Microenvironment Awareness Pragmatic Analysis for Generative Enzyme Function Reasoning.

MPA-EC, a generative enzyme function reasoning framework incorporating microenvironment context, improves F1 score by 0.1068 over the state-of-the-art GraphEC model on benchmark data.

• Why it matters: Accurate enzyme function annotation is essential for interpreting expanding sequence data, but existing computational methods struggle to generalize to new enzymatic activities and often ignore microenvironment factors like pH and temperature, which are crucial for functional inference.
• What they did: The study developed MPA-EC by fine-tuning a large language model on a newly created Enzyme Microenvironment Database, enabling it to infer microenvironment from sequence and generate EC labels with reasoning chains, imputing missing pH and temperature values as needed.
• The result: MPA-EC demonstrates practical utility by generating reference-worthy function labels under cold-start conditions, advancing enzyme annotation for emerging activities and providing accessible source code and data for further research.