Accumulation
Millions of tons of PET accumulate in landfills and oceans, while natural degradation can take centuries.
This is a crisis in the making
A Study on the Rational Evolution and Degradation Mechanism of PET Hydrolases Using Graph-Based Learning
We combine structure-aware AI, experimental validation, and mechanistic diagnosis to engineer PET hydrolases for more efficient plastic biodegradation.
What is at stake?
PET waste is everywhere
01PET waste
02AI screening
03Enzyme design
04Mechanism
Microplastics
PET waste breaks into microplastics that can spread through marine ecosystems and food chains.
Recycling gap
Traditional physical and chemical recycling can be energy-intensive and may create secondary pollution.
The Bottlenecks
From waste to resource, PET degradation faces critical bottlenecks
Despite advances in biocatalysis, industrial-scale plastic recycling is hindered by severe limitations.
Fragile at High Temps
Weak Natural Activity
Enzyme
Limits
Limits
Process Limits
Inefficient Screening
Intermediate Accumulation
Current methods are too slow for a fast-growing crisis
AI-driven rational design unlocks PET hydrolase potential
We replace random mutations with Graph Neural Networks for precise active-site engineering. Furthermore, we deploy a synergistic dual-enzyme system (PETase + BHETase) to ensure the complete conversion of plastic waste into valuable TPA.
Key results & impact
A scalable route toward more efficient PET biodegradation
Our AI-guided framework integrates graph learning, meta-learning, experimental validation, and mechanistic simulation to engineer PET hydrolases with improved catalytic performance and thermal robustness.
Engineered variants achieved up to 5.5x higher hydrolytic activity than their wild-type counterparts at 40-70C. An ML-optimized PETase/BHETase dual-enzyme system further improved intermediate conversion, increasing TPA yield by up to 1.5x.
Let's explore PIM-NEBS