CoChat

DeepEnzyme: a robust deep learning model for improved enzyme turnover number prediction by utilizing features of protein 3D structures

Open access
Wang, Tong, Xiang, Guangming, He, Siwei, Su, Liyun, Yan, Xuefeng, Lu, Hongzhong · 2023 · Briefings in Bioinformatics

Preprint — published as 10.1093/bib/bbae409 · Paper page

The DOI matches its Crossref record; 2 of the 2 references it cites were found in a registry or an index.

Abstract

Turnover numbers (kcat), which indicate an enzyme’s catalytic efficiency, have a wide range of applications in fields including protein engineering and synthetic biology. Experimentally measuring the enzymes’ kcat is always time-consuming. Recently, the prediction of kcat using deep learning models has mitigated this problem. However, the accuracy and robustness in kcat prediction still needs to be improved significantly, particularly when dealing with enzymes with low sequence similarity compared to those within the training dataset. Herein, we present DeepEnzyme, a cutting-edge deep learning model that combines the most recent Transformer and Graph Convolutional Network (GCN) architectures. To improve the prediction accuracy, DeepEnzyme was trained by leveraging the integrated features from both sequences and 3D structures. Consequently, our model exhibits remarkable robustness when processing enzymes with low sequence similarity compared to those in the training dataset by utilizing additional features from high-quality protein 3D structures. DeepEnzyme also makes it possible to evaluate how point mutations affect the catalytic activity of the enzyme, which helps identify residue sites that are crucial for the catalytic function. In summary, DeepEnzyme represents a pioneering effort in predicting enzymes’ kcat values with superior accuracy and robustness compared to previous algorithms. This advancement will significantly contribute to our comprehension of enzyme function and its evolutionary patterns across species.

