eDNA-based microbial communities enable rapid screening of sediment pollution.

Microbial communities are increasingly used as bioindicators of coastal ecosystem condition, yet many studies remain limited to descriptive taxonomic analyses. This study investigated relationships between sediment chemistry and eDNA-based microbial community structure across 82 stations in Gyeonggi Bay. Specific links between pollutant concentrations and indicator taxa were identified. Shannon diversity showed no linear trend with contamination levels, whereas the Firmicutes/Proteobacteria ratio was markedly elevated at highly contaminated sites (mean 55.9, up to 136.0) compared with cleaner sites (mean 0.51). In this dataset, these taxonomic shifts provided contaminant-specific signals that the diversity index did not, indicating that taxon-based and diversity-based metrics offer complementary rather than interchangeable information. As these patterns derive from a single regional dataset, their relative utility may vary under other environmental settings. Key indicator families, including Peptostreptococcaceae and Pirellulaceae, were identified using predefined criteria combining statistical significance (p < 0.05), correlation strength (|ρ| > 0.4), and field consistency (p < 0.05). LEfSe analysis further detected 17 genera with significant discriminative power (LDA score > 3.0). Anaerobic Romboutsia and Trichococcus characterized highly contaminated sites, whereas aerobic Woeseia, Blastopirellula, Rhodopirellula, Rubripirellula, and Sulfitobacter characterized stable low-contamination sites. Spatial mapping showed strong correspondence between microbial indicators and SQG-Q gradients. These findings demonstrate that specific microbial taxa can serve as complementary biological indicators of chemical pollution and support their integration into sediment quality assessment frameworks.