Selectively Quantify Toxic Pollutants in Water by Machine Learning Empowered Electrochemical Biosensors

Area of Science:

• Environmental Science
• Biosensor Technology
• Machine Learning

Background:

• Electroactive biofilm (EAB) sensors offer high sensitivity for water quality monitoring.
• Detecting multiple toxicants in complex water bodies concurrently remains a significant challenge.

Purpose of the Study:

• To develop an innovative biosensor detection strategy combined with machine learning for simultaneous multi-toxicant identification.
• To create a prediction model (MEA-ANN) for quantifying and qualitatively predicting toxins in multitoxic systems.

Main Methods:

• Developed a machine learning model (MEA-ANN) analyzing electrochemical toxicity response parameters of toxicants (Cd2+, Cr6+, triclosan, trichloroacetic acid).
• Utilized mean impact value to filter characteristic response parameters, enhancing model accuracy and efficiency (OMEA-ANN).
• Validated the model using real and spiked natural water samples.

Main Results:

• The optimized model (OMEA-ANN) showed strong performance in predicting target toxicants within complex interference systems.
• Model validation using real water samples achieved R2 > 0.9, demonstrating practicability and feasibility.
• The strategy effectively expands EAB sensor applicability for multi-toxicant detection.

Conclusions:

• The novel, eco-friendly, and intelligent strategy overcomes limitations of traditional EAB sensors.
• This approach significantly advances water quality monitoring and provides insights for intelligent sewage management.
• Enhanced EAB sensors offer a powerful tool for early ecological risk warnings in aquatic environments.