(A) Study on New Approach Methods (NAMs) for Water Quality Risk Assessment Based on Multi-Meta-Omics of Daphnia magna

다국어 초록 (Multilingual Abstract)

The increasing occurrence of micropollutants and complex chemical mixtures in freshwater ecosystems poses substantial challenges for conventional water quality monitoring frameworks, which largely depend on target-specific chemical analyses and often fail to detect sublethal, cumulative, or delayed biological effects. These limitations have intensified the need for New Approach Methods (NAMs) capable of providing sensitive, mechanism-informed, and predictive assessments of ecological risk. This dissertation develops, validates, and integrates multiple NAMs—including multi-meta-omics, exo-metabolomics, and deep learning–based behavioral analytics—using Daphnia magna as a sentinel organism to advance effect-based water quality risk assessment. First, an integrative multi-meta-omics platform was established by combining non-targeted LC-QTOF-MS metabolomics with microbial community profiling via DNA metabarcoding. This framework enabled the discovery of coordinated biochemical and microbial shifts under environmental stress. A proof-of-concept application to the edible red alga Pyropia yezoensis demonstrated strong associations between disease severity, dysregulated metabolic pathways (e.g., purine metabolism), and restructuring of surface microbiota—including pathogen- associated taxa such as Pythium. These results validate the capability of multi- omics indicators to capture early ecological disturbances that may not be detectable by conventional monitoring methods. Second, this dissertation evaluates exo-metabolomics as a non-invasive and time- resolved monitoring strategy within BEWS through controlled copper exposures and river-water experiments using a flow-through chamber system. A total of 1,228 extracellular metabolic features were detected, with 736 significantly enriched in the presence of D. magna, confirming active release of exo-metabolites. PCA distinguished Daphnia-exposed water from controls and revealed temporal structuring linked to reproductive cycles. Copper treatments induced clear metabolic shifts across pre-, exposure-, and post-exposure periods, producing reversible (1–2–1) and persistent (1–2–2) response patterns. Concentration– response analyses further identified 129 copper-associated metabolites and 86 features strongly correlated with exposure levels, including 12 high-confidence candidates displaying dose-dependent trajectories. Notably, sublethal metabolic responses were detectable at concentrations as low as 5 μg/L—well below lethal thresholds—demonstrating the superior sensitivity of exo-metabolomics and its compatibility with early-warning, non-destructive monitoring frameworks in complex aquatic environments. Third, this dissertation advances exo-metabolomics as a non-invasive and dynamic monitoring tool by characterizing the extracellular metabolite profiles of D. magna exposed to sublethal copper concentrations and natural river water. Across two field sites with distinct physicochemical properties, 2,440 metabolic features were detected, and hundreds displayed site-specific, concentration- dependent, and temporally reproducible patterns. Metabolite responses included reversible trajectories (1–2–1 patterns) indicative of organismal recovery, as well as persistent disturbances (1–2–2 patterns) reflective of lasting toxic impairment. Several metabolite markers demonstrated strong correlations with copper concentration and partial reproducibility across independent studies, supporting their utility as early-warning indicators suitable for BEWS deployment and for identifying molecular points of departure (PODs). Fourth, a deep learning–based behavioral monitoring framework was developed to quantify subtle yet ecologically meaningful changes in D. magna swimming behavior. The YOLOv5-based tracking model enabled automated, high-resolution behavioral extraction, overcoming the limitations of manual observation and improving reproducibility under complex environmental conditions. When behavioral endpoints were compared with exo-metabolomic indicators, the combined evidence strengthened diagnostic confidence and illustrated the complementary nature of molecular and organism-level responses under toxic exposure. This integration supports the development of next-generation, AI- enhanced BEWS capable of real-time and high-throughput early-warning functions. Collectively, the findings of this dissertation provide a comprehensive demonstration of how multi-meta-omics, exo-metabolomics, and AI-assisted behavioral analytics can be harnessed as NAMs to detect sublethal toxicity, characterize mixture effects, and reveal mechanistic pathways of stress responses in freshwater environments. By offering mechanistic interpretability, concentration-dependent biomarkers, and automated behavioral assessment, these NAM-based approaches address key gaps in existing water quality monitoring frameworks and support a shift toward proactive, predictive, and effect-based regulatory strategies. Future research should expand field validation across seasons and river basins, improve metabolite annotation pipelines, and develop standardized biomarker libraries to facilitate operational deployment in national water governance systems.

