PreTKcat: A pre-trained representation learning and machine learning framework for predicting enzyme turnover number.

Area of Science:

• Biochemistry and Molecular Biology
• Computational Chemistry
• Enzyme Engineering

Background:

• Enzyme turnover number (kcat) is a critical parameter in enzyme kinetics, essential for optimizing biotechnological applications.
• Experimental determination of kcat is resource-intensive, creating a need for efficient computational prediction methods.
• Accurate prediction of enzyme kinetics parameters like kcat and km is vital for enzyme engineering and industrial enzyme applications.

Purpose of the Study:

• To develop and validate a novel computational framework, PreTKcat, for predicting enzyme turnover number (kcat).
• To integrate pre-trained protein and molecular representation learning with machine learning for enhanced prediction accuracy.
• To assess the capability of PreTKcat in predicting enzyme-substrate affinity (km) and the influence of temperature on kcat.

Main Methods:

• Utilized ProtT5 for encoding enzyme sequences and MolGNet for encoding substrate molecular graphs.
• Employed the ExtraTrees machine learning model to integrate sequence and structural information for kcat prediction.
• Incorporated temperature as a variable in the prediction model and extended the framework for km prediction.

Main Results:

• PreTKcat demonstrated superior performance compared to existing state-of-the-art models in predicting kcat values.
• The framework effectively integrates enzyme sequence and substrate molecular features for accurate kinetic predictions.
• PreTKcat successfully predicted enzyme-substrate affinity (km) and accounted for temperature effects on kcat.

Conclusions:

• PreTKcat provides a powerful and accurate computational tool for predicting enzyme kinetics parameters, specifically kcat and km.
• The framework offers a cost-effective alternative to experimental measurements, accelerating enzyme engineering and development.
• PreTKcat opens new avenues for enzyme discovery, optimization, and industrial application by providing reliable kinetic insights.