Non-invasive continuous biological monitoring method based on non-targeted environmental metabonomics

Abstract

The invention discloses a non-invasive continuous biological monitoring method based on non-targeted environmental metabonomics, and belongs to the technical field of environmental monitoring and metabonomics. Comprising the following steps: sample collection: randomly dividing zebrafish into a control group and a pollutant exposure group, and collecting male and female whole zebrafish as a tissue sample; metabolite extraction: effectively extracting the metabolite in the water sample and the zebra fish tissue sample through solid-phase extraction and concentration; data acquisition: performing characteristic peak detection and acquisition on metabolites in the sample by adopting LC-MS (Liquid Chromatography-Mass Spectrometer); preprocessing data: extracting, aligning, denoising, background removing and annotating metabolic characteristic peaks by adopting MS-DIAL software; and data analysis: carrying out classification identification, differential metabolite screening, pathway enrichment and PLS-DA on the metabolic data, and carrying out dynamic response monitoring on the core metabolite. The non-invasive continuous biological monitoring method has the advantages of being non-invasive, continuous, high in flux, low in cost and the like, and dynamic monitoring of the physiological response process of the non-invasive continuous biological monitoring method is achieved.

Description

Technical Field

[0001] This invention belongs to the field of environmental monitoring and metabolomics technology, specifically relating to a non-invasive continuous biological monitoring method based on non-targeted environmental metabolomics. Background Technology

[0002] Traditional environmental toxicology studies widely employ laboratory animals (such as zebrafish, mice, and nematodes) for pollutant exposure experiments, assessing the toxic effects of pollutants on organisms by analyzing changes in their tissues, body fluids, or behavior. However, this method generally suffers from problems such as large sample size requirements, long experimental cycles, complex procedures, high costs, and harm to laboratory animals. Furthermore, since laboratory animals can usually only be sampled at a single time point, the data obtained is mostly static information, making it difficult to capture the dynamic characteristics of the entire metabolic process, thus limiting in-depth analysis of pollutant exposure mechanisms.

[0003] In recent years, various alternative research methods have emerged, such as in vitro cell models, organoid culture, molecular docking, computational toxicology simulations, and machine learning-based toxicity prediction methods, which have shown certain advantages in improving research efficiency and reducing the use of laboratory animals. However, these methods generally suffer from problems such as a lack of complete physiological background, weak biological relevance, and strong dependence on large-scale training data, making it difficult to comprehensively and accurately reflect the actual metabolic response process of organisms under complex environmental conditions. Therefore, the field of environmental toxicology urgently needs a highly efficient biomonitoring technology that can dynamically and continuously monitor the metabolic state of animals without interfering with their normal physiological metabolism.

[0004] With the development of high-throughput detection technologies and systems biology, omics technologies have become important tools for studying the overall characteristics of living systems, encompassing multiple levels such as genomics, transcriptomics, proteomics, and metabolomics. Among them, metabolomics, through comprehensive qualitative and quantitative analysis of small molecule metabolites in biological systems, especially non-targeted metabolomics, can systematically capture the entire metabolic network without pre-defined targets, sensitively reflecting the overall response of organisms to environmental stresses or physiological disturbances. As the omics level closest to the phenotype, metabolomics has wide-ranging applications in areas such as early disease screening, toxicological mechanism analysis, pollutant effect monitoring, and ecological health assessment. Summary of the Invention

[0005] Zebrafish, as a widely used aquatic model organism, produce small-molecule metabolites during their metabolism that can naturally diffuse into the aquaculture water, forming analyzable environmental metabolic signals. Based on this, the purpose of this invention is to provide a non-invasive, continuous biomonitoring method based on non-targeted environmental metabolomics. Through systematic analysis of non-targeted metabolites in zebrafish aquaculture water, this method indirectly reflects the endogenous metabolic state of the organism, enabling dynamic monitoring of its physiological response processes. This provides a scientific, feasible, and sustainable technical approach for studying the effects of pollutant exposure and assessing environmental health risks in aquatic ecosystems.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] A first aspect of the present invention provides a non-invasive continuous biomonitoring system based on non-targeted environmental metabolomics, comprising a sample collection module, a metabolite extraction module, a data acquisition module, a data preprocessing module, and a data analysis module; wherein:

[0008] The sample collection module is used to randomly divide zebrafish into a control group and a pollutant exposure group, and a continuous drip water exchange system is used to change the exposure solution daily. During the exposure process, 1.5L of aquaculture water sample is collected every 24 hours. At the end of the exposure, whole male and female zebrafish are collected as tissue samples.

[0009] The metabolite extraction module is used to effectively extract metabolites from water samples and zebrafish tissue samples.

[0010] The data acquisition module uses LC-MS to detect and acquire characteristic peaks of metabolites in water samples and zebrafish tissue samples.

[0011] The data preprocessing module uses MS-DIAL (v4.9) software to extract, align, denoise, remove background and annotate metabolic feature peaks, and corrects instrument signal drift using the QC-RFSC method in the R package statTarget. Feature peaks with poor stability are removed based on the standard of coefficient of variation (CV) <30% to ensure data quality.

[0012] The data analysis module uses Class Fire, Metabo Analyst, R language, and Origin tools to classify and identify processed metabolic data, screen for differential metabolites, enrich pathways, and perform partial least squares discriminant analysis (PLS-DA). Based on time series, it identifies key metabolites that play a dominant role in metabolic differences at each time point and dynamically monitors the response of core metabolites that repeatedly appear in pathway enrichment at multiple time points and represent the core regulation of the metabolic network. This allows for in-depth exploration of the metabolic characteristics and potential biological significance of water samples and zebrafish tissue samples.

[0013] Preferably, the continuous drip water exchange system adopts a closed-loop control method to dynamically balance and stabilize the concentration of the exposure solution during the renewal process. The pre-prepared exposure solution is continuously delivered to the zebrafish breeding tank at a set flow rate by a peristaltic pump to ensure that the supply process is stable and controllable. At the same time, the other end of the breeding tank is connected to a drain pipe to discharge the solution at the same speed and synchronously. The volume of the discharged liquid is consistent with the volume of the inlet water, so as to maintain a constant total amount of liquid in the breeding tank.

