prediction,

    K. Huang, C. Xiao, L. M. Glass, and J. Sun, “Moltrans: molecular interaction transformer for drug–target interaction prediction,”Bioinfor- matics, vol. 37, no. 6, pp. 830–836, 2021

    2021

  8. [16]

    Gene ontology: tool for the unification of biology,

    M. Ashburner, C. A. Ball, J. A. Blake, D. Botstein, H. Butler, J. M. Cherry, A. P. Davis, K. Dolinski, S. S. Dwight, J. T. Eppiget al., “Gene ontology: tool for the unification of biology,”Nature Genetics, vol. 25, no. 1, pp. 25–29, 2000

    2000

  9. [17]

    A unified drug–target interaction prediction framework based on knowledge graph and recommendation system,

    Q. Ye, C.-Y . Hsieh, Z. Yang, Y . Kang, J. Chen, D. Cao, S. He, and T. Hou, “A unified drug–target interaction prediction framework based on knowledge graph and recommendation system,”Nature Communica- tions, vol. 12, no. 1, p. 6775, 2021

    2021

  10. [18]

    Knowledge graph embed- ding: A survey of approaches and applications,

    Q. Wang, Z. Mao, B. Wang, and L. Guo, “Knowledge graph embed- ding: A survey of approaches and applications,”IEEE Transactions on Knowledge and Data Engineering, vol. 29, no. 12, pp. 2724–2743, 2017

    2017

  11. [19]

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

    Z. Lin, H. Akin, R. Rao, B. Hie, Z. Zhu, W. Lu, N. Smetanin, R. Verkuil, O. Kabeli, Y . Shmueliet al., “Evolutionary-scale prediction of atomic- level protein structure with a language model,”Science, vol. 379, no. 6637, pp. 1123–1130, 2023

    2023

  12. [20]

    Learning characteristics of graph neural networks predicting protein–ligand affinities,

    A. Mastropietro, G. Pasculli, and J. Bajorath, “Learning characteristics of graph neural networks predicting protein–ligand affinities,”Nature Machine Intelligence, vol. 5, no. 12, pp. 1427–1436, 2023

    2023

  13. [21]

    Cross attention network for few-shot classification,

    R. Hou, H. Chang, B. Ma, S. Shan, and X. Chen, “Cross attention network for few-shot classification,”International Conference on Neural Information Processing Systems, pp. 4003–4014, 2019

    2019

  14. [22]

    A machine learning approach to predicting protein–ligand binding affinity with applications to molecular docking,

    P. J. Ballester and J. B. Mitchell, “A machine learning approach to predicting protein–ligand binding affinity with applications to molecular docking,”Bioinformatics, vol. 26, no. 9, pp. 1169–1175, 2010

    2010

  15. [23]

    Graphdta: predicting drug–target binding affinity with graph neural networks,

    T. Nguyen, H. Le, T. P. Quinn, T. Nguyen, T. D. Le, and S. Venkatesh, “Graphdta: predicting drug–target binding affinity with graph neural networks,”Bioinformatics, vol. 37, no. 8, pp. 1140–1147, 2021

    2021

  16. [24]

    Interpretable bilinear atten- tion network with domain adaptation improves drug–target prediction,

    P. Bai, F. Miljkovi ´c, B. John, and H. Lu, “Interpretable bilinear atten- tion network with domain adaptation improves drug–target prediction,” Nature Machine Intelligence, vol. 5, no. 2, pp. 126–136, 2023

    2023

  17. [25]

    Onionnet: a multiple-layer intermolecular- contact-based convolutional neural network for protein–ligand binding affinity prediction,

    L. Zheng, J. Fan, and Y . Mu, “Onionnet: a multiple-layer intermolecular- contact-based convolutional neural network for protein–ligand binding affinity prediction,”ACS Omega, vol. 4, no. 14, pp. 15 956–15 965, 2019

    2019

  18. [26]

    Predicting drug–target interaction using a novel graph neural network with 3d structure-embedded graph representation,

