a CAS Framework for Predicting the Synthetic Evolution of Anti-Plastic Enzymes

Actions correspond to mutation operations:

Point mutation at residue i with amino acid j

Segmental crossover (if modular evolution is enabled)

Structural fine-tuning (e.g., loop reshaping or bond reorientation)

Each action alters the protein's genotype, triggering re-evaluation by folding and functional modules.

c. Reward Function

The reward balances multiple objectives:

R=1Fitness+2Stability3Mutation CostR = \lambda_1 \cdot \Delta \text{Fitness} + \lambda_2 \cdot \Delta \text{Stability} - \lambda_3 \cdot \text{Mutation Cost}

Where:

Fitness reflects improvement in catalytic efficiency or substrate affinity.

Stability measures structural integrity changes.