[0014] A second aspect of the present invention provides a non-invasive continuous biomonitoring method based on non-targeted environmental metabolomics, comprising the following steps:

[0015] S1. Sample collection; including: randomly dividing zebrafish into a control group and a pollutant exposure group, using a continuous drip water exchange system, and changing the exposure solution daily; during the exposure process, collecting 1.5L of aquaculture water samples every 24 hours; at the end of the exposure, collecting whole male and female zebrafish as tissue samples;

[0016] S2. Metabolite extraction; including: effective extraction of metabolites from water samples and zebrafish tissue samples using solid-phase extraction and concentration methods;

[0017] S3. Data Acquisition; including: using LC-MS to detect and acquire characteristic peaks of metabolites in water samples and zebrafish tissue samples;

[0018] S4. Data preprocessing; including: using MS-DIAL software to extract, align, denoise, remove background and annotate metabolic characteristic peaks, and using the QC-RFSC method in the R package statTarget to correct instrument signal drift, and removing characteristic peaks with poor stability according to the standard of CV<30% to ensure data quality;

[0019] S5. Data Analysis: This includes classifying and identifying processed metabolic data, screening for differential metabolites, enriching pathways, and performing PLS-DA based on Class Fire, Metabo Analyst, R language, and Origin tools. It also involves identifying key metabolites that play a dominant role in metabolic differences at each time point based on time series data, and dynamically monitoring the core metabolites that repeatedly appear in pathway enrichment at multiple time points and represent the core regulation of the metabolic network.

[0020] Preferably, in step S1, the continuous drip water exchange system uses a closed-loop control method to control the dynamic balance and concentration stability of the exposure solution renewal process. The exposure solution is continuously delivered to the zebrafish breeding tank at a set flow rate by a peristaltic pump to ensure that the supply process is stable and controllable. At the same time, the other end of the breeding tank is connected to a drain pipe to discharge the solution at the same speed and synchronously. The volume of the discharged liquid is consistent with the volume of the inlet water, so as to maintain a constant total amount of liquid in the breeding tank.

[0021] Preferably, in step S2, the metabolite extraction includes: first filtering the water sample through a 0.22μm glass fiber membrane to remove suspended particles and microbial impurities, and then enriching the metabolites using an HLB solid-phase extraction column.

[0022] More preferably, step S2, the metabolite extraction step, includes:

[0023] For water sample processing, the water sample was loaded into a pre-activated HLB column. After extraction, the residual liquid in the column was dried with high-purity nitrogen. The sample was then washed with 10 mL of LC-MS grade ultrapure water and eluted with 15 mL of LC-MS grade methanol. The eluent was dried with nitrogen and concentrated to near-dry state. It was then reconstituted with 150 μL of acetonitrile-water mixture at a volume ratio of 1:1 and transferred to a glass sample vial with an inner support tube for storage.

[0024] For zebrafish tissue sample processing, male and female zebrafish were mixed in equal proportions and thoroughly ground. Equal amounts of mixed tissue samples were taken from each group and 1 mL of a mixed extraction solution with a volume ratio of methanol:acetonitrile:water = 4:4:2 was added. After ultrasonic extraction in an ice-water bath for 10 min, the samples were allowed to stand at -20℃ to precipitate. Then, the samples were centrifuged at 13000 rpm and 4℃ for 15 min. The supernatant was collected and concentrated to dryness by vacuum centrifugation. The samples were then reconstituted with 150 μL of an acetonitrile-water mixture with a volume ratio of 1:1 and transferred to glass sample vials for storage.

[0025] Preferably, in step S3, LC-MS is used to detect and collect characteristic peaks of metabolites in water samples and zebrafish tissue samples, covering both hydrophilic interaction chromatography (HILIC) and reversed-phase chromatography (RPLC) modes; wherein:

[0026] The HILIC mode parameters are as follows: a Waters BEH amide column (1.7 μm, 2.1 × 100 mm) was used, with a column temperature of 45 °C; mobile phase A was an aqueous solution containing 25 mM ammonium acetate and ammonium hydroxide, and mobile phase B was acetonitrile; the flow rate was 0.5 mL / min; the gradient was set as follows: phase B decreased from 95% to 65% within 0-7 min, decreased to 40% within 1 min, and was held for 1 min before returning to the initial conditions;

[0027] The RPLC mode parameters are as follows: using Phenomenex C18 column (2.6 μm, 2.1 × 100 mm), column temperature 25 °C; mobile phase A was an aqueous solution containing 0.01% acetic acid, and mobile phase B was isopropanol:acetonitrile (1:1, v / v); flow rate was 0.3 mL / min; gradient: phase B increased from 1% to 99% within 0-7 min, retained for 1 min, and then returned to the initial conditions.

[0028] Preferably, in step S4, MS-DIAL (v4.9) software is used to extract, align, denoise, remove background and annotate metabolic characteristic peaks. The steps include: before importing into MS-DIAL, the raw mass spectrometry data file (.raw) is converted to ABF format (.abf) using Abfconverter ver.1.3.

[0029] The key parameters in MS-DIAL are set as follows: the quality tolerances of MS1 and MS2 are 0.01 Da and 0.025 Da, respectively; the acquisition time range is 0.5-13.5 minutes, the mass-to-charge ratio (m / z) range is set to 80-1200 Da; the quality slice width is 0.05 Da; the minimum peak height threshold is set to 20000, and the peak width filtering threshold is set to 5 scans to remove background interference and low-quality signals, thereby improving the accuracy and stability of peak identification.

[0030] Metabolite annotation was performed based on MS spectrum matching, which was achieved by comparing the characteristic ion peaks in the experimental samples with the MS1 and MS2 spectra recorded in the real standard database.

[0031] The standard database used is the total database provided by MS-DIAL, including: the MS / MS database in positive ion mode (ESI(+)-MS / MS from authentic standards (16232 unique compounds) and the MS / MS database in negative ion mode (ESI(-)-MS / MS from authentic standards (8887 unique compounds)). During the matching process, the precise mass tolerances of MS1 and MS2 were set to 0.01 Da and 0.05 Da, respectively, to ensure the accuracy and reliability of metabolite annotation.