목차 (Table of Contents)

• 1. Introduction 1
• 1.1 Background and Rationale 1
• 1.1.1 Sources and pathways of micropollutants in aquatic environments 1
• 1.1.2 Limitations of conventional chemical-based water quality management 4
• 1.1.3 Combined exposure and delayed adverse effects on aquatic organisms and human health 6
• 1.1.4 Need for effect-based early warning indicators using multi-meta-omics approaches 7
• 1.2 Literature Review 10
• 1.2.1 Challenge in early warning of aquatic micropollutants 10
• 1.2.2 NAMs in Risk Assessment and Environmental Monitoring 12
• 1.2.3 Meta-omics technologies 14
• 1.2.4 Trends in integrative multi-omics research 16
• 1.3 Research Aims and Scope 17
• 1.3.1 Research questions and hypotheses 17
• 1.3.2 Scope and structure of the dissertation 18
• 1.3.3 Academic significance and expected contributions 19
• 2. Development of Meta-metabolomics and DNA Metabarcoding Integration Methodology 20
• 2.1 Research Objectives and Methodology 20
• 2.1.1 Overall research design and framework 20
• 2.1.2 Sample collection and preprocessing 21
• 2.1.3 Development of microbiome DNA metabarcoding techniques 28
• 2.1.4 Establishment of meta-metabolomics platform (non-targeted mass spectrometry) 29
• 2.1.5 Data preprocessing and multi-omics integration strategy 30
• 2.2 Results and Discussion 33
• 2.2.1 Microbial community profiling results 33
• 2.2.2 Metabolite profiling from environmental samples (e.g., aquatic plants, P. yezoensis) 56
• 2.2.3 Discovery of metabolite and microbial indicators through integrative analysis 63
• 2.2.4 Discussion: Methodological advantages and comparison with conventional approaches 67
• 2.3 Conclusion and Future Perspectives 70
• 2.3.1 Summary of findings 70
• 2.3.2 Limitations and future directions 70
• 3. Exo-metabolomic Profiling of Daphnia magna for Biomarker Discovery 72
• 3.1 Research Objectives and Methodology 72
• 3.1.1 Research Background and Rationale 72
• 3.1.2 Research Objectives 73
• 3.1.3 Preparation of materials and test organisms 74
• 3.1.4 Exposure experiments and sampling of Daphnia exo-metabolites 75
• 3.1.5 Sample Preparation and Exo-metabolite Collection 80
• 3.1.6 Analytical Platform and Data Acquisition 80
• 3.1.7 Data Processing and Statistical Analyses 81
• 3.2 Results and Discussion 82
• 3.2.1 Presumed exo-metabolome characteristics of Daphnia 82
• 3.2.2 Regulation of Daphnia magna exo-metabolites under copper ion exposure 85
• 3.2.3 Concentration–response relationships between copper exposure and exo-metabolites of Daphnia magna 91
• 3.2.4 Biological Interpretation of Biomarker Responses 99
• 3.2.5 Comparison with Previous Studies 99
• 3.2.6 Implications for Non-invasive Water Quality Assessment 100
• 3.3 Conclusion and Future Perspectives 101
• 3.3.1 Summary of Key Findings 101
• 3.3.2 Strengths and Limitations of Exo-metabolomics 101
• 3.3.3 Future Research Directions 101
• 4. Non-invasive exo-metabolomics reveals candidate early-warning indicators across multiple river sites 103
• 4.1 Research Objectives and Methodology 103
• 4.1.1. Preparation of materials and test organisms 103
• 4.1.2. Exposure experiments in the BEWS system 106
• 4.1.3 Extraction and Global Profiling of the Exo-metabolome of Daphnia 110
• 4.1.4. Data Processing and Statistical Analyses of the Global Metabolome 111
• 4.2 Results and Discussion 113
• 4.2.1. Detection of exo-metabolites of Daphnia magna consistently observed across two river sites 113
• 4.2.2. Shared temporal metabolomic responses before, during, and after exposure in both rivers 120
• 4.2.3. Copper concentration-dependent metabolite alterations and comparison with previous findings 124
• 4.2.4 Implications for exo-metabolomics-based BEWS applications 136
• 4.3 Conclusion and Future Perspectives 139
• 4.3.1 Summary of validation outcomes 139
• 4.3.2 Future research and implementation perspectives 140
• 5. Deep Learning Applications to Daphnia magna Behavioral Indicators for Early Warning 141
• 5.1 Research Objectives and Methodology 141
• 5.1.1 Current status and limitations of behavior-based early warning systems 141
• 5.1.2 Development of a high-resolution behavioral monitoring platform using Daphnia magna 142
• 5.1.3 Deep learning approaches for behavioral data analysis and pattern recognition 146
• 5.1.4 Behavioral parameter extraction and change point detection 147
• 5.1.5 Composite behavioral parameter selection 148
• 5.2 Results and Discussion 150
• 5.2.1 Quantification of subtle behavioral changes under pollutant exposure 150
• 5.2.2 Behavioral responses of Daphnia magna to chemical exposure 153
• 5.2.3 Selection of Behavioral Indicators and Response Criteria 162
• 5.2.4 Optimization of Composite Behavioral Parameter Combinations 166
• 5.2.5 Discussion: Environmental Context and Site-Specific Variability 169
• 5.3 Conclusion and Future Perspectives 171
• 5.3.1 Key findings 171
• 5.3.2 Future perspectives 171
• 6. General Conclusion 173
• 6.1 Overall summary of research findings 173
• 6.2 Integrated implications for water quality monitoring and risk assessment 175
• 6.3 Academic contributions and novelty 176
• 6.4 Future research outlook 177
• References 178