CoChat
Sign Up
PreprintCited by 13Computational Drug Discovery MethodsMachine Learning in BioinformaticsProtein Degradation and Inhibitors

The DOI matches its Crossref record; 1 of the 1 references it cites were found in a registry or an index.

Escaping the drug-bias trap: using debiasing design to improve interpretability and generalization of drug-target interaction prediction

Open access
Zhang, Pei-Dong, Ma, Jianzhu, Chen, Ting·2024·IEEE Transactions on Computational Biology and Bioinformatics
DOI 10.1101/2024.09.12.612771

Preprint — published as 10.1109/tcbbio.2025.3576488 · Paper page

Open this paper in CoChat — free account

Abstract

Considering the high cost associated with determining reaction affinities through in-vitro experiments, virtual screening of potential drugs bound with specific protein pockets from vast compounds is critical in AI-assisted drug discovery. Deep-leaning approaches have been proposed for Drug-Target Interaction (DTI) prediction. However, they have shown overestimated accuracy because of the drug-bias trap, a challenge that results from excessive reliance on the drug branch in the traditional drug-protein dual-branch network approach. This casts doubt on the interpretability and generalizability of existing Drug-Target Interaction (DTI) models. Therefore, we introduce UdanDTI, an innovative deep-learning architecture designed specifically for predicting drug-protein interactions. UdanDTI applies an unbalanced dual-branch system and an attentive aggregation module to enhance interpretability from a biological perspective. Across various public datasets, UdanDTI demonstrates outstanding performance, outperforming state-of-the-art models under in-domain, cross-domain, and structural interpretability settings. Notably, it demonstrates exceptional accuracy in predicting drug responses of two crucial subgroups of Epidermal Growth Factor Receptor (EGFR) mutations associated with non-small cell lung cancer, consistent with experimental results. Meanwhile, UdanDTI could complement the advanced molecular docking software DiffDock. The codes and datasets of UdanDTI are available at https://github.com/CQ-zhang-2016/UdanDTI .

Overview

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1101/2024.09.12.612771

Cites

1 of 1 references verified0 citation errors · 0 not checkable

Statistics

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Flagged first, then by citations