[0032] As a preferred option, in step S4, the correction of instrument signal drift includes: using the QC-RFSC algorithm in the statTarget software package under the R language environment, taking the quality control (QC) samples injected at intervals during the experiment as a reference, fitting and correcting the signal change trend of each metabolic feature in different sample injection sequences, and eliminating systematic errors caused by batch effects or signal fluctuations generated during instrument detection.

[0033] Based on the CV < 30%, feature peaks with poor stability were removed as follows: the relative standard deviation (CV value) of each metabolic feature in the QC sample was calculated based on the signal intensity of each feature in multiple tests, and feature variables with high data stability were selected and retained with a threshold of 30%; feature peaks with CV ≥ 30% were regarded as having large detection fluctuations and poor reproducibility and were removed to reduce data noise and false positives.

[0034] Preferably, in step S5, the classification and identification are implemented using the Classy Fire tool. The standardized structural identifiers InChIKey corresponding to the metabolites obtained through mass spectrometry annotation are input into the Classy Fire system, and the built-in chemical ontology system Chem Ont is used to automatically classify the metabolites structurally. Based on the molecular structural features represented by InChIKey and combined with a rule-driven classification algorithm, Classy Fire identifies the superclass, class, and subclass chemical classification information of metabolites layer by layer, thereby realizing the systematic classification of various metabolites at the structural level.

[0035] Preferably, in step S5, the differential metabolite screening is performed by the StatisticalAnalysis module in the Metabo Analyst platform. The preprocessed metabolomics data is imported into this module, and the significance of metabolic characteristic peaks between different groups is tested using its built-in statistical methods. The fold change (FC) of metabolite expression between groups is then used for comprehensive evaluation. The screening criteria for differential metabolites are set as follows: FC ≥ 1.5 or ≤ 0.67, and the corresponding p value < 0.05. Metabolites that meet this criterion are determined to be significantly differentially expressed between the exposure group and the control group and are included in the biological function analysis and metabolic pathway analysis.

[0036] Preferably, in step S5, the pathway enrichment analysis is performed using the PathwayAnalysis module in the Metabo Analyst platform. The screened differential metabolites are imported into this module, and metabolic pathway matching and enrichment analysis are performed using the corresponding species' metabolic pathway database. The system outputs potential metabolic pathways based on the coverage and enrichment level of metabolites in each pathway, combined with topological structure and statistical significance. Among them, pathways with p-value < 0.05 are identified as significantly enriched pathways.

[0037] Preferably, in step S5, the partial least squares discriminant analysis is performed by associating fish metabolomics and aquatic metabolomics data with sample grouping information to model the difference between the two in different treatment groups and their consistency.

[0038] As a preferred option, in step S5, key metabolites that play a dominant role in metabolic differences at each time point are identified based on time series data. Multivariate statistical analysis is performed on metabolomics data collected at different time points within the exposure period. PLS-DA is used to calculate the variable importance projection value (VIP) of each metabolite, and the trend of VIP value changes of each metabolite at each time point is dynamically displayed.

[0039] Preferably, in step S5, the core metabolite is a metabolite that repeatedly participates in key signaling pathways at multiple time points. Monitoring its dynamic response includes: conducting signal intensity tracking analysis at multiple time points, continuously monitoring the dynamic changes of the core metabolite in the water body, and real-time assessment and monitoring of metabolite response, fish physiological state, and water environment conditions.

[0040] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0041] This invention proposes a non-invasive, continuous biomonitoring method based on non-targeted environmental metabolomics, which has the advantages of being non-invasive, continuous, high-throughput, and low-cost. It can achieve dynamic monitoring of the physiological response process of experimental animals without sampling or harming them. It provides a scientific, feasible, and sustainable new technical approach for the study of pollutant exposure effects and environmental health risk assessment in aquatic ecosystems, and is applicable to fields such as environmental pollution biological response detection, aquatic ecotoxicology research, and environmental health risk assessment. Attached Figure Description

[0042] Figure 1 This is a schematic diagram of the workflow of a non-invasive continuous biomonitoring system based on non-targeted environmental metabolomics in the embodiment.

[0043] Figure 2 This is a schematic diagram of the metabolite extraction module in the embodiment.

[0044] Figure 3 This is a schematic diagram of the data preprocessing module in the embodiment.

[0045] Figure 4 This is a schematic diagram of the data analysis module in the embodiment.

[0046] Figure 5 This is a schematic diagram showing the categories and proportions of metabolites in the water metabolome and fish metabolome at the exposure endpoint in the examples.

[0047] Figure 6 The PLS-DA plots of the water metabolome and fish metabolome at the exposure endpoint are shown in the examples.

[0048] Figure 7 The results show the intersection of significantly enriched signaling pathways in the water metabolome and fish metabolome at the exposure endpoint in the examples.

[0049] Figure 8 The PLS-DA diagrams for each treatment group in the water metabolome during the exposure process are shown in the examples.

[0050] Figure 9 The PLS-DA plots are shown at various time points of the water metabolome during the exposure process in the examples.

[0051] Figure 10This is a graph showing the dynamic changes in the response intensity of the core metabolite at different time points in the example. Detailed Implementation

[0052] To more fully understand and demonstrate the technical solutions, objectives, and advantages of the present invention, the technical effects produced by the present invention will be further described in detail and completely below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. It should be noted that other embodiments obtained by those skilled in the art without departing from the concept of the present invention are all within the protection scope of the present invention.

