AUKAT: Conditional VAE-Driven Augmentation and Neural Modeling of Enzyme Turnover Numbers

Area of Science:

• Biochemistry
• Computational Biology
• Bioinformatics

Background:

• Accurate enzyme turnover number (kcat) prediction is crucial for systems biology, metabolic engineering, and drug discovery.
• Current prediction methods face challenges due to limited and unevenly distributed experimental data.

Purpose of the Study:

• To develop an integrated framework, AUKAT, for improved kcat estimation using conditional generative modeling and deep neural prediction.
• To address data scarcity and enhance the reliability of kcat predictions.

Main Methods:

• AUKAT utilizes a conditional variational autoencoder to generate synthetic training data, ensuring reliability through an agreement-based selection pipeline.
• A hybrid convolutional neural network and transformer architecture predicts kcat using substrate, enzyme functional, and species embeddings.
• A specialized model, AUKAT-human, was developed using pre-training and fine-tuning for human enzyme kinetics.

Main Results:

• Incorporating synthetic data improved predictive performance for both random forest and neural network models, with significant gains for the neural network architecture.
• AUKAT demonstrated comparable accuracy to DLKcat on standard datasets and improved generalization on unseen subsets, particularly for low-similarity substrates and enzymes.
• Feature importance analysis revealed AUKAT balances substrate, enzyme functional, and species information for prediction.

Conclusions:

• AUKAT offers a scalable approach to enzyme kinetics prediction, effectively mitigating challenges posed by data scarcity in biochemical modeling.
• The framework provides a practical solution for enhancing the accuracy and reliability of kcat predictions in various biological applications.