Overview

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

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

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

Integrity signals

Registry

Crossref · 10.1101/2023.12.09.570923

Cites

2 of 2 references verified0 citation errors · 0 not checkable

Statistics

Figures

Figure 1 The framework of DeepEnzyme for kcat prediction. DeepEnzyme integrates transformer and GCN models to distill features from both the enzyme and substrate for predicting kcat. GCN is employed to extract structural features based on protein 3D-structures and substrate adjacency matrices; transformer is utilized to extract sequence features from protein sequences.
Figure 1 The framework of DeepEnzyme for kcat prediction. DeepEnzyme integrates transformer and GCN models to distill features from both the enzyme and substrate for predicting kcat. GCN is employed to extract structural features based on protein 3D-structures and substrate adjacency matrices; transformer is utilized to extract sequence features from protein sequences.CC BY · PubMed Central
Figure 2 Evaluation of DeepEnzyme performance in kcat prediction. (a) the performance of DeepEnzyme on the test dataset was evaluated by the Pearson correlation coefficient (PCC) and P values calculated from the predicted and experimental kcat values. (b-c) DeepEnzyme prediction performance on wild-type (b) and mutant (c) enzymes in the test dataset. (d) DeepEnzyme prediction performance for enzymes classified by different EC numbers in the test dataset. (e) the performance of the model with different types of datasets as input. DeepEnzyme: Enzyme sequence, enzyme structure, and substrate information are used as inputs; Only-structure: Substrate information and enzyme structure information as inputs; Only-sequence: Substrate and enzyme sequence as inputs. In (a)–(e) specific seed was adopted during the calculation. (f) Comparison in average coefficient of determination (R2) values for testing dataset and training dataset from five rounds of training.
Figure 2 Evaluation of DeepEnzyme performance in kcat prediction. (a) the performance of DeepEnzyme on the test dataset was evaluated by the Pearson correlation coefficient (PCC) and P values calculated from the predicted and experimental kcat values. (b-c) DeepEnzyme prediction performance on wild-type (b) and mutant (c) enzymes in the test dataset. (d) DeepEnzyme prediction performance for enzymes classified by different EC numbers in the test dataset. (e) the performance of the model with different types of datasets as input. DeepEnzyme: Enzyme sequence, enzyme structure, and substrate information are used as inputs; Only-structure: Substrate information and enzyme structure information as inputs; Only-sequence: Substrate and enzyme sequence as inputs. In (a)–(e) specific seed was adopted during the calculation. (f) Comparison in average coefficient of determination (R2) values for testing dataset and training dataset from five rounds of training.CC BY · PubMed Central
Figure 3 Improved performances of DeepEnzyme in kcat prediction compared to existing models, even for protein sequences in the test dataset exhibiting lower similarity compared to those in the training dataset. (a) Comparison of R2 values on the test dataset for different models. (b) Comparison of RMSE values on the test dataset for different models. (c) Comparison of R2 in kcat value prediction for enzymes in the test dataset at different levels of sequence similarity by DeepEnzyme, TurNuP, DLKcat, and DLTKcat. (d) Two enzymes from Myxococcus xanthus and Bacillus subtilis, both with EC numbers 1.3.3.4, are highly similar in protein 3D-structure (TM-score = 0.8762), gray for enzyme from M. xanthus and red for enzyme from B. subtilis. (e) the similarity of the amino acid sequences for the above two enzymes is 27% (Q for enzyme from M. xanthus, T for enzyme from B. subtilis).
Figure 3 Improved performances of DeepEnzyme in kcat prediction compared to existing models, even for protein sequences in the test dataset exhibiting lower similarity compared to those in the training dataset. (a) Comparison of R2 values on the test dataset for different models. (b) Comparison of RMSE values on the test dataset for different models. (c) Comparison of R2 in kcat value prediction for enzymes in the test dataset at different levels of sequence similarity by DeepEnzyme, TurNuP, DLKcat, and DLTKcat. (d) Two enzymes from Myxococcus xanthus and Bacillus subtilis, both with EC numbers 1.3.3.4, are highly similar in protein 3D-structure (TM-score = 0.8762), gray for enzyme from M. xanthus and red for enzyme from B. subtilis. (e) the similarity of the amino acid sequences for the above two enzymes is 27% (Q for enzyme from M. xanthus, T for enzyme from B. subtilis).CC BY · PubMed Central
Figure 4 Analysis of the prediction ability of DeepEnzyme for two enzymes with saturation mutagenesis datasets. (a) Comparison of predicted results for different CYP2C9 variants: Red for missense variants, green for nonsense variants. (b) Comparison of experimental activity score for different CYP2C9 variants: Red for missense variants, green for nonsense variants [40]. (c) Comparison of predicted kcat values for different PafA mutations: Green for low kcat mutations, red for high kcat mutations [41]. (d) Comparison of experimentally measured kcat values for different PafA mutations: Green for low kcat mutations, red for high kcat mutations [41].
Figure 4 Analysis of the prediction ability of DeepEnzyme for two enzymes with saturation mutagenesis datasets. (a) Comparison of predicted results for different CYP2C9 variants: Red for missense variants, green for nonsense variants. (b) Comparison of experimental activity score for different CYP2C9 variants: Red for missense variants, green for nonsense variants [40]. (c) Comparison of predicted kcat values for different PafA mutations: Green for low kcat mutations, red for high kcat mutations [41]. (d) Comparison of experimentally measured kcat values for different PafA mutations: Green for low kcat mutations, red for high kcat mutations [41].CC BY · PubMed Central
Figure 5 Comparison between the binding/active site and high-weight site (these 5% residues sites with the highest weight scores calculated by DeepEnzyme) within protein 3D-structures. (a) the weight scores of different residue sites in PafA; the red points are binding/active sites. (b) Comparison of the weight scores between the binding/active sites and general sites in PafA; the green box line indicates general sites, and the red box line indicates binding/active sites. (c) the distribution of the binding/active sites and high-weight sites regions within the 3D-structure of PafA, the red region represents for binding/active sites, the blue region for high-weight sites. (d) the weights of different residue sites in P00558; the red points are binding/active sites. (e) Comparison of the weights between the binding/active sites and general sites in P00558, where the green box line indicates general sites, and the red box line indicates binding/active sites. (f) the distribution of the binding/active sites and high-weight sites regions within the 3D-structure of P00558, the red region represents for binding/active sites, the blue region for high-weight sites, the yellow region for the overlap between the two.
Figure 5 Comparison between the binding/active site and high-weight site (these 5% residues sites with the highest weight scores calculated by DeepEnzyme) within protein 3D-structures. (a) the weight scores of different residue sites in PafA; the red points are binding/active sites. (b) Comparison of the weight scores between the binding/active sites and general sites in PafA; the green box line indicates general sites, and the red box line indicates binding/active sites. (c) the distribution of the binding/active sites and high-weight sites regions within the 3D-structure of PafA, the red region represents for binding/active sites, the blue region for high-weight sites. (d) the weights of different residue sites in P00558; the red points are binding/active sites. (e) Comparison of the weights between the binding/active sites and general sites in P00558, where the green box line indicates general sites, and the red box line indicates binding/active sites. (f) the distribution of the binding/active sites and high-weight sites regions within the 3D-structure of P00558, the red region represents for binding/active sites, the blue region for high-weight sites, the yellow region for the overlap between the two.CC BY · PubMed Central
Figure 6 Predicted kcat values for enzyme-catalyzed reactions in genome-scale metabolic models. (a) Distribution of kcat values predicted by DeepEnzyme for enzyme-catalyzed reactions in metabolic models including those for Homo sapiens, Mus musculus, Saccharomyces cerevisiae, and E. Coli [42]. (b) Distribution of kcat values predicted by DeepEnzyme for enzyme-catalyzed reactions from the GEMs of Geobacter metallireducens GS-15 (BiGG ID: iAF987).
Figure 6 Predicted kcat values for enzyme-catalyzed reactions in genome-scale metabolic models. (a) Distribution of kcat values predicted by DeepEnzyme for enzyme-catalyzed reactions in metabolic models including those for Homo sapiens, Mus musculus, Saccharomyces cerevisiae, and E. Coli [42]. (b) Distribution of kcat values predicted by DeepEnzyme for enzyme-catalyzed reactions from the GEMs of Geobacter metallireducens GS-15 (BiGG ID: iAF987).CC BY · PubMed Central