[0053] like Figure 1-4 As shown in the following embodiments, a non-invasive continuous biomonitoring system based on non-targeted environmental metabolomics is proposed, including a sample collection module, a metabolite extraction module, a data acquisition module, a data preprocessing module, and a data analysis module; wherein:

[0054] The sample collection module was used to randomly divide zebrafish into a control group and a pollutant exposure group. A continuous drip water exchange system was used to change the exposure solution daily. During the exposure process, 1.5L of culture water samples were collected every 24 hours. At the end of the exposure, whole male and female zebrafish were collected as tissue samples. The continuous drip water exchange system adopted a closed-loop control method to maintain the dynamic balance and concentration stability of the exposure solution during the renewal process. A peristaltic pump continuously delivered the pre-prepared exposure solution to the zebrafish culture tank at a set flow rate to ensure a stable and controllable supply process. At the same time, the other end of the culture tank was connected to a drain pipe to drain the solution at the same speed and synchronously, so that the volume of the drained liquid was the same as the volume of the inlet water, maintaining a constant total liquid volume in the culture tank.

[0055] The metabolite extraction module is used for the effective extraction of metabolites from water samples and zebrafish tissue samples;

[0056] The data acquisition module uses LC-MS to detect and acquire characteristic peaks of metabolites in water samples and zebrafish tissue samples;

[0057] The data preprocessing module uses MS-DIAL (v4.9) software to extract, align, denoise, remove background and annotate metabolic characteristic peaks, and corrects instrument signal drift using the QC-RFSC method in the R package statTarget. Characteristic peaks with poor stability are removed according to the standard of CV < 30% to ensure data quality.

[0058] The data analysis module uses Class Fire, Metabo Analyst, R language, and Origin tools to classify and identify processed metabolic data, screen for differential metabolites, enrich pathways, and perform PLS-DA. Based on time series analysis, it identifies key metabolites that play a dominant role in metabolic differences at each time point and dynamically monitors the responses of core metabolites that repeatedly appear in pathway enrichment at multiple time points and represent the core regulation of the metabolic network. This allows for in-depth exploration of the metabolic characteristics and potential biological significance of water and zebrafish tissue samples.

[0059] The following embodiments propose a non-invasive continuous biomonitoring method based on non-targeted environmental metabolomics, comprising the following steps:

[0060] S1. Sample collection; including: randomly dividing zebrafish into a control group and a pollutant exposure group, using a continuous drip water exchange system, and changing the exposure solution daily; during the exposure process, collecting 1.5L of aquaculture water samples every 24 hours; at the end of the exposure, collecting whole male and female zebrafish as tissue samples;

[0061] S2. Metabolite extraction; including: effective extraction of metabolites from water samples and zebrafish tissue samples using solid-phase extraction and concentration methods;

[0062] S3. Data Acquisition; including: using LC-MS to detect and acquire characteristic peaks of metabolites in water samples and zebrafish tissue samples;

[0063] S4. Data preprocessing; including: using MS-DIAL software to extract, align, denoise, remove background and annotate metabolic characteristic peaks, and using the QC-RFSC method in the R package statTarget to correct instrument signal drift, and removing characteristic peaks with poor stability according to the standard of CV<30% to ensure data quality;

[0064] S5. Data Analysis: This includes classifying and identifying processed metabolic data, screening for differential metabolites, enriching pathways, and performing PLS-DA based on Class Fire, Metabo Analyst, R language, and Origin tools. It also involves identifying key metabolites that play a dominant role in metabolic differences at each time point based on time series data, and dynamically monitoring the core metabolites that repeatedly appear in pathway enrichment at multiple time points and represent the core regulation of the metabolic network.

[0065] In some embodiments, metabolite extraction includes: For water sample treatment, the water sample is first filtered through a 0.22 μm glass fiber membrane to remove suspended particles and microbial impurities, followed by enrichment of metabolites using an HLB solid-phase extraction column. Specific steps include: loading the water sample into a pre-activated HLB column; after extraction, drying the column residue with high-purity nitrogen; washing with 10 mL of LC-MS grade ultrapure water and eluting with 15 mL of LC-MS grade methanol. The resulting eluent is dried under nitrogen to near-dryness, then reconstituted with 150 μL of a 1:1 volume ratio acetonitrile-water mixture and transferred to a glass sample vial with an inner support tube for later use. For zebrafish tissue sample treatment, male and female zebrafish are mixed in equal proportions and thoroughly ground. Equal amounts of mixed tissue samples are taken from each group, and 1 mL of a methanol:acetonitrile:water mixture with a volume ratio of 4:4:2 is added. The samples are then ultrasonically extracted in an ice-water bath for 10 min, followed by sedimentation at -20°C. The sample was then centrifuged at 13,000 rpm and 4°C for 15 min. The supernatant was collected and concentrated to dryness by vacuum centrifugation. It was then reconstituted with 150 μL of acetonitrile-water mixture (1:1 volume ratio) and transferred to a glass sample vial for storage for subsequent analysis.

[0066] In some embodiments, LC-MS was used to detect and acquire characteristic peaks of metabolites in water samples and zebrafish tissue samples, covering both HILIC and RPLC modes. The HILIC mode parameters were as follows: a Waters BEH amide column (1.7 μm, 2.1 × 100 mm) was used, with a column temperature of 45 °C; mobile phase A was an aqueous solution containing 25 mM ammonium acetate and ammonium hydroxide, and mobile phase B was acetonitrile; the flow rate was 0.5 mL / min; the gradient was set as follows: phase B decreased from 95% to 65% within 0-7 min, then to 40% within 1 min, held for 1 min, and returned to the initial conditions. The RPLC mode parameters were as follows: a Phenomenex column was used. C18 column (2.6 μm, 2.1 × 100 mm), column temperature 25 °C; mobile phase A is an aqueous solution containing 0.01% acetic acid, and mobile phase B is isopropanol:acetonitrile (1:1, v / v); flow rate 0.3 mL / min; gradient: phase B increases from 1% to 99% within 0-7 min, retains for 1 min, and then returns to the initial conditions.

