Fish occupy pivotal trophic positions in aquatic ecosystems, mediating energy transfer and shaping community structure through their feeding interactions. Unraveling these dietary relationships is therefore fundamental for understanding ecosystem functioning and supporting sustainable fisheries management. Traditional morphological analyses, while informative, often fall short in resolving fine-scale prey diversity and trophic linkages. In contrast, DNA metabarcoding has revolutionized dietary studies by enabling comprehensive, high-resolution, and non-invasive characterization of prey assemblages. This review synthesizes recent progress in applying DNA metabarcoding to fish trophic ecology, emphasizing technical innovations, methodological standardization, and ecological insights. We discuss how DNA metabarcoding has advanced the understanding of food web complexity, species interactions, and ecological responses to environmental change. However, challenges persist in quantification accuracy, reference database completeness, and cross-source contamination. Future research integrating multi-marker approaches, standardized workflows, and multi-method integration holds promise for transforming DNA metabarcoding into a powerful, reliable and mechanistic tool for trophic ecology. Collectively, these developments will bridge molecular data with ecological theory, strengthening the scientific foundation for ecosystem-based fisheries management.

Coastal ecosystems, particularly semi-enclosed gulfs, are increasing anthropogenic pressures from urbanization and industrialization with profound impacts on biodiversity maintenance, energy transfer, and biogeochemical cycling. However, how metazoan communities—key components of marine food webs—respond to spatiotemporal variability and human disturbance remains insufficiently understood. This study applied eDNA metabarcoding targeting the mitochondrial COI gene to investigate metazoan communities across 68 stations in the Beibu Gulf, spanning bay, coastal, and island regions, during wet and dry seasons. In total, 878 metazoan ASVs from 13 phyla were detected. Arthropoda dominated both seasons (wet: 85%; dry: 55%), whereas Chordata increased during the dry season (wet: 0.16%; dry: 37%). At the α-diversity level, diversity peaked in the bay region during the dry season and shifted toward the coastal region during the wet season. At the β-diversity level, community composition differed significantly between seasons and spatial regions, with seasonal variation exerting a stronger influence than spatial differentiation. Co-occurrence networks revealed higher complexity during the dry season. β-diversity was overwhelmingly driven by species turnover (94.4%). The island region exhibited the highest community uniqueness, while the human-impacted bay region showed reduced distinctiveness. Redundancy analysis further identified anthropogenically influenced inorganic nitrogen, together with water temperature, transparency, and salinity, as key environmental drivers shaping community structure. βNTI analysis indicated that community assembly was governed by the combined effects of deterministic and stochastic processes. Overall, this study highlights how environmental gradients and human pressures jointly regulate metazoan dynamics, providing insights for biodiversity conservation in human-impacted coastal seas.

The widespread use of antimicrobial biocides and metals has led to the continuous accumulation of biocide and metal resistance genes (BMRGs) in the environment. The issue is of growing concern, as it reduces the efficacy of these agents and poses a potential threat to coastal ecological security. However, the extent of coastal BMRG pollution, its transmission mechanisms, and the influence of seasonal variations on its assembly remain poorly understood. In this study, metagenomic sequencing was employed to investigate BMRGs, microbiomes, and mobile genetic elements (MGEs) within the subtropical nearshore ecosystem of the Beibu Gulf during the autumn and winter seasons. A total of 33 BMRG types and 457 subtypes were detected, with higher subtype diversity in winter than in autumn (440 vs. 326 subtypes). Notably, genes resistant to multi-biocides exhibited the highest diversity, whereas those resistant to both biocides and metals were the most abundant. Co-occurrence network analysis showed that 22 of the 23 detected BMRGs in the winter network were associated with MGEs, especially transposase-related elements such as tnpA. Path modeling indicated that BMRG abundance was more strongly associated with bacterial community composition in autumn, whereas MGE-related variables showed stronger associations in winter. These findings suggest a pronounced seasonal shift in the underlying mechanisms shaping BMRG dynamics, with bacterial communities playing a dominant role in autumn and MGEs playing a more critical role in winter. This seasonal shift highlights the need for season-specific monitoring of BMRGs, coastal pollution control, and resistance-risk management in subtropical coastal ecosystems.