DeltaCata: Geometric Deep Learning for Enzyme Mutation Effects

Nature Communications China

Overview

Researchers have developed DeltaCata, a Siamese geometric graph neural network that predicts changes in enzyme kinetic parameters induced by residue substitutions. The tool achieved up to a 7.0-fold improvement in catalytic efficiency for in vitro icaritin synthesis using the methyltransferase sGmOMT2. This approach enables more effective industrial enzyme engineering by focusing on mutation effects rather than absolute values.

In Depth

Key Findings

DeltaCata demonstrates substantial performance improvements over existing methods in predicting mutation-induced changes in kinetic parameters, particularly for enzymes not present in the training sets. In a practical application to engineer the methyltransferase sGmOMT2, the model achieved up to a 7.0-fold improvement in catalytic efficiency for the in vitro synthesis of icaritin. The utility of the system was further validated through in silico beneficial mutation screening using deep mutational scanning datasets.

Technical Details

The researchers developed a high-quality enzyme mutation effect dataset by pairing wild-type and mutant kinetic parameters measured under identical experimental conditions. This dataset was created using a semi-automatic procedure that combined manual verification with language model-assisted parsing. DeltaCata utilizes a Siamese geometric graph neural network to directly model the changes in turnover numbers (Δk cat) and Michaelis constants (ΔK m) resulting from residue substitutions, leveraging protein structural geometry to improve accuracy.

Background & Context

Quantifying how mutations affect enzyme kinetic parameters is critical for accelerating enzyme evolution. Previous computational methods focused on predicting the absolute values of k cat and K m, but these approaches often failed to accurately capture the specific effects of mutations. A primary obstacle to developing a direct mutation-effect model has been the scarcity of training data that pairs wild-type and mutant parameters under consistent conditions.

Significance & Outlook

By shifting the focus from absolute parameter prediction to the modeling of mutation effects (Δk cat and ΔK m), DeltaCata addresses a significant gap in enzyme engineering. The ability to accurately predict the impact of residue substitutions on catalytic efficiency suggests high potential for the tool’s application in industrial enzyme engineering, where optimizing enzyme performance is essential for efficient chemical synthesis.

Get our weekly technology intelligence — free

Receive an infographic that lets you judge at a glance whether each field’s analysis report is worth reading.

Subscribe Free — Weekly Tech Intelligence

By subscribing, you’ll receive Troy-Technical’s weekly technology intelligence newsletter.

• Your email and selected fields are used only to deliver the newsletter.
• We never share your information with third parties.
• You can unsubscribe anytime via the link in each email.

See our Privacy Policy for details.

Agree & Continue ▶

Takes about a minute · Unsubscribe anytime