[0067] In some embodiments, MS-DIAL (v4.9) software was used to extract, align, denoise, remove background, and annotate metabolic characteristic peaks. Before importing into MS-DIAL, the raw mass spectrometry data file (.raw) was converted to ABF format (.abf) using Abfconverter ver.1.3. Key parameters in MS-DIAL were set as follows: mass tolerances for MS1 and MS2 were 0.01 Da and 0.025 Da, respectively; acquisition time range was 0.5–13.5 minutes; mass-to-charge ratio (m / z) range was set to 80–1200 Da; mass slice width was 0.05 Da; minimum peak height threshold was set to 20000; and peak width filtering threshold was set to 5 scans to remove background interference and low-quality signals, improving the accuracy and stability of peak identification. Metabolite annotation was performed based on MS spectrum matching, identifying peaks by comparing characteristic ion peaks in experimental samples with MS1 and MS2 spectra recorded in the real standard database. The standard database used was the total database provided by MS-DIAL, including the MS / MS database in positive ion mode (ESI(+)-MS / MS from authentic standards (16232 unique compounds) and the MS / MS database in negative ion mode (ESI(-)-MS / MS from authentic standards (8887 unique compounds)). During the matching process, the precise mass tolerances of MS1 and MS2 were set to 0.01 Da and 0.05 Da, respectively, to ensure the accuracy and reliability of metabolite annotation.

[0068] In some embodiments, correcting instrument signal drift refers to using the QC-RFSC algorithm in the statTarget software package under the R language environment. Using the quality control (QC) samples injected at intervals during the experiment as a reference, the signal change trend of each metabolic feature in different sample injection sequences is fitted and corrected, thereby eliminating systematic errors caused by batch effects or signal fluctuations generated during instrument detection.

[0069] In some embodiments, removing feature peaks with poor stability based on a CV < 30% means calculating the relative standard deviation (CV) of each metabolic feature in multiple tests based on the signal intensity of each feature in the QC sample, and using 30% as a threshold to select and retain feature variables with higher data stability. Feature peaks with a CV ≥ 30% are considered to have large detection fluctuations and poor reproducibility, and are therefore removed to reduce data noise and false positives, and improve the accuracy and reliability of subsequent statistical analysis and biological interpretation.

[0070] In some embodiments, classification and identification are achieved using the Classy Fire tool. The standardized structural identifiers (InChIKeys) corresponding to metabolites, obtained through mass spectrometry annotation, are input into the ClassyFire system. Its built-in ChemOnt chemical ontology system is then used to automatically classify the metabolites structurally. Based on the molecular structural features represented by the InChIKeys, and combined with a rule-driven classification algorithm, ClassyFire identifies the superclass, class, and subclass of metabolites hierarchically, achieving a systematic classification of various metabolites at the structural level.

[0071] In some embodiments, differential metabolite screening is performed using the StatisticalAnalysis module in the Metabo Analyst platform. Preprocessed metabolomics data is imported into this module, and its built-in statistical methods (such as one-way ANOVA and t-tests) are used to test the significance of metabolic characteristic peaks between different groups. This is combined with a comprehensive evaluation based on the fold change (FC) of metabolite expression between groups. The screening criteria for differential metabolites are set as follows: FC ≥ 1.5 or ≤ 0.67, with a corresponding p-value < 0.05. Metabolites meeting these criteria are considered to be significantly differentially expressed between the exposure group and the control group and are included in subsequent biological function analysis and metabolic pathway profiling.

[0072] In some embodiments, pathway enrichment analysis is performed using the Pathway Analysis module in the MetaboAnalyst platform. The screened differentially expressed metabolites are imported into this module, and metabolic pathway matching and enrichment analysis are performed using the corresponding species' metabolic pathway database. Based on the coverage and enrichment level of metabolites in each pathway, combined with topological structure and statistical significance, the system outputs potentially affected metabolic pathways. Pathways with a p-value < 0.05 are identified as significantly enriched pathways.

[0073] In some embodiments, the PLS-DA method effectively evaluates the distinguishing effect and consistency of fish metabolomics and aquatic metabolomics data across different treatment groups by associating them with sample grouping information.

[0074] In some embodiments, key metabolites that play a dominant role in metabolic differences at each time point are identified based on time series analysis. Multivariate statistical analysis is performed on metabolomics data collected at different time points within the exposure period. PLS-DA is used to calculate the variable importance projection value (VIP) of each metabolite, and the trend of VIP value changes of each metabolite at each time point is dynamically displayed.

[0075] In some embodiments, dynamic response monitoring of core metabolites involves tracking and analyzing the signal intensity of metabolites that repeatedly participate in key signaling pathways at multiple time points (i.e., core metabolites). By continuously monitoring the dynamic changes of these metabolites in the water, real-time assessment and monitoring of metabolite responses, fish physiological states, and aquatic environmental conditions can be achieved.

[0076] Example 1

[0077] The experimental animals used in this embodiment were healthy 4-month-old AB strain zebrafish (Danio rerio), and the pollutant was triclosan (TCS). Based on the typical distribution level of TCS in the environment, experimental groups were set up, including a control group and three exposure groups with TCS concentrations of 4 ppb, 20 ppb, and 100 ppb, where 4 ppb and 20 ppb were environmentally relevant concentrations, and 100 ppb was a high-dose exposure concentration. The control group did not receive any TCS, only dimethyl sulfoxide (DMSO) at a volume ratio of 1:10000; each exposure group had TCS dissolved in an equal proportion of DMSO to achieve the set concentration. Each treatment had three replicates, with 10 pairs of male and female zebrafish in each replicate, cultured in 10 L of exposure solution. To maintain the stability of triclosan concentration during exposure, a continuous drip water exchange system was used, with the exposure solution changed daily.

[0078] During the 7-day exposure experiment, 1.5 L of water samples were collected every 24 hours. The collected water samples were first filtered through a 0.22 μm glass fiber membrane to remove suspended particles and any microbial impurities that might be present in the water. The filtered samples were then subjected to HLB solid-phase extraction for metabolite enrichment. The specific steps included: loading the pretreated water sample into a pre-activated HLB column; after extraction, drying the column with high-purity nitrogen to remove any residual liquid; washing with 10 mL of LC-MS grade ultrapure water; and eluting with 15 mL of LC-MS grade methanol. The eluent was dried to near dryness under nitrogen and then reconstituted with 150 μL of a 1:1 acetonitrile-water mixture. The reconstituted sample was transferred to a glass vial with an inner support tube for subsequent metabolomics analysis.