Research works printed but not linked: 70

  1. 10.1093/bioinformatics/bty593 · DOI
  2. 10.1038/384014a0 · DOI
  3. 2022, Synthon-based ligand discovery in virtual libraries of over 11 billion compounds, Nature, 601, 452, 10.1038/s41586-021-04220-9 · DOI · Paper page
  4. 10.1093/bioinformatics/btm580 · DOI
  5. 10.1093/bioinformatics/btn162 · DOI
  6. 10.1093/bioinformatics/btn409 · DOI
  7. 10.1038/nature08506 · DOI · Paper page
  8. 2009, Supervised prediction of drug–target interactions using bipartite local models, 25, 2397
  9. 10.1093/bioinformatics/btr500 · DOI
  10. Wang, W. , S. Yang , and J. Li . Drug target predictions based on heterogeneous graph inference. in Pacific symposium on biocomputing. 2013. World Scientific.
  11. Shi, Y. , et al., Protein-chemical interaction prediction via kernelized sparse learning svm, in Biocomputing 2013. 2013, World Scientific. p. 41–52. · DOI
  12. 10.1016/j.gpb.2019.04.003 · DOI · Paper page
  13. 2022, HyperAttentionDTI: improving drug-protein interaction prediction by sequence-based deep learning with attention mechanism, Bioinformatics, 38, 655, 10.1093/bioinformatics/btab715 · DOI
  14. 2019, Compound-protein interaction prediction with end-to-end learning of neural networks for graphs and sequences, Bioinformatics, 35, 309, 10.1093/bioinformatics/bty535 · DOI · Paper page
  15. 2023, Interpretable bilinear attention network with domain adaptation improves drug–target prediction, Nat Mach Intell, 5, 126, 10.1038/s42256-022-00605-1 · DOI
  16. Bian, J. , et al., MCANet: shared-weight-based MultiheadCrossAttention network for drug–target interaction prediction. Brief Bioinform, 2023. 24(2). · DOI
  17. 2021, Identifying drug-target interactions based on graph convolutional network and deep neural network, Brief Bioinform, 22, 2141, 10.1093/bib/bbaa044 · DOI
  18. 2022, Drug-target interaction predication via multi-channel graph neural networks, Brief Bioinform, 23, bbab346, 10.1093/bib/bbab346 · DOI · Paper page
  19. 2022, BridgeDPI: a novel Graph Neural Network for predicting drug–protein interactions, Bioinformatics, 38, 2571, 10.1093/bioinformatics/btac155 · DOI
  20. 2023, CPInformer for Efficient and Robust Compound-Protein Interaction Prediction, IEEE/ACM Trans Comput Biol Bioinform, 20, 285, 10.1109/TCBB.2022.3144008 · DOI
  21. 2021, GraphDTA: predicting drug-target binding affinity with graph neural networks, Bioinformatics, 37, 1140, 10.1093/bioinformatics/btaa921 · DOI
  22. 2022, Yuel: Improving the Generalizability of Structure-Free Compound-Protein Interaction Prediction, J Chem Inf Model, 62, 463, 10.1021/acs.jcim.1c01531 · DOI
  23. 2021, DTI-Voodoo: machine learning over interaction networks and ontology-based background knowledge predicts drug-target interactions, Bioinformatics, 37, 4835, 10.1093/bioinformatics/btab548 · DOI
  24. 10.1371/journal.pcbi.1007129 · DOI
  25. 10.1093/bioinformatics/btaa880 · DOI
  26. 10.1093/bioinformatics/btaa524 · DOI · Paper page
  27. Gao, B. , et al., DrugCLIP: Contrasive Protein-Molecule Representation Learning for Virtual Screening, in Advances in Neural Information Processing Systems. 2024.
  28. 2024, CCL-DTI: contributing the contrastive loss in drug-target interaction prediction, BMC Bioinformatics, 25, 48, 10.1186/s12859-024-05671-3 · DOI
  29. Dehghan, A. , et al., TripletMultiDTI: Multimodal representation learning in drug-target interaction prediction with triplet loss function. Expert Syst Appl, 2023. 232. · DOI
  30. 10.1073/pnas.2220778120 · DOI
  31. 10.3390/pharmaceutics14081710 · DOI
  32. 2022, Affinity2Vec: drug-target binding affinity prediction through representation learning, graph mining, and machine learning, Sci Rep, 12, 4751, 10.1038/s41598-022-08787-9 · DOI
  33. 2023, Improving the generalizability of protein-ligand binding predictions with AI-Bind, Nat Commun, 14, 1989, 10.1038/s41467-023-37572-z · DOI
  34. 10.1021/acs.jcim.8b00712 · DOI
  35. 2021, Systematic auditing is essential to debiasing machine learning in biology, Commun Biol, 4, 183, 10.1038/s42003-021-01674-5 · DOI
  36. 2022, MGraphDTA: deep multiscale graph neural network for explainable drug-target binding affinity prediction, Chem Sci, 13, 816, 10.1039/D1SC05180F · DOI
  37. 10.1126/science.ade2574 · DOI · Paper page
  38. 2022, ProtTrans: Toward Understanding the Language of Life Through Self-Supervised Learning, IEEE Trans Pattern Anal Mach Intell, 44, 7112, 10.1109/TPAMI.2021.3095381 · DOI · Paper page
  39. Saito, K. , et al. Maximum classifier discrepancy for unsupervised domain adaptation. in Proceedings of the IEEE conference on computer vision and pattern recognition. 2018. · DOI
  40. Bai, P. , et al. Hierarchical clustering split for low-bias evaluation of drug-target interaction prediction. in 2021 IEEE International Conference on Bioinformatics and Biomedicine (BIBM). 2021. IEEE. · DOI
  41. 10.1093/bioinformatics/btv256 · DOI
  42. 2021, Structure-based classification predicts drug response in EGFR-mutant NSCLC, Nature, 597, 732, 10.1038/s41586-021-03898-1 · DOI · Paper page
  43. 2021, Splitting chemical structure data sets for federated privacy-preserving machine learning, J Cheminform, 13, 96, 10.1186/s13321-021-00576-2 · DOI
  44. Vaswani, A. , et al., Attention is all you need, in Advances in neural information processing systems. 2017.
  45. Hou, R. , et al., Cross attention network for few-shot classification, in Advances in neural information processing systems. 2019.
  46. Lin, H. , et al. Cat: Cross attention in vision transformer. in 2022 IEEE International Conference on Multimedia and Expo (ICME). 2022. IEEE. · DOI
  47. Chen, C.-F.R. , Q. Fan , and R. Panda . Crossvit: Cross-attention multi-scale vision transformer for image classification. in Proceedings of the IEEE/CVF international conference on computer vision. 2021. · DOI
  48. 10.1007/s10822-020-00314-0 · DOI · Paper page
  49. 2022, Explainable deep drug-target representations for binding affinity prediction, BMC Bioinformatics, 23, 237, 10.1186/s12859-022-04767-y · DOI
  50. Corso, G. , et al., Diffdock: Diffusion steps, twists, and turns for molecular docking, in International Conference on Learning Representations (ICLR). 2022.
  51. 2016, BindingDB in 2015: a public database for medicinal chemistry, computational chemistry and systems pharmacology, 44, D1045
  52. Bai, P. , et al. Hierarchical clustering split for low-bias evaluation of drug-target interaction prediction. in 2021 IEEE International Conference on Bioinformatics and Biomedicine (BIBM). 2021. IEEE. · DOI
  53. 2021, MolTrans: molecular interaction transformer for drug–target interaction prediction, 37, 830
  54. 10.1093/bioinformatics/btv256 · DOI
  55. 2021, Structure-based classification predicts drug response in EGFR-mutant NSCLC, Nature, 597, 732, 10.1038/s41586-021-03898-1 · DOI · Paper page
  56. 2010, Hahn, and modeling, Extended-connectivity fingerprints, 50, 742
  57. 2013, propy: a tool to generate various modes of Chou’s PseAAC, 29, 960
  58. 2019, DeepConv-DTI: Prediction of drug-target interactions via deep learning with convolution on protein sequences, 15, e1007129
  59. 2021, GraphDTA: Predicting drug–target binding affinity with graph neural networks, 37, 1140
  60. 2022, MGraphDTA: deep multiscale graph neural network for explainable drug-target binding affinity prediction, Chem Sci, 13, 816, 10.1039/D1SC05180F · DOI
  61. Bian, J. , et al., MCANet: shared-weight-based MultiheadCrossAttention network for drug– target interaction prediction. Briefings in Bioinformatics, 2023. 24(2). · DOI
  62. 2023, Interpretable bilinear attention network with domain adaptation improves drug–target prediction, Nature Machine Intelligence, 5, 126, 10.1038/s42256-022-00605-1 · DOI
  63. 2022, Yuel: Improving the Generalizability of Structure-Free Compound-Protein Interaction Prediction, J Chem Inf Model, 62, 463, 10.1021/acs.jcim.1c01531 · DOI
  64. 2021, Splitting chemical structure data sets for federated privacy-preserving machine learning, 13, 1
  65. Madani, A. , et al., Large language models generate functional protein sequences across diverse families. 2023: p. 1–8. · DOI · Paper page
  66. 2021, Learning the protein language: Evolution, structure, and function, 12, 654
  67. 2023, Evolutionary-scale prediction of atomic-level protein structure with a language model, 379, 1123
  68. 2022, ProtTrans: Toward Understanding the Language of Life Through Self-Supervised Learning, IEEE Trans Pattern Anal Mach Intell, 44, 7112, 10.1109/TPAMI.2021.3095381 · DOI · Paper page
  69. 2022, Large-scale chemical language representations capture molecular structure and properties, 4, 1256
  70. 10.1093/nar/gkaa1100 · DOI · Paper page

