Water bloom biomarker determination method, water bloom early warning method and device

Abstract

The invention discloses a water bloom biomarker determination method and a water bloom early warning method and device.The water bloom biomarker determination method comprises the steps that a single algae sample data set and a mixed algae sample data set are obtained, and a machine learning model is constructed with mass spectrum characteristic data as input characteristic data and first index data as an output target; training and checking a machine learning model by using the single-algae sample data set to obtain a trained first machine learning model; training and checking the machine learning model by using the mixed algae sample data set to obtain a trained second machine learning model; performing importance evaluation on the input feature data of the first machine learning model and the second machine learning model to extract volatile organic compounds corresponding to one or more important mass-to-charge ratio positions as water bloom biomarkers, the problem that an existing screened biomarker cannot accurately reflect the algal bloom trend of a multi-algae coexistence scene is solved, and a multi-index comprehensive decision basis is provided for accurate management of algal bloom.

Description

Technical Field

[0001] This disclosure relates to, but is not limited to, the field of water environment protection technology, and particularly to a method for determining biomarkers of algal blooms, an algal bloom early warning method, and an apparatus. Background Technology

[0002] Harmful algal blooms (HABs) are characterized by the rapid and excessive growth of algae in water bodies, producing various toxins that lead to unpleasant odors, turbidity, dissolved oxygen depletion, and toxicity to aquatic animals and humans. Due to their significant threats to water quality, aquatic ecosystems, and human health, harmful algal blooms have become a global environmental challenge. With the increasing severity of global warming and nutrient overload, large-scale algal blooms in lakes and reservoirs are becoming more frequent. Therefore, achieving rapid and accurate prediction of algal density is crucial for mitigating the risks posed by harmful algal blooms.

[0003] Currently, the monitoring and early warning of algal blooms mainly rely on data-driven models, such as prediction methods based on water quality physicochemical parameters or remote sensing data. Although such methods can achieve early warning to a certain extent, they often depend on the acquisition and processing of a large number of parameters, resulting in problems such as complex detection methods, time-consuming and labor-intensive data collection, and high monitoring costs. Summary of the Invention

[0004] The following is an overview of the subject matter described in detail herein. This overview is not intended to limit the scope of the claims.

[0005] This disclosure provides a method for determining biomarkers of algal blooms, including: Obtain a single algae sample dataset and a mixed algae sample dataset. The single algae sample dataset includes mass spectrometry feature data and first index data of volatile organic compounds released during the growth of at least one algae. The mixed algae sample dataset includes mass spectrometry feature data and first index data of volatile organic compounds released during the mixed growth of multiple algae. The first index data is algal density data or can be used to calculate the algal density data. Using the mass spectrometry feature data as input feature data and the first index data as output target, a machine learning model is constructed; the machine learning model is trained and tested using the single algae sample dataset to obtain a trained first machine learning model; the machine learning model is trained and tested using the mixed algae sample dataset to obtain a trained second machine learning model. The importance of the input feature data of the first machine learning model and the second machine learning model is evaluated respectively, so as to extract one or more volatile organic compounds corresponding to important mass-to-charge ratio positions as biomarkers of algal blooms.

[0006] This disclosure also provides an apparatus for determining algal bloom biomarkers, including a memory; and a processor connected to the memory, the memory being used to store instructions, the processor being configured to perform the steps of the algal bloom biomarker determination method as described in any embodiment of this disclosure based on the instructions stored in the memory.

[0007] This disclosure also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for determining algal bloom biomarkers according to any embodiment of this disclosure.

[0008] This disclosure also provides a program product including instructions that, when executed by a computer, perform the method for determining algal bloom biomarkers as described in any embodiment of this disclosure.

[0009] This disclosure also provides a method for early warning of algal blooms, including: Based on predetermined algal bloom biomarkers, a biomarker smoothing contribution function is constructed, wherein the biomarker smoothing contribution function is used to represent the relationship between the mass spectrometry signal intensity of the algal bloom biomarker and the algal bloom risk contribution, wherein the algal bloom risk contribution is used to represent the impact on the fluctuation of the first indicator data of the water body, wherein the first indicator data is algal density data or can be used to calculate the algal density data; Mass spectrometry feature data of volatile organic compounds released by algae in the target water body are obtained, and the mass spectrometry signal intensity of the algal bloom biomarker in the mass spectrometry feature data is extracted. Substituting the mass spectrometry signal intensity of the algal bloom biomarker into the smoothing contribution function of the biomarker, the algal bloom risk contribution amount corresponding to the algal bloom biomarker is obtained, and the algal bloom risk probability of the target water body is calculated based on the algal bloom risk contribution amount.