[0079] All samples were analyzed by LC-MS, involving both HILIC and RPLC modes. The HILIC mode parameters were as follows: Waters BEH amide column (1.7 μm, 2.1 × 100 mm), column temperature 45℃; mobile phase A was an aqueous solution containing 25 mM ammonium acetate and ammonium hydroxide, and mobile phase B was acetonitrile; flow rate 0.5 mL / min; gradient settings: phase B decreased from 95% to 65% within 0-7 min, then to 40% within 1 min, held for 1 min, and returned to initial conditions. The RPLC mode parameters were: Phenomenex... C18 column (2.6 μm, 2.1 × 100 mm), column temperature 25 °C; mobile phase A is an aqueous solution containing 0.01% acetic acid, and mobile phase B is isopropanol:acetonitrile (1:1, v / v); flow rate 0.3 mL / min; gradient: phase B increases from 1% to 99% within 0-7 min, retains for 1 min, and then returns to the initial conditions.

[0080] Metabolic characteristic peaks were extracted, aligned, denoised, de-backgrounded, and annotated using MS-DIAL (v4.9) software. Instrument signal drift was corrected using the QC-RFSC method in the R package statTarget. Characteristic peaks with poor stability were removed based on a CV < 30%. Key parameters in MS-DIAL were set as follows: mass tolerances for MS1 and MS2 were 0.01 Da and 0.025 Da, respectively; acquisition time range was 0.5–13.5 minutes; mass-to-charge ratio (m / z) range was 80–1200 Da; mass slice width was 0.05 Da; minimum peak height threshold was set to 20000; and peak width filtering threshold was set to 5 scans to remove background interference and low-quality signals, improving the accuracy and stability of peak identification. Metabolite annotation was performed based on MS spectrum matching, identifying peaks by comparing characteristic ion peaks in experimental samples with MS1 and MS2 spectra recorded in the real standard database. The standard database used was the official MS-DIAL database, including the MS / MS database in positive ion mode (ESI(+)-MS / MS from authentic standards (16232 unique compounds) and the MS / MS database in negative ion mode (ESI(-)-MS / MS from authentic standards (8887 unique compounds)). During the matching process, the precise mass tolerances for MS1 and MS2 were set to 0.01 Da and 0.05 Da, respectively, to ensure the accuracy and reliability of metabolite annotation. Subsequently, the QC-RFSC algorithm in the statTarget package under R language was used, with the quality control (QC) samples injected at intervals during the experiment as a reference, to fit and correct the signal change trend of each metabolic feature in different sample injection sequences, thereby eliminating systematic errors caused by batch effects or signal fluctuations during instrument detection.

[0081] A data matrix was constructed based on the obtained metabolite names and their corresponding peak intensities. The PLS-DA method was used to model and analyze the metabolomes of fish and water to evaluate the ability to distinguish metabolic characteristics under different treatments, the differences between groups, and the consistency of the distinguishing effect between the metabolomes of fish and water.

[0082] After metabolite annotation, pairwise comparisons of metabolic levels between treatment and control groups were performed. FC (Functionally Validated) and p-value were used as screening criteria to identify differentially expressed metabolites caused by exposure. The specific procedure was as follows: the FC value of each metabolite between the treatment and control groups was calculated, and a two-tailed Student's test was performed. Metabolites meeting the criteria of FC ≥ 1.5 or FC ≤ 0.67 with p < 0.05 were selected as significantly differentially expressed metabolites. The screening results were visualized using a volcano plot, clearly showing the number and distribution characteristics of upregulated and downregulated metabolites, providing a basis for subsequent pathway enrichment analysis and biomarker screening.

[0083] The differentially identified metabolites were mapped and annotated using the Metabo Analyst platform and the KEGG database, and enrichment analysis was performed to reveal the metabolic pathways and biological functional modules involved, which helps to determine the systemic metabolic disturbances that may be caused by pollutant exposure.

[0084] To gain a deeper understanding of the differences in metabolic responses at different exposure time points, a time-series data matrix was constructed based on daily collected water sample metabolomics data. PLS-DA analysis was performed at each time point, and VIP values ​​were calculated to dynamically track the changes in the VIP values ​​of each metabolite over time, visually demonstrating the gradual evolution of metabolic differences within the exposure period.

[0085] Based on differential metabolite and KEGG pathway enrichment analyses at various time points, co-metabolites (core metabolites) that repeatedly participated in pathway enrichment at multiple time points were further screened. Their relative abundance throughout the entire exposure period was tracked and visualized to form a time-dynamic response map, which was used to characterize the perturbation patterns of stable metabolic networks in organisms under pollutant intervention.

[0086] The experimental process of Example 1 was tested and characterized, and the results are as follows: Figure 5-10 As shown.

[0087] like Figure 5 As shown, chemical classification of annotated metabolites using the Classy Fire tool revealed that the main categories in both water and fish matrices are organic acids and their derivatives, lipids and lipid molecules, and organic heterocyclic compounds. These metabolites are likely involved in various fundamental metabolic processes within organisms, such as energy metabolism, membrane composition and remodeling, and nucleotide metabolism. Notably, the specific categories and relative distribution of these metabolites are highly consistent between water and fish.

[0088] like Figure 6 As shown, PLS-DA analysis of water and fish metabolomics data at the exposure endpoint demonstrated that the water metabolomics exhibited a very similar ability to distinguish between the treatment and control groups as the fish metabolomics.

[0089] like Figure 7 As shown, metabolic pathway enrichment analyses of the aquatic and fish metabolomics revealed significant alterations in 18 metabolic pathways in the water samples and significant enrichment in 23 pathways in the fish samples at the three exposure concentrations. Eleven of these pathways were co-occurring in both matrices. This overlap indicates that although water and fish belong to different biological and environmental media, metabolites excreted by fish into the water can effectively indicate the interference effects of pollutant exposure.

