Machine learning models reveal how polycyclic aromatic hydrocarbons influence environmental bacterial communities.

Area of Science:

• Environmental Science
• Microbiology
• Ecotoxicology

Background:

• Polycyclic aromatic hydrocarbons (PAHs) are pervasive environmental pollutants with ecological risks.
• Understanding PAH impacts on diverse bacterial communities (soil, water, sediment) is crucial but challenging.

Purpose of the Study:

• To investigate the effects of PAHs on bacterial community structure and function in different environmental habitats.
• To identify potential PAH-degrading bacteria using machine learning models.

Main Methods:

• Reanalysis of 1924 16S rRNA sequencing samples from various habitats.
• Application of machine learning, specifically the random forest model, to analyze bacterial community shifts and predict degrading species.
• Assessment of bacterial community functions in response to PAH contamination.

Main Results:

• PAH contamination significantly altered bacterial community structures differently across soil, water, and sediment.
• Proteobacteria abundance decreased in soil and sediment but increased in water.
• The random forest model achieved high accuracy (97.72–100%) in identifying PAH effects at the genus level.
• 70 PAH-responsive biomarkers, including potential degraders like Bacillus and Flavobacterium, were identified.
• Bacterial community functions remained largely unaffected by PAH contamination.

Conclusions:

• PAHs induce habitat-specific shifts in bacterial community structure.
• Machine learning effectively identifies bacterial biomarkers and potential PAH degraders.
• The identified microbial candidates are valuable for future bioremediation efforts.
• Bacterial communities exhibit functional stability despite structural changes induced by PAHs.