    J. Lim, S. Ryu, K. Park, Y . J. Choe, J. Ham, and W. Y . Kim, “Predicting drug–target interaction using a novel graph neural network with 3d structure-embedded graph representation,”Journal of Chemical Information and Modeling, vol. 59, no. 9, pp. 3981–3988, 2019

    2019

  19. [27]

    Analyzing learned molecular representations for property prediction,

    K. Yang, K. Swanson, W. Jin, C. Coley, P. Eiden, H. Gao, A. Guzman- Perez, T. Hopper, B. Kelley, M. Matheaet al., “Analyzing learned molecular representations for property prediction,”Journal of Chemical Information and Modeling, vol. 59, no. 8, pp. 3370–3388, 2019

    2019

  20. [28]

    Development and evaluation of a deep learning model for protein– ligand binding affinity prediction,

    M. M. Stepniewska-Dziubinska, P. Zielenkiewicz, and P. Siedlecki, “Development and evaluation of a deep learning model for protein– ligand binding affinity prediction,”Bioinformatics, vol. 34, no. 21, pp. 3666–3674, 2018

    2018

  21. [29]

    Structure-aware interactive graph neural networks for the prediction of protein-ligand binding affinity,

    S. Li, J. Zhou, T. Xu, L. Huang, F. Wang, H. Xiong, W. Huang, D. Dou, and H. Xiong, “Structure-aware interactive graph neural networks for the prediction of protein-ligand binding affinity,”ACM SIGKDD Conference on Knowledge Discovery and Data Mining, pp. 975–985, 2021

    2021

  22. [30]

    Giant: Protein-ligand binding affinity prediction via geometry- aware interactive graph neural network,

    ——, “Giant: Protein-ligand binding affinity prediction via geometry- aware interactive graph neural network,”IEEE Transactions on Knowl- edge and Data Engineering, vol. 36, no. 5, pp. 1991–2008, 2024

    1991

  23. [31]

    Biomedical knowledge graph embedding with capsule network for multi-label drug-drug interaction prediction,

    X. Su, Z. You, D. Huang, L. Wang, L. Wong, B. Ji, and B. Zhao, “Biomedical knowledge graph embedding with capsule network for multi-label drug-drug interaction prediction,”IEEE Transactions on Knowledge and Data Engineering, vol. 35, no. 6, pp. 5640–5651, 2022

    2022

  24. [32]

    Kg-mtl: knowledge graph enhanced multi-task learning for molecular interaction,

    T. Ma, X. Lin, B. Song, P. S. Yu, and X. Zeng, “Kg-mtl: knowledge graph enhanced multi-task learning for molecular interaction,”IEEE Transactions on Knowledge and Data Engineering, vol. 35, no. 7, pp. 7068–7081, 2022

    2022

  25. [33]

    Dgl-lifesci: An open-source toolkit for deep learning on graphs in life science,

    M. Li, J. Zhou, J. Hu, W. Fan, Y . Zhang, Y . Gu, and G. Karypis, “Dgl-lifesci: An open-source toolkit for deep learning on graphs in life science,”ACS Omega, vol. 6, no. 41, pp. 27 233–27 238, 2021

    2021

  26. [34]

    Translating embeddings for modeling multi-relational data,

    A. Bordes, N. Usunier, A. Garcia-Durán, J. Weston, and O. Yakhnenko, “Translating embeddings for modeling multi-relational data,”Interna- tional Conference on Neural Information Processing Systems, pp. 2787– 2795, 2013

    2013

  27. [35]

    Rotate: Knowledge graph embedding by relational rotation in complex space,

    Z. Sun, Z.-H. Deng, J.-Y . Nie, and J. Tang, “Rotate: Knowledge graph embedding by relational rotation in complex space,”International Conference on Learning Representations

  28. [36]

    Adam: A method for stochastic optimization,

    D. P. Kingma and J. Ba, “Adam: A method for stochastic optimization,” International Conference on Learning Representations, 2015

    2015

  29. [37]