[0090] To assess the sensitivity of water metabolomics to different exposure durations and doses, a systematic comparative metabolomics analysis was performed on water samples exposed at multiple time points and with multiple dose gradients. PLS-DA results showed that, at the same exposure dose, samples from different durations could be clearly distinguished. Figure 8 ); and under the same exposure time, different dose groups also showed a clear separation trend ( Figure 9 The above results indicate that the water metabolome is highly sensitive to changes in exposure conditions and has a good discriminative ability.

[0091] like Figure 10 As shown, taking key metabolites involved in KEGG enrichment as an example, it is possible to visualize the dynamic expression pattern of any metabolite throughout the exposure period. This not only helps to reveal the response patterns of metabolites during exposure, but also provides a powerful tool for understanding the persistent biochemical disturbances caused by pollutants in aquatic ecosystems.

[0092] The above are merely preferred embodiments of the present invention and are not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A non-invasive continuous biomonitoring system based on non-targeted environmental metabolomics, characterized in that, It includes a sample collection module, a metabolite extraction module, a data acquisition module, a data preprocessing module, and a data analysis module; among which: The sample collection module is used to randomly divide zebrafish into a control group and a pollutant exposure group, and to expose them using a continuous drip water exchange system, changing the exposure solution daily; during the exposure process, aquaculture water samples are collected every 24 hours; at the end of the exposure, whole male and female zebrafish are collected as tissue samples. The metabolite extraction module is used to effectively extract metabolites from water samples and zebrafish tissue samples. The data acquisition module uses LC-MS to detect and acquire characteristic peaks of metabolites in water samples and zebrafish tissue samples. The data preprocessing module uses MS-DIAL software to extract, align, denoise, remove background and annotate metabolic feature peaks, and corrects instrument signal drift using the QC-RFSC method in the R package statTarget. Feature peaks with poor stability are removed according to the standard of coefficient of variation CV < 30% to ensure data quality. The data analysis module uses Class Fire, Metabo Analyst, R language, and Origin tools to classify and identify processed metabolic data, screen for differential metabolites, enrich pathways, and perform partial least squares discriminant analysis (PLS-DA). Based on time series, it identifies key metabolites that play a dominant role in metabolic differences at each time point and dynamically monitors the core responses of core metabolites that repeatedly appear in pathway enrichment at multiple time points and represent the core regulation of the metabolic network.

2. The non-invasive continuous biomonitoring system based on non-targeted environmental metabolomics according to claim 1, characterized in that, The continuous drip water exchange system adopts a closed-loop control method to dynamically balance and stabilize the concentration of the exposure solution during the renewal process. The exposure solution is continuously delivered to the zebrafish rearing tank through a peristaltic pump to ensure a stable and controllable supply process. At the same time, the other end of the rearing tank is connected to a drain pipe to discharge the solution at the same speed and synchronously. The volume of the discharged liquid is consistent with the volume of the inlet water, maintaining a constant total liquid volume in the rearing tank.

3. A non-invasive continuous biomonitoring method based on non-targeted environmental metabolomics, characterized in that, Includes the following steps: S1. Sample collection; including: randomly dividing zebrafish into a control group and a pollutant exposure group, using a continuous drip water exchange system for exposure, and changing the exposure solution daily; during the exposure process, collecting aquaculture water samples every 24 hours; at the end of the exposure, collecting whole male and female zebrafish as tissue samples; S2. Metabolite extraction; including: extraction of metabolites from water samples and zebrafish tissue samples by solid-phase extraction and concentration. S3. Data Acquisition; including: using LC-MS to detect and acquire characteristic peaks of metabolites in water samples and zebrafish tissue samples; S4. Data preprocessing; including: using MS-DIAL software to extract, align, denoise, remove background and annotate metabolic characteristic peaks, and using the QC-RFSC method in the R package statTarget to correct instrument signal drift, and removing characteristic peaks with poor stability according to the standard of CV<30% to ensure data quality; S5. Data Analysis: This includes classifying and identifying processed metabolic data, screening for differential metabolites, enriching pathways, and performing PLS-DA based on Class Fire, Metabo Analyst, R language, and Origin tools. It also involves identifying key metabolites that play a dominant role in metabolic differences at each time point based on time series data, and dynamically monitoring the core metabolites that repeatedly appear in pathway enrichment at multiple time points and represent the core regulation of the metabolic network.

4. The non-invasive continuous biomonitoring method based on non-targeted environmental metabolomics according to claim 3, characterized in that, In step S1, the continuous drip water exchange system adopts a closed-loop control method to control the dynamic balance and concentration stability of the exposure solution renewal process. The exposure solution is continuously delivered to the zebrafish breeding tank by a peristaltic pump to ensure that the supply process is stable and controllable. At the same time, the other end of the breeding tank is connected to the drain pipe to drain the solution at the same speed and synchronously. The volume of the drained liquid is consistent with the volume of the inlet water, so as to maintain a constant total amount of liquid in the breeding tank.

5. The non-invasive continuous biomonitoring method based on non-targeted environmental metabolomics according to claim 3, characterized in that, In step S2, the metabolite extraction includes: first filtering the water sample through a 0.22 μm glass fiber membrane to remove suspended particles and microbial impurities, and then enriching the metabolites using an HLB solid-phase extraction column.

6. The non-invasive continuous biomonitoring method based on non-targeted environmental metabolomics according to claim 5, characterized in that, In step S2, the metabolite extraction step includes: For water sample processing, the water sample was loaded into a pre-activated HLB column. After extraction, the residual liquid in the column was dried with high-purity nitrogen. The sample was then washed with 10 mL of LC-MS grade ultrapure water and eluted with 15 mL of LC-MS grade methanol. The eluent was dried with nitrogen and concentrated to near-dry state. It was then reconstituted with 150 μL of acetonitrile-water mixture at a volume ratio of 1:1 and transferred to a glass sample vial with an inner support tube for storage. For zebrafish tissue sample processing, male and female zebrafish were mixed in equal proportions and thoroughly ground. Equal amounts of mixed tissue samples were taken from each group and 1 mL of a mixed extraction solution with a volume ratio of methanol:acetonitrile:water = 4:4:2 was added. After ultrasonic extraction in an ice-water bath for 10 min, the samples were allowed to stand at -20℃ to precipitate. Then, the samples were centrifuged at 13000 rpm and 4℃ for 15 min. The supernatant was collected and concentrated to dryness by vacuum centrifugation. The samples were then reconstituted with 150 μL of an acetonitrile-water mixture with a volume ratio of 1:1 and transferred to glass sample vials for storage.