Verification

Claims the paper makes

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

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

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

Show all 2 references

Research works printed but not linked: 38

  1. 2023, Data integration across conditions improves turnover number estimates and metabolic predictions, Nature Communications, 14, 1485, 10.1038/s41467-023-37151-2 · DOI · Paper page
  2. 10.1073/pnas.1514240113 · DOI · Paper page
  3. 2017, Metabolic models of protein allocation call for the kinetome, Cell Systems, 5, 538, 10.1016/j.cels.2017.11.013 · DOI · Paper page
  4. Sánchez, B.J. , et al. Improving the phenotype predictions of a yeast genome -scale metabolic model by incorporating enzymatic constraints. Molecular systems biology 13, 935 (2017). · DOI · Paper page
  5. 10.1016/j.mib.2018.01.002 · DOI · Paper page
  6. 2017, The BRENDA enzyme information system–From a database to an expert system, Journal of biotechnology, 261, 194, 10.1016/j.jbiotec.2017.04.020 · DOI · Paper page
  7. 10.1093/nar/gkx1065 · DOI · Paper page
  8. 10.1093/NAR/GKAC1052 · DOI · Paper page
  9. 2021, Structure-based protein function prediction using graph convolutional networks, Nature communications, 12, 3168, 10.1038/s41467-021-23303-9 · DOI · Paper page
  10. 2020, Deep learning suggests that gene expression is encoded in all parts of a co-evolving interacting gene regulatory structure, Nature Communications, 11, 6141, 10.1038/s41467-020-19921-4 · DOI · Paper page
  11. 2018, Machine learning applied to enzyme turnover numbers reveals protein structural correlates and improves metabolic models, Nature communications, 9, 5252, 10.1038/s41467-018-07652-6 · DOI · Paper page
  12. Kroll, A. , Hu, X.-P. , Liebrand, N.A. & Lercher, M.J . Turnover number predictions for kinetically uncharacterized enzymes using machine and deep learning. bioRxiv, 2022.2011. 2010.516024 (2022). · DOI
  13. Goodsell, D.S. The machinery of life. (2009). · DOI · Paper page
  14. 2019, The importance of catalytic promiscuity for enzyme design and evolution, Nature Reviews Chemistry, 3, 687, 10.1038/s41570-019-0143-x · DOI · Paper page
  15. 2019, Elucidating structure–performance relationships in whole-cell cooperative enzyme catalysis, Nature catalysis, 2, 809, 10.1038/s41929-019-0321-8 · DOI
  16. 10.1371/journal.pcbi.1008291 · DOI · Paper page
  17. 2018, High precision protein functional site detection using 3D convolutional neural networks, Bioinformatics, 35, 1503
  18. 10.1126/science.abj8754 · doi:10.1126/science.abj8754 · Paper page
  19. 2022, ColabFold: making protein folding accessible to all, Nature methods, 19, 679, 10.1038/s41592-022-01488-1 · doi:10.1038/s41592-022-01488-1 · Paper page
  20. Kipf, T.N. & Welling, M. Semi-supervised classification with graph convolutional networks. arXiv preprint arXiv:1609.02907 (2016). · arXiv
  21. Duvenaud, D.K. , et al. Convolutional networks on graphs for learning molecular fingerprints. Advances in neural information processing systems 28 (2015).
  22. 2017, Convolutional embedding of attributed molecular graphs for physical property prediction, Journal of chemical information and modeling, 57, 1757, 10.1021/acs.jcim.6b00601 · DOI · Paper page
  23. Vaswani, A. , et al. Attention is all you need. Advances in neural information processing systems 30 (2017).
  24. 10.1126/science.ade2574 · doi:10.1126/science.ade2574 · Paper page
  25. Madani, A. , et al. Large language models generate functional protein sequences across diverse families. Nature Biotechnology, 1–8 (2023). · DOI · Paper page
  26. Landrum, G . RDKit: A software suite for cheminformatics, computational chemistry, and predictive modeling. Greg Landrum 8, 31 (2013).
  27. 2023, A general model to predict small molecule substrates of enzymes based on machine and deep learning, Nature Communications, 14, 2787, 10.1038/s41467-023-38347-2 · DOI · Paper page
  28. 10.1038/nbt.3988 · DOI · Paper page
  29. van Kempen, M. , et al. Fast and accurate protein structure search with Foldseek. Nature Biotechnology (2023). · DOI · Paper page
  30. 2022, US-align: universal structure alignments of proteins, nucleic acids, and macromolecular complexes, Nature Methods, 19, 1109, 10.1038/s41592-022-01585-1 · DOI · Paper page
  31. 10.1146/annurev.pharmtox.45.120403.095821 · DOI · Paper page
  32. 2013, Cytochrome P450 enzymes in drug metabolism: regulation of gene expression, enzyme activities, and impact of genetic variation, Pharmacology & therapeutics, 138, 103, 10.1016/j.pharmthera.2012.12.007 · DOI · Paper page
  33. 10.1097/00008571-199710000-00004 · DOI · Paper page
  34. 10.1016/j.ajhg.2021.07.001 · DOI
  35. 10.1371/journal.pbio.3001402 · DOI · Paper page
  36. Yu, H. , Deng, H. , He, J. , Keasling, J. & Luo, X. Highly accurate enzyme turnover number prediction and enzyme engineering with PreKcat. (2023). · DOI · Paper page
  37. 2022, Lm-gvp: an extensible sequence and structure informed deep learning framework for protein property prediction, Scientific reports, 12, 6832, 10.1038/s41598-022-10775-y · DOI · Paper page
  38. 2020, SciPy 1.0: fundamental algorithms for scientific computing in Python, Nature methods, 17, 261, 10.1038/s41592-019-0686-2 · DOI · Paper page

