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

  "parent": "V120",

  "mutations": ["S238F", "T140A"],

  "G_folding": -7.8,

  "G_binding": -8.4,

  "S_dyn": 0.31,

  "score_total": 0.88}

3. Mutation Simulation Mechanics

Mutation operators: Point mutations at hotspots (e.g., binding cleft residues, flexible loops). Insertions/deletions (rare, with high penalty unless stabilizing). Directed mutations based on RL policy gradient.

Trajectory depth: ~200 generations

Population size: 500--1000 variants per generation

Exploration strategy: -greedy with episodic random seeding