Background Enzyme turnover numbers (<InlineEquation ID="IEq1"> <EquationSource Format="TEX">${\text{K}}_{\text{cat}}$</EquationSource> <EquationSource Format="MATHML"><math> <msub> <mtext>K</mtext> <mtext>cat</mtext> </msub> </math></EquationSource> </InlineEquation>) are fundamental kinetic constants that quantify enzymatic efficiency. Systematic studies of <InlineEquation ID="IEq2"> <EquationSource Format="TEX">${\text{K}}_{\text{cat}}$</EquationSource> <EquationSource Format="MATHML"><math> <msub> <mtext>K</mtext> <mtext>cat</mtext> </msub> </math></EquationSource> </InlineEquation> are essential for characterizing the mechanisms underlying proteomic composition and cellular metabolism. However, experimental measurements of <InlineEquation ID="IEq3"> <EquationSource Format="TEX">${\text{K}}_{\text{cat}}$</EquationSource> <EquationSource Format="MATHML"><math> <msub> <mtext>K</mtext> <mtext>cat</mtext> </msub> </math></EquationSource> </InlineEquation> remain limited and prone to noise. Results To address this, we present KcatNet, a geometric deep learning model designed for high-throughput prediction of <InlineEquation ID="IEq4"> <EquationSource Format="TEX">${\text{K}}_{\text{cat}}$</EquationSource> <EquationSource Format="MATHML"><math> <msub> <mtext>K</mtext> <mtext>cat</mtext> </msub> </math></EquationSource> </InlineEquation> in metabolic enzymes across all organisms, leveraging paired enzyme sequence and substrate representations. KcatNet consistently outperforms existing predictors, particularly for enzymes with high catalytic efficiency, and demonstrates strong generalization to enzymes that are dissimilar to those in the training set. Furthermore, KcatNet uncovers structural mechanisms and interaction patterns within enzyme–substrate complexes, providing insights into architectural principles that are inaccessible with existing methods by harnessing the representational power of large-scale protein language models. We apply KcatNet to genome-scale <InlineEquation ID="IEq5"> <EquationSource Format="TEX">${\text{K}}_{\text{cat}}$</EquationSource> <EquationSource Format="MATHML"><math> <msub> <mtext>K</mtext> <mtext>cat</mtext> </msub> </math></EquationSource> </InlineEquation> prediction across diverse yeast species, improving proteome allocation predictions by integrating its outputs into metabolic models. Experimental validation confirms the model's ability to identify enzyme mutants with enhanced activity. Conclusion By bridging the gap between sequence, structure, and function, KcatNet establishes a robust foundation for advancing understanding of molecular-level mechanisms and accelerating enzyme engineering efforts.

A geometric deep learning framework for genome-wide prediction of enzyme turnover number

期刊论文
出版模式：

开放获取
发表日期：

2026年2月11日

查看全文

Tong Pan,
Xin Cui,
Huan Yee Koh,
Yue Bi,
Xiaoyu Wang,
Yumeng Zhang,
Shantong Hu,
Geoffrey I. Webb,
Lukasz Kurgan,
Guimin Zhang,
Jiangning Song

摘要

Background

Enzyme turnover numbers ( ${\text{K}}_{\text{cat}}$ $K_{cat}$ ) are fundamental kinetic constants that quantify enzymatic efficiency. Systematic studies of ${\text{K}}_{\text{cat}}$ $K_{cat}$ are essential for characterizing the mechanisms underlying proteomic composition and cellular metabolism. However, experimental measurements of ${\text{K}}_{\text{cat}}$ $K_{cat}$ remain limited and prone to noise.

Results

To address this, we present KcatNet, a geometric deep learning model designed for high-throughput prediction of ${\text{K}}_{\text{cat}}$ $K_{cat}$ in metabolic enzymes across all organisms, leveraging paired enzyme sequence and substrate representations. KcatNet consistently outperforms existing predictors, particularly for enzymes with high catalytic efficiency, and demonstrates strong generalization to enzymes that are dissimilar to those in the training set. Furthermore, KcatNet uncovers structural mechanisms and interaction patterns within enzyme–substrate complexes, providing insights into architectural principles that are inaccessible with existing methods by harnessing the representational power of large-scale protein language models. We apply KcatNet to genome-scale ${\text{K}}_{\text{cat}}$ $K_{cat}$ prediction across diverse yeast species, improving proteome allocation predictions by integrating its outputs into metabolic models. Experimental validation confirms the model's ability to identify enzyme mutants with enhanced activity.

Conclusion

By bridging the gap between sequence, structure, and function, KcatNet establishes a robust foundation for advancing understanding of molecular-level mechanisms and accelerating enzyme engineering efforts.