Artificial intelligence is filtering through millions of unexplored biological databases to help scientists identify naturally occurring enzymes capable of breaking down stubborn plastics, per- and polyfluoroalkyl substances (PFAS), and other persistent environmental pollutants, according to a Nanowerk report published on October 4, 2026.
Murdoch University Machine Learning Pipelines Target Environmental Pollutants
Scientists at Murdoch University’s Bioplastics Innovation Hub are utilizing machine learning tools to sort through millions of data entries containing potential enzyme candidates. Manually sorting through this vast digital library would present an insurmountable task for human researchers. Instead, automated pipelines leverage data from previously characterized enzymes to predict how specific biological structures interact with synthetic target pollutants. Joseph Boctor, a PhD candidate working with the hub, outlined these processes in a recent review published in Nature Reviews Earth & Environment, titled “Using machine learning with biochemical analysis to identify suitable enzymes for bioremediation applications.”
I strongly advocate that overengineering enzymes is a bad starting point that overlooks millions of years of evolution that have already produced lots of potential solutions to these contaminants,
Rather than engineering synthetic proteins from scratch, the research team focuses on mining existing biological datasets to discover functional matches.
The Ubiquity of Microplastics in Agricultural Soils and Food Supplies
Persistent industrial compounds present distinct challenges because they are both durable and biologically active. These contaminants often mimic natural human hormones, creating documented disruptions to overall health. Beyond industrial persistence, agricultural soils have become major accumulation zones for these synthetic materials. A comprehensive review conducted by Mr Boctor during the prior year revealed that agricultural soils hold roughly 23 times more microplastics than the world’s oceans. Studies have detected microplastics and nanoplastics inside common food crops, including lettuce, wheat, and carrots. Other hazardous soil additives identified in the research include phthalates, which are linked to reproductive issues, alongside polybrominated diphenyl ethers (PBDEs), which act as neurotoxic flame retardants associated with neurodegenerative diseases, increased risks of stroke and heart attack, and premature mortality.
Machine Learning Accelerates the Scaling of Bioremediation Enzymes
While separate engineering groups continue developing future bioplastic alternatives, machine learning algorithms are accelerating the remediation of pollutants already contaminating natural ecosystems. After computational models narrow down millions of enzyme candidates to the most promising biological profiles, the research team transitions from digital screening to physical testing. The primary focus of the Bioplastics Innovation Hub now centers on testing, validating, and scaling these identified enzymes to ensure they can operate effectively in real-world bioremediation environments.
