Predicting antifouling paint particle contamination based on 16S rRNA gene sequencing data using random forest-based...

Area of Science:

• Environmental Science
• Microbiology
• Machine Learning

Background:

• Antifouling paint particles (APPs) impact marine sediment microbial communities.
• Current methods for APP detection are specialized and challenging.
• 16S rRNA amplicon sequencing offers a universal approach for microbial analysis.

Purpose of the Study:

• To develop and validate a machine learning model for predicting APP presence in sediment using 16S rRNA microbial community data.
• To assess the model's applicability in real-world environmental monitoring scenarios.

Main Methods:

• A supervised random forest machine learning model was trained using 16S rRNA amplicon sequencing data from mesocosm experiments.
• The model's predictive performance was evaluated on independent test sets and field samples from the Baltic Sea and Warnow estuary.
• APP presence in field samples was confirmed using scanning electron microscopy-energy-dispersive X-ray spectroscopy.

Main Results:

• The model achieved high accuracy in predicting APP presence (100% in test set) and absence (83.3% in test set).
• The model correctly identified all APP-absent field sites and three out of five APP-contaminated sites.
• The model demonstrated robustness against geographic and seasonal variations in microbial communities.

Conclusions:

• Predicting antifouling paint particle presence in marine sediments using microbial community data is feasible.
• This study provides a proof of concept for machine learning-based predictive tools in environmental monitoring.
• The developed approach offers a blueprint for creating novel pollution detection models.