About This Project

We propose a generative model to accomplish Programmable Enzymatic Degradation of plastics. The ultimate goal is to input a SMILES notation of any plastic, and return a novel functional enzyme sequence capable of degrading that plastic. Our model allows for de-novo design of enzymes for polymers without naturally occurring degradation enzymes, such as PP and PE. Resultant enzymatic “hits” will then undergo experimental validation to confirm their degradation efficacy.

Ask the Scientists

What is the context of this research?

Plastic has become a ubiquitous symbol of human advancement and paradoxically, our environmental shortcoming. The world produces over 300 million tons of plastic yearly, a figure that's steadily rising. Plastic pollution poses one of the most insidious threats to the environment and marine ecosystems. Originating from a variety of sources including single-use plastics, packaging, and discarded goods, these plastic pollutants persist in the environment for centuries, leaching harmful chemicals, exacerbating climate change, disrupting food chains, and physically harming aquatic life. Methods like recycling and waste management fall short due to economic, infrastructural, and scalability challenges. A transformative solution is essential—targeting plastics at their molecular level.

What is the significance of this project?

Nature's enzymes present a sustainable avenue to combat plastic pollution. Recent advances have seen progress in targeting polymers such as polyethylene terephthalate (PET), polyurethane (PUR), and polyamide (PA), with a total of 127 naturally-occuring degradation enzymes. However, these polymers account for only 14% of annual plastic waste. In stark contrast, polyethylene (PE) and polypropylene (PP), which represent a whopping 49% (193 million metric tons) of global plastic waste, are characterized by only four enzymes. Our focus must pivot to designing enzymes that degrade these polymers to truly make a dent in the plastic crisis. Therefore, we propose a generative model that designs novel enzymes tailored to degrade any plastic polymer, allowing us to degrade PP and PE.

What are the goals of the project?

Using deep learning and generative AI, we will:

1. Create a model in which we input any plastic’s polymer structure and design an enzyme able to degrade the respective plastic. This will allow us to design enzymes to degrade plastics, like PP, which has zero biochemically characterized degrading enzymes.

2. Curate a publicly available database of existing plastic structures and current/generated enzyme sequence pairs, along with information on the enzyme’s stability and degradation confidence score.

3. Create an experimental screening pipeline to validate our designed enzymes in the wet-lab.

Attaining these goals will lay the groundwork for a robust method to produce enzymes fit for degrading various plastics.