[0010] This disclosure also provides an algal bloom early warning device, including a memory; and a processor connected to the memory, the memory being used to store instructions, the processor being configured to execute the steps of the algal bloom early warning method as described in any embodiment of this disclosure based on the instructions stored in the memory.

[0011] This disclosure also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the algal bloom early warning method described in any embodiment of this disclosure.

[0012] This disclosure also provides a program product including instructions that, when executed by a computer, perform the algal bloom early warning method as described in any embodiment of this disclosure.

[0013] The method, method, and apparatus for determining algal bloom biomarkers in this disclosure utilize single algal sample datasets and mixed algal sample datasets to train and test machine learning models, resulting in a trained first machine learning model and a second machine learning model. The importance of the input feature data of the first and second machine learning models is evaluated to extract volatile organic compounds (VOCs) corresponding to one or more important mass-to-charge ratio positions as algal bloom biomarkers. This solves the problem that existing biomarkers screened for AVOCs cannot accurately reflect the algal bloom trend in multi-algal coexistence scenarios. It can be used for early and accurate identification and warning of algal blooms, and can be applied to the monitoring and control of algal blooms in freshwater lakes, reservoirs, and nearshore waters. This disclosure utilizes an interpretable machine learning model to efficiently screen reliable algal bloom biomarkers from numerous algal VOCs, enabling early identification of algal blooms, improving monitoring sensitivity and timeliness, and providing strong technical support for accurate early warning and scientific control of algal blooms. It has significant application value in the fields of environmental monitoring and ecological protection.

[0014] Other features and advantages of this disclosure will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the disclosure. Other advantages of this disclosure may be realized and obtained by means of the methods described in the description and the accompanying drawings. Attached Figure Description

[0015] The accompanying drawings are used to provide an understanding of the technical solutions of this disclosure and form part of the specification. They are used together with the embodiments of this disclosure to explain the technical solutions of this disclosure and do not constitute a limitation on the technical solutions of this disclosure.

[0016] Figure 1 This is a flowchart illustrating a method for determining biomarkers of algal blooms, provided as an exemplary embodiment of this disclosure.

[0017] Figure 2 Fluorescence and signal intensity distribution of Microcystis aeruginosa and Chlorella vulgaris in a co-culture system provided for exemplary embodiments of this disclosure.

[0018] Figure 3 Growth curves of algal samples grown under different conditions for 15 days, provided as an exemplary embodiment of this disclosure.

[0019] Figure 4A The image shows the changes in mass spectrometry characteristics of AVOCs in groups SM, SC, 1M1C, 2M1C, and 1M2C under hypertrophic conditions provided in an exemplary embodiment of this disclosure, as a result of the growth process.

[0020] Figure 4BThe following are mass spectrometry characteristic data of AVOCs in the SM group, SC group, 1M1C group, 2M1C group and 1M2C group under various other different nutritional conditions provided for exemplary embodiments of this disclosure, as shown in the graphs.

[0021] Figure 5 A schematic diagram illustrating the performance of a machine learning model for predicting algal density, provided as an exemplary embodiment of this disclosure.

[0022] Figure 6 This diagram illustrates the top 10 AVOC species in single-culture and co-culture systems, obtained from feature importance algorithm SHAP analysis, as an exemplary embodiment of this disclosure.

[0023] Figure 7 A schematic diagram illustrating the results of the analysis of the biological activity and physiological correlation of interaction-related AVOCs provided for exemplary embodiments of this disclosure.

[0024] Figure 8A A schematic diagram illustrating the predictive effect of using basal metabolic bloom biomarkers at different stages of algal growth, provided as an exemplary embodiment of this disclosure.