Metadata

Authors
Wang, T., Xiang, G., He, S., Su, L., Yan, X., & Lu, H.
Affiliations
  • Shanghai Jiao Tong University (CN)
  • Chongqing University of Technology (CN)
  • East China University of Science and Technology (CN)
Corresponding author
Lu, Hongzhong
Venue
Briefings in Bioinformatics
Publisher
openRxiv
Published
2023-12-10
Volume
25
Issue
5
ISSN
1467-5463 (ISSN-L)
OpenAlex
W4401715368
Cited by
57
References
2
Citations by year
2026: 30 · 2025: 22 · 2024: 5
FWCI
7.74
Type
journal article
Language
en
Open access
Licence
cc-by
Locations
  • Briefings in Bioinformatics (publishedVersion, cc-by)
  • bioRxiv (Cold Spring Harbor Laboratory) (acceptedVersion)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
Funders
  • National Natural Science Foundation of China (22208211, 22378263)
  • Shanghai Jiao Tong University
  • National Key Research and Development Program of China (2022YFA0913000)
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Machine Learning in Bioinformatics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Protein Structure and Dynamics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Structural Biology › Enzyme Structure and Function
Keywords
turnover number, catalytic efficiency, protein 3D structure, protein sequences, transformer models, graph convolutional networks, low sequence similarity, catalytic residues
MeSH terms
Deep Learning, Algorithms, Enzymes, Models, Molecular, Protein Conformation, Proteins, Computational Biology

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

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: DeepEnzyme: a robust deep learning model for improved enzyme turnover number prediction by utilizing features of protein 3D structures. https://app.cochat.ai/papers/10.1101/2023.12.09.570923
The paper
Wang, T., Xiang, G., He, S., Su, L., Yan, X., & Lu, H. (2023). DeepEnzyme: a robust deep learning model for improved enzyme turnover number prediction by utilizing features of protein 3D structures. Briefings in Bioinformatics. https://doi.org/10.1101/2023.12.09.570923

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.