CoChat

The DOI matches its Crossref record; 73 of the 73 references it cites were found in a registry or an index.

Machine learning applied to enzyme turnover numbers reveals protein structural correlates and improves metabolic models

Open access

Abstract

Knowing the catalytic turnover numbers of enzymes is essential for understanding the growth rate, proteome composition, and physiology of organisms, but experimental data on enzyme turnover numbers is sparse and noisy. Here, we demonstrate that machine learning can successfully predict catalytic turnover numbers in Escherichia coli based on integrated data on enzyme biochemistry, protein structure, and network context. We identify a diverse set of features that are consistently predictive for both in vivo and in vitro enzyme turnover rates, revealing novel protein structural correlates of catalytic turnover. We use our predictions to parameterize two mechanistic genome-scale modelling frameworks for proteome-limited metabolism, leading to significantly higher accuracy in the prediction of quantitative proteome data than previous approaches. The presented machine learning models thus provide a valuable tool for understanding metabolism and the proteome at the genome scale, and elucidate structural, biochemical, and network properties that underlie enzyme kinetics.

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 · 10.1038/s41467-018-07652-6

Cites

73 of 73 references verified0 citation errors · 0 not checkable

Statistics

Figures

Fig. 1 Machine learning of catalytic turnover numbers for genome-scale metabolic model (GEM) parameterization. A feature set from diverse classes is curated and mapped to independently build machine learning (ML) models of both kcat in vitro (f(x)) and kapp,max in vivo (g(x)). The inferred ML models are used to predict kcat in vitro or kapp,max at the genome-scale to parameterize GEMs
Fig. 1 Machine learning of catalytic turnover numbers for genome-scale metabolic model (GEM) parameterization. A feature set from diverse classes is curated and mapped to independently build machine learning (ML) models of both kcat in vitro (f(x)) and kapp,max in vivo (g(x)). The inferred ML models are used to predict kcat in vitro or kapp,max at the genome-scale to parameterize GEMsCC BY · PubMed Central
Fig. 2 Machine learning model performances for kapp,max and kcat in vitro. Center lines show the median R2 across five times repeated five-fold cross-validation (25 validations), except for the deep learning case, where the median for a single round of five-fold cross-validation (five validations) is shown. Box limits represent the 1st and 3rd quartiles, whiskers extend to values that lie within the 1.5x interquartile range, and the remaining points are shown as outliers (marked x). Circles show R2 for a test set consisting of 20% of the available samples that were not used for hyperparameter optimization. This resulted in a training set of 172 observations of kcat in vitro and 106 observations of kapp,max. For the test set, 43 and 27 observations were used for kcat in vitro and kapp,max, respectively. See Methods for details on hyperparameter optimization
Fig. 2 Machine learning model performances for kapp,max and kcat in vitro. Center lines show the median R2 across five times repeated five-fold cross-validation (25 validations), except for the deep learning case, where the median for a single round of five-fold cross-validation (five validations) is shown. Box limits represent the 1st and 3rd quartiles, whiskers extend to values that lie within the 1.5x interquartile range, and the remaining points are shown as outliers (marked x). Circles show R2 for a test set consisting of 20% of the available samples that were not used for hyperparameter optimization. This resulted in a training set of 172 observations of kcat in vitro and 106 observations of kapp,max. For the test set, 43 and 27 observations were used for kcat in vitro and kapp,max, respectively. See Methods for details on hyperparameter optimizationCC BY · PubMed Central
Fig. 3 Feature importance in random forest models for in vivo and in vitro turnover numbers. The relative importance as measured by the average decrease in out-of-bag mean squared error (MSE) across trees that results from randomly permuting a given feature (scaled by the standard deviation) is shown. Missing bars indicate permutation importance smaller or equal to zero. The statistical significance of feature importance was evaluated using a permutation test based on 500 permutations of the response variable per model; *p-value < 0.05, **p-value < 0.005. Spearman rank correlation between the importance estimates of the two models is 0.47 (p < 0.021, n = 24, S = 1214, see Methods), ignoring assay-related features that are not used in the model for kapp,max
Fig. 3 Feature importance in random forest models for in vivo and in vitro turnover numbers. The relative importance as measured by the average decrease in out-of-bag mean squared error (MSE) across trees that results from randomly permuting a given feature (scaled by the standard deviation) is shown. Missing bars indicate permutation importance smaller or equal to zero. The statistical significance of feature importance was evaluated using a permutation test based on 500 permutations of the response variable per model; *p-value < 0.05, **p-value < 0.005. Spearman rank correlation between the importance estimates of the two models is 0.47 (p < 0.021, n = 24, S = 1214, see Methods), ignoring assay-related features that are not used in the model for kapp,maxCC BY · PubMed Central
Fig. 4 Performance of vectors of catalytic turnover numbers in predicting quantitative proteome data. Performance for two different genome-scale metabolic modelling frameworks, MOMENT and the ME model, are shown. Model predictions are compared to quantitative proteomics data in Schmidt et al.37 through the root mean squared error (RMSE) for metabolic proteome fractions on log10 scale. Comparisons use proteins that are both found in proteomics data and are expressed in the model predictions. To allow comparison of different parameterization strategies, the intersection of the sets of comparable proteins is used in each condition-model combination resulting in the number of comparisons n. The performance of the two modelling frameworks, MOMENT and ME, is thus not comparable, as different sets of proteins are underlying the error calculations. See Methods and Supplementary Figure 8 for details
Fig. 4 Performance of vectors of catalytic turnover numbers in predicting quantitative proteome data. Performance for two different genome-scale metabolic modelling frameworks, MOMENT and the ME model, are shown. Model predictions are compared to quantitative proteomics data in Schmidt et al.37 through the root mean squared error (RMSE) for metabolic proteome fractions on log10 scale. Comparisons use proteins that are both found in proteomics data and are expressed in the model predictions. To allow comparison of different parameterization strategies, the intersection of the sets of comparable proteins is used in each condition-model combination resulting in the number of comparisons n. The performance of the two modelling frameworks, MOMENT and ME, is thus not comparable, as different sets of proteins are underlying the error calculations. See Methods and Supplementary Figure 8 for detailsCC 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