    Pdb-wide collection of binding data: current status of the pdbbind database,

    Z. Liu, Y . Li, L. Han, J. Li, J. Liu, Z. Zhao, W. Nie, Y . Liu, and R. Wang, “Pdb-wide collection of binding data: current status of the pdbbind database,”Bioinformatics, vol. 31, no. 3, pp. 405–412, 2015

    2015

  30. [38]

    Csar benchmark exercise of 2010: combined evaluation across all submitted scoring functions,

    R. D. Smith, J. B. Dunbar Jr, P. M.-U. Ung, E. X. Esposito, C.-Y . Yang, S. Wang, and H. A. Carlson, “Csar benchmark exercise of 2010: combined evaluation across all submitted scoring functions,”Journal of Chemical Information and Modeling, vol. 51, no. 9, pp. 2115–2131, 2011

    2010

  31. [39]

    Spatial graph convolutional networks,

    T. Danel, P. Spurek, J. Tabor, M. ´Smieja, Ł. Struski, A. Słowik, and Ł. Maziarka, “Spatial graph convolutional networks,”International Conference on Neural Information Processing, pp. 668–675, 2020

    2020

  32. [40]

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

    Z. Wu, P. Jain, M. Wright, A. Mirhoseini, J. E. Gonzalez, and I. Stoica, “Representing long-range context for graph neural networks with global attention,”International Conference on Neural Information Processing Systems, vol. 34, pp. 13 266–13 279, 2021

    2021

  33. [41]

    Non-local graph neural networks,

    M. Liu, Z. Wang, and S. Ji, “Non-local graph neural networks,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 44, no. 12, pp. 10 270–10 276, 2021

    2021

  34. [42]

    Molecule attention transformer,

    Ł. Maziarka, T. Danel, S. Mucha, K. Rataj, J. Tabor, and S. Jastrz˛ eb- ski, “Molecule attention transformer,”arXiv preprint arXiv:2002.08264, 2020

    2002 arXiv

  35. [43]

    Directional message passing for molecular graphs,

    J. Gasteiger, J. Groß, and S. Günnemann, “Directional message passing for molecular graphs,”International Conference on Learning Represen- tations, 2020

    2020

  36. [44]

    Com- municative representation learning on attributed molecular graphs,

    Y . Song, S. Zheng, Z. Niu, Z.-H. Fu, Y . Lu, and Y . Yang, “Com- municative representation learning on attributed molecular graphs,” International Joint Conference on Artificial Intelligence, pp. 2831–2838, 2020

    2020

  37. [45]

    Compound–protein interaction prediction with end-to-end learning of neural networks for graphs and sequences,

    M. Tsubaki, K. Tomii, and J. Sese, “Compound–protein interaction prediction with end-to-end learning of neural networks for graphs and sequences,”Bioinformatics, vol. 35, no. 2, pp. 309–318, 2019

    2019

  38. [46]

    The protein data bank,

    H. M. Berman, J. Westbrook, Z. Feng, G. Gilliland, T. N. Bhat, H. Weissig, I. N. Shindyalov, and P. E. Bourne, “The protein data bank,” Nucleic Acids Research, vol. 28, no. 1, pp. 235–242, 2000

    2000

  39. [47]

    Protein x-ray crystallography and drug discovery,

    L. Maveyraud and L. Mourey, “Protein x-ray crystallography and drug discovery,”Molecules, vol. 25, no. 5, p. 1030, 2020

    2020

  40. [48]

    Accurate structure prediction of biomolecular interactions with alphafold 3,

    J. Abramson, J. Adler, J. Dunger, R. Evans, T. Green, A. Pritzel, O. Ronneberger, L. Willmore, A. J. Ballard, J. Bambricket al., “Accurate structure prediction of biomolecular interactions with alphafold 3,” Nature, vol. 630, no. 8016, pp. 493–500, 2024

    2024

  41. [49]

    Survey: Functional module detection from protein-protein interaction networks,

    J. Ji, A. Zhang, C. Liu, X. Quan, and Z. Liu, “Survey: Functional module detection from protein-protein interaction networks,”IEEE Transactions on Knowledge and Data Engineering, vol. 26, no. 2, pp. 261–277, 2012

    2012

Pith tools

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