DLTKcat: deep learning-based prediction of temperature-dependent enzyme turnover rates

Area of Science:

• Biochemistry
• Computational Biology
• Enzyme Kinetics

Background:

• Enzyme turnover rate (kcat) is crucial for enzyme kinetics, measuring catalytic efficiency.
• Experimental kcat measurements are costly and data is scarce in public databases.
• Temperature significantly influences enzyme activity and kcat values.

Purpose of the Study:

• To develop a predictive model, DLTKcat, for enzyme turnover rates (kcat).
• To incorporate the temperature dependence of kcat into the predictive model.
• To evaluate DLTKcat's performance against existing models and in case studies.

Main Methods:

• Development of the DLTKcat computational model.
• Training and validation using existing enzyme kinetic datasets.
• Assessment of model performance using metrics like RMSE and R-squared.
• Application of DLTKcat in case studies involving mutations and temperature variations.

Main Results:

• DLTKcat demonstrated superior predictive performance compared to previous models (RMSE = 0.88, R-squared = 0.66).
• The model successfully predicted the impact of protein mutations on kcat.
• DLTKcat accurately predicted kcat changes in response to temperature variations.
• Case studies validated the model's ability to capture mutation and temperature effects.

Conclusions:

• DLTKcat offers a promising computational approach for predicting enzyme turnover rates and their temperature dependence.
• The model has potential for future applications in understanding biological system responses to temperature.
• Further improvements in quantitative accuracy are needed for direct cellular metabolism modeling.