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

5 more claims about this paper

Create a free account to see every claim other papers make about this one.

See all claims — free account

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 73 references

Cited by

See all 218

Metadata

Authors
Heckmann, D., Lloyd, C. J., Mih, N., Ha, Y., Zielinski, D. C., Haiman, Z. B., Desouki, A. A., Lercher, M. J., & Palsson, B. O.
Affiliations
  • University of California San Diego (US)
  • Heinrich Heine University Düsseldorf (DE)
Corresponding author
Heckmann, David
Venue
Nature Communications
Published
2018-12-07
Volume
9
Issue
1
Pages
5252-5252
ISSN
2041-1723 (ISSN-L)
OpenAlex
W2903357856
Cited by
218
References
73
Citations by year
2026: 21 · 2025: 34 · 2024: 31 · 2023: 31 · 2022: 32 · 2021: 30 · 2020: 24 · 2019: 15
FWCI
6.13
Type
journal article
Language
en
Open access
Licence
Locations
  • Nature Communications (publishedVersion, cc-by)
  • Europe PMC (PubMed Central) (publishedVersion, other-oa)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • RePEc: Research Papers in Economics (submittedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
  • eScholarship (California Digital Library) (submittedVersion)
  • eScholarship (California Digital Library) (submittedVersion, cc-by)
  • Technical University of Denmark, DTU Orbit (Technical University of Denmark, DTU) (submittedVersion, other-oa)
Funders
  • Deutsche Forschungsgemeinschaft (CRC1310)
  • Novo Nordisk (NNF10CC1016517)
  • Novo Nordisk Fonden (NNF10CC1016517)
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Microbial Metabolic Engineering and Bioproduction
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Bioinformatics and Genomic Networks
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Gene Regulatory Network Analysis
Keywords
enzyme turnover number, Escherichia coli, protein structure, enzyme kinetics, genome-scale metabolic models, machine learning prediction
MeSH terms
Machine Learning, Algorithms, Escherichia coli, Kinetics, Models, Biological, Proteome, Escherichia coli Proteins, Metabolic Networks and Pathways, Biocatalysis

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: Machine learning applied to enzyme turnover numbers reveals protein structural correlates and improves metabolic models. https://app.cochat.ai/papers/10.1038/s41467-018-07652-6
The paper
Heckmann, D., Lloyd, C. J., Mih, N., Ha, Y., Zielinski, D. C., Haiman, Z. B., Desouki, A. A., Lercher, M. J., & Palsson, B. O. (2018). Machine learning applied to enzyme turnover numbers reveals protein structural correlates and improves metabolic models. Nature Communications. https://doi.org/10.1038/s41467-018-07652-6

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.