[0025] Figure 8B A schematic diagram illustrating the predictive effects of combining basal metabolic bloom biomarkers and interacting bloom biomarkers at different stages of algal growth, provided as an exemplary embodiment of this disclosure.

[0026] Figure 9 A schematic diagram of the nonlinear fitting curve of the interaction between algal bloom biomarkers and algal biomass, provided as an exemplary embodiment of this disclosure.

[0027] Figure 10 A schematic diagram illustrating the quantitative analysis results of interactions between biomarkers of algal blooms, provided as an exemplary embodiment of this disclosure.

[0028] Figure 11 A flowchart illustrating an algal bloom early warning method provided as an exemplary embodiment of this disclosure.

[0029] Figure 12 A schematic diagram of a device for determining algal bloom biomarkers provided as an exemplary embodiment of this disclosure.

[0030] Figure 13 This is a schematic diagram of the structure of an algal bloom early warning device provided as an exemplary embodiment of the present disclosure. Detailed Implementation

[0031] This disclosure describes several embodiments, but these descriptions are exemplary and not limiting, and it will be apparent to those skilled in the art that many more embodiments and implementations are possible within the scope of the embodiments described herein. Although many possible combinations of features are shown in the drawings and discussed in the detailed description, many other combinations of the disclosed features are also possible. Unless specifically limited, any feature or element of any embodiment may be used in combination with, or may replace, any feature or element of any other embodiment.

[0032] This disclosure includes and contemplates combinations of features and elements known to those skilled in the art. The embodiments, features, and elements disclosed in this disclosure may also be combined with any conventional features or elements to form a unique inventive scheme as defined by the claims. Any feature or element of any embodiment may also be combined with features or elements from other inventive schemes to form another unique inventive scheme as defined by the claims. Therefore, it should be understood that any feature shown and / or discussed in this disclosure may be implemented individually or in any suitable combination. Therefore, the embodiments are not limited except by the limitations imposed by the appended claims and their equivalents. Furthermore, various modifications and changes may be made within the scope of the appended claims.

[0033] Furthermore, in describing representative embodiments, the specification may have presented methods and / or processes as a specific sequence of steps. However, the method or process should not be limited to the specific order of steps described herein, to the extent that the method or process does not depend on the specific order of steps described herein. As will be understood by those skilled in the art, other sequences of steps are also possible. Therefore, the specific order of steps set forth in the specification should not be construed as a limitation of the claims. Moreover, the claims relating to the method and / or process should not be limited to the steps performed in the order written, and those skilled in the art will readily understand that these orders can be varied and still remain within the spirit and scope of the embodiments disclosed herein.

[0034] Biomarkers are considered valuable for early warning of algal blooms because they directly reflect the physiological state and growth trend of algae. However, existing research on biomarkers for algal bloom identification mainly focuses on non-volatile substances such as chlorophyll a and algal toxins. While chlorophyll a can serve as an indirect indicator of algal biomass, its changes are not significant before algal proliferation, lacking early warning capabilities. Algal toxins, although directly related to harmful algal blooms, are typically produced in large quantities in the middle and late stages of blooms, exhibiting a response lag, and the detection process is relatively cumbersome and time-consuming, limiting their application effectiveness. In contrast, volatile organic compounds (AVOCs) released by algae during metabolism have attracted attention as potential biomarkers because they can more directly reflect the metabolic state of algae.

[0035] Current research on AVOCs still has the following shortcomings: Most existing methods are based on single-algal culture systems, failing to consider the core characteristic of algal blooms in actual aquatic ecosystems—driven by interactions (competition, inhibition, etc.) among multiple algae; they also fail to distinguish between "basal metabolic AVOCs" and "algal-interaction-related AVOCs"; and the screened biomarkers cannot accurately reflect the bloom trends in multi-algal coexistence scenarios. Therefore, current monitoring technologies are insufficient to meet the sensitivity and specificity requirements for early identification of algal blooms.