Metadata

Authors
Zhang, P.-D., Ma, J., & Chen, T.
Affiliations
  • Tsinghua University (CN)
  • Beijing National Research Center for Information Science and Technology (CN)
Corresponding author
  • Ma, Jianzhu
  • Chen, Ting
ORCID
Venue
IEEE Transactions on Computational Biology and Bioinformatics
Publisher
openRxiv
Published
2024-09-15
Volume
22
Issue
4
Pages
1902-1911
ISSN
1545-5963 (ISSN-L)
DOI
10.1101/2024.09.12.612771
OpenAlex
W4411019816
Cited by
13
References
1
Citations by year
2026: 6 · 2025: 7
FWCI
6.72
Type
journal article
Language
en
Open access
Licence
Locations
  • IEEE Transactions on Computational Biology and Bioinformatics (publishedVersion, cc-by-nc-nd)
  • bioRxiv (Cold Spring Harbor Laboratory) (acceptedVersion, cc-by-nc)
  • PubMed (publishedVersion)
Funders
  • Guoqiang Institute, Tsinghua University
  • Beijing National Research Center For Information Science And Technology
  • National Key Research and Development Program of China (2024YFF1207103, 2024YFF1207100, 2022YFC2703100, 2022YFC2703105)
Topics
  • Life Sciences › Pharmacology, Toxicology and Pharmaceutics › Drug Discovery › Computational Drug Discovery Methods
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Machine Learning in Bioinformatics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Protein Degradation and Inhibitors
Keywords
drug-target interaction prediction, interpretability, generalization, dual-branch architecture, virtual screening, EGFR mutation, non-small cell lung cancer, drug response prediction, DiffDock

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review