7. The non-invasive continuous biomonitoring method based on non-targeted environmental metabolomics according to claim 3, characterized in that, In step S3, LC-MS was used to detect and acquire characteristic peaks of metabolites in water samples and zebrafish tissue samples, covering two modes: hydrophilic interaction chromatography (HILIC) and reversed-phase chromatography (RPLC); wherein: The HILIC mode parameters are as follows: a Waters BEH amide column was used, with a column temperature of 45℃; mobile phase A was an aqueous solution containing 25mM ammonium acetate and ammonium hydroxide, and mobile phase B was acetonitrile; the flow rate was 0.5mL / min; the gradient was set as follows: phase B decreased from 95% to 65% within 0-7min, decreased to 40% within 1min, and was held for 1min before returning to the initial conditions; The RPLC mode parameters are as follows: using Phenomenex C18 column, column temperature 25℃; mobile phase A is an aqueous solution containing 0.01% acetic acid, mobile phase B is isopropanol:acetonitrile, 1:1, v / v; flow rate is 0.3 mL / min; gradient is: phase B increases from 1% to 99% within 0-7 min, retains for 1 min and then returns to the initial conditions.

8. The non-invasive continuous biomonitoring method based on non-targeted environmental metabolomics according to claim 3, characterized in that, In step S4, before importing MS-DIAL, the raw mass spectrometry data file .raw is converted to ABF format .abf using Abfconverter ver.1.3; The key parameters in MS-DIAL are set as follows: the quality tolerances of MS1 and MS2 are 0.01 Da and 0.025 Da, respectively; the acquisition time range is 0.5-13.5 minutes, the mass-to-charge ratio (m / z) range is 80-1200 Da; the quality slice width is 0.05 Da; the minimum peak height threshold is 20000, and the peak width filtering threshold is 5 scans to remove background interference and low-quality signals, thereby improving the accuracy and stability of peak identification. Metabolite annotation is based on MS spectrum matching, which is achieved by comparing the characteristic ion peaks in the sample with the MS1 and MS2 spectra recorded in the real standard database; The standard database includes: an MS / MS database in positive ion mode (ESI(+)-MS / MS from authentic standards, 16232 unique compounds) and an MS / MS database in negative ion mode (ESI(-)-MS / MS from authentic standards, 8887 unique compounds). During the matching process, the precise mass tolerances for MS1 and MS2 are 0.01 Da and 0.05 Da, respectively.

9. The non-invasive continuous biomonitoring method based on non-targeted environmental metabolomics according to claim 3, characterized in that, In step S4, the correction of instrument signal drift includes: using the QC-RFSC algorithm in the statTarget software package under the R language environment, taking the quality control QC samples injected at intervals during the experiment as a reference, fitting and correcting the signal change trend of each metabolic feature in different sample injection sequences, and eliminating systematic errors caused by batch effects or signal fluctuations generated during instrument detection. The standard for removing feature peaks with CV < 30% is as follows: Based on the signal intensity of each metabolic feature in multiple tests in the QC sample, the relative standard deviation (CV) value of each feature is calculated, and a threshold of 30% is used to screen and retain feature variables with high data stability; feature peaks with CV ≥ 30% are considered to have large detection fluctuations and poor reproducibility, and are removed to reduce data noise and false positives.

10. The non-invasive continuous biomonitoring method based on non-targeted environmental metabolomics according to claim 3, characterized in that, In step S5, the classification and identification are achieved using the Classy Fire tool. The standardized structural identifiers InChIKey corresponding to the metabolites obtained through mass spectrometry annotation are input into the Classy Fire system, and the built-in chemical ontology system Chem Ont is used to automatically classify the metabolites structurally. Based on the molecular structural features represented by InChIKey and combined with a rule-driven classification algorithm, Classy Fire identifies the superclass, class, and subclass chemical classification information of metabolites hierarchically, thereby achieving a systematic classification of various metabolites at the structural level. The differential metabolite screening was performed using the Statistical Analysis module in the Metabo Analyst platform. Preprocessed metabolomics data were imported into this module, and its built-in statistical methods were used to perform significance tests on metabolic characteristic peaks between different groups. The fold change (FC) of metabolite expression between groups was also used for comprehensive evaluation. The screening criteria for differential metabolites were: FC ≥ 1.5 or ≤ 0.67, with a corresponding p-value < 0.

05. Metabolites meeting these criteria were determined to be significantly differentially expressed between the exposure group and the control group and were included in biological function analysis and metabolic pathway profiling. The pathway enrichment analysis was performed using the Pathway Analysis module in the Metabo Analyst platform. The screened differential metabolites were imported into this module, and metabolic pathway matching and enrichment analysis were performed using the corresponding species' metabolic pathway database. The system outputs potential metabolic pathways based on the coverage and enrichment level of metabolites in each pathway, combined with topological structure and statistical significance. Among them, pathways with p-value < 0.05 were identified as significantly enriched pathways. The PLS-DA model correlates fish metabolomics and aquatic metabolomics data with sample grouping information to evaluate the distinguishing effect and consistency of the two between different treatment groups. Based on time series identification, key metabolites that play a dominant role in metabolic differences at each time point are identified. Multivariate statistical analysis is performed on metabolomics data collected at different time points within the exposure period. PLS-DA is used to calculate the variable importance projection value (VIP) of each metabolite, and the VIP value change trend of each metabolite at each time point is dynamically displayed. The core metabolites are those that repeatedly participate in key signaling pathways at multiple time points. Monitoring their dynamic response includes: conducting signal intensity tracking analysis at multiple time points, continuously monitoring the dynamic changes of core metabolites in the water body, and conducting real-time assessment and monitoring of metabolite response, fish physiological state, and aquatic environmental conditions.

Images