Information leakage in enzyme-substrate prediction
5 affiliations
- 1Drug Bioinformatics, Helmholtz Institute for Pharmaceutical Research Saarland , Saarbruecken, Saarland 66123,
- 2Centre for Bioinformatics, Saarland University , Saarbruecken, Saarland 66123,
- 3Department of Chemistry and Molecular Biology, University of Gothenburg , Gothenburg 41390,
- 4Wallenberg Centre for Molecular and Translational Medicine, University of Gothenburg , Gothenburg 41390,
- 5Medical Faculty, Saarland University , Homburg, Saarland 66421,
Abstract
Motivation: Enzymes are essential catalysts in many cellular processes. Understanding their interactions with small molecules, such as regulators, cofactors, and most importantly, substrates, is crucial for understanding the biochemical processes that occur in cells. Correctly interpreting the roles of small molecules that interact with enzymes is key to elucidating enzyme function. Recently, enzyme–small-molecule interaction prediction has attracted growing interest from computational methods, especially deep learning. As a result, researchers have published several datasets and numerous models with remarkable performance.
Results: In this work, we critically examine one of the most popular datasets and four models trained on it, identifying similarity-induced information leakage that may overinflate reported model performance in out-of-distribution applications. We show that the inspected models are susceptible to information leakage, and their performance is not better than statistical baselines when the leakage is removed. Furthermore, controlling for leakage due to protein sequence similarity alone still yields high predictive performance, whereas performance decreases substantially when common substrates shared between training and test sets are removed from one of them, and approaches random prediction when structurally similar ligands are separated between splits. Thus, the investigated models generalize considerably better across enzyme sequence space than across ligand chemical space, indicating that their reported performance depends strongly on the small-molecule similarity between training and evaluation data.
Availability and implementation: All data splits we calculated in this study are available on Zenodo (DOI: 10.5281/zenodo.18786394). The code is available on GitHub and backed up on Zenodo (DOI: 10.5281/zenodo.18788609).
Related
- DeepEnzyKat: Turnover number prediction through enzyme-substrate interaction modeling and adaptive expert ensembles
- Decoding enzyme-substrate interaction topology reveals principles underlying catalytic efficiency and mutational outcomes
- miR1919–MCD1 Regulation of Debneyol Biosynthesis in Nicotiana: Enzyme Organisation, Catalytic Evidence and the Limits of Substrate Channelling
- Evolution promotes transition-state-like conformations with enhanced basicity and electric fields in the substrate complex of a designer enzyme
- Joint enzyme–reaction retrieval and catalytic optima prediction via multimodal fusion
BibTeX
@article{elmi2026information,
title = {{Information leakage in enzyme-substrate prediction}},
author = {Vahid Atabaigi Elmi and Roman Joeres and Olga V Kalinina},
journal = {Bioinformatics},
year = {2026},
month = oct,
doi = {10.1093/bioinformatics/btag733},
url = {https://academic.oup.com/bioinformatics/article/doi/10.1093/bioinformatics/btag733/8860137},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Information leakage in enzyme-substrate prediction
AU - Vahid Atabaigi Elmi
AU - Roman Joeres
AU - Olga V Kalinina
PY - 2026
DA - 2026/10/01
JO - Bioinformatics
DO - 10.1093/bioinformatics/btag733
UR - https://academic.oup.com/bioinformatics/article/doi/10.1093/bioinformatics/btag733/8860137
AB - Motivation: Enzymes are essential catalysts in many cellular processes. Understanding their interactions with small molecules, such as regulators, cofactors, and most importantly, substrates, is crucial for understanding the biochemical processes that occur in cells. Correctly interpreting the roles of small molecules that interact with enzymes is key to elucidating enzyme function. Recently, enzyme–small-molecule interaction prediction has attracted growing interest from computational methods, especially deep learning. As a result, researchers have published several datasets and numerous models with remarkable performance. Results: In this work, we critically examine one of the most popular datasets and four models trained on it, identifying similarity-induced information leakage that may overinflate reported model performance in out-of-distribution applications. We show that the inspected models are susceptible to information leakage, and their performance is not better than statistical baselines when the leakage is removed. Furthermore, controlling for leakage due to protein sequence similarity alone still yields high predictive performance, whereas performance decreases substantially when common substrates shared between training and test sets are removed from one of them, and approaches random prediction when structurally similar ligands are separated between splits. Thus, the investigated models generalize considerably better across enzyme sequence space than across ligand chemical space, indicating that their reported performance depends strongly on the small-molecule similarity between training and evaluation data. Availability and implementation: All data splits we calculated in this study are available on Zenodo (DOI: 10.5281/zenodo.18786394). The code is available on GitHub and backed up on Zenodo (DOI: 10.5281/zenodo.18788609).
SN - 1367-4803
SN - 1367-4811
ER -