[0036] like Figure 1 As shown in the embodiments of this disclosure, a method for determining biomarkers of algal blooms is provided, including: Step 101: Obtain a single algae sample dataset and a mixed algae sample dataset. The single algae sample dataset includes mass spectrometry feature data and first index data of volatile organic compounds released during the growth of at least one type of algae. The mixed algae sample dataset includes mass spectrometry feature data and first index data of volatile organic compounds released during the mixed growth of multiple types of algae. The first index data is algal density data or can be used to calculate algal density data. Step 102: Using mass spectrometry feature data as input feature data and the first index data as output target, construct a machine learning model; use a single algae sample dataset to train and test the machine learning model to obtain a trained first machine learning model; use a mixed algae sample dataset to train and test the machine learning model to obtain a trained second machine learning model. Step 103: Evaluate the importance of the input feature data of the first machine learning model and the second machine learning model respectively, and extract one or more volatile organic compounds corresponding to important mass-to-charge ratio positions as biomarkers of algal blooms.

[0037] The method for determining algal bloom biomarkers provided in this disclosure trains and validates a machine learning model using single algal sample datasets and mixed algal sample datasets, respectively, to obtain a trained first machine learning model and a second machine learning model. The importance of the input feature data of the first and second machine learning models is evaluated to extract volatile organic compounds corresponding to one or more important mass-to-charge ratio positions as algal bloom biomarkers. This method solves the problem that existing biomarkers screened for AVOCs cannot accurately reflect the algal bloom trend in multi-algal coexistence scenarios. It can be used for early and accurate identification and warning of algal blooms, and can be applied to the monitoring and control of algal blooms in freshwater lakes, reservoirs, and nearshore waters. The algal bloom biomarkers determined through this disclosure enable early identification of algal blooms, improve monitoring sensitivity and timeliness, and provide strong technical support for accurate early warning and scientific control of algal blooms, demonstrating significant application value in environmental monitoring and ecological protection.

[0038] The machine learning model in this embodiment uses mass spectrometry data of AVOCs released during algal growth as input feature data and a first indicator data as the output target to achieve early warning based on algal metabolic signals. AVOCs, as secondary metabolites constantly released during algal growth, can reflect algal growth status signals more promptly and comprehensively compared to other algal growth status indicators (such as chlorophyll a or algal toxins). Online monitoring of AVOCs can be achieved using a proton charge transfer time-of-flight mass spectrometer. This embodiment uses mass spectrometry data of AVOCs as the core indicator for model training, identifies the probability of algal blooms and predicts future trends, and fully integrates the intrinsic indicators of algal growth status. It breaks through the limitations of the traditional "passive treatment after algal bloom" and achieves "active early warning before the bloom" based on the metabolic signals of algal interactions. It avoids the secondary pollution risk of physical harvesting, the residual hazards of chemical algaecides, and the lag in the effect of biological algae control. It reduces the environmental and economic costs of algal bloom control from the source and helps the sustainable management of aquatic ecosystems. Compared with previous algal bloom early warning schemes, it can achieve true early warning of algal blooms faster and better.

[0039] In some exemplary embodiments, when the first indicator data can be used to calculate algal density data, the first indicator is any one of the following: chlorophyll a content, absorbance at a preset wavelength. However, this disclosure is not limited thereto. The first indicator data can also be set to other indicator data that can be used to calculate algal density data as needed.

[0040] In some exemplary embodiments, the preset wavelength can be 680 nm; however, this disclosure does not limit this. The following description uses absorbance (OD680) with a first index data of 680 nm as an example.

[0041] In some exemplary embodiments, the method further includes, prior to: The relationship curve between algal density and the first index is pre-constructed using the following method: multiple preset algal solutions of different concentrations are prepared; the number of algal cells in each preset algal solution and the value of the first index of each preset algal solution are measured; based on the number of algal cells in each preset algal solution and the value of the first index corresponding to each preset algal solution, the relationship curve between algal density and the first index of the preset algal solution is fitted.

[0042] In this embodiment of the present disclosure, the algae contained in the preset algae solution, single algae sample, and mixed algae sample can be set as needed. For example, the single algae sample may include a single algae sample of Microcystis aeruginosa and a single algae sample of Chlorella vulgaris. The algae contained in the preset algae solution and mixed algae sample may be a mixed algae of Microcystis aeruginosa and Chlorella vulgaris. However, the present disclosure does not limit this.

[0043] In some exemplary implementations, the sample dataset can be constructed using the following methods: Under multiple different combinations of initial experimental parameters, pre-set single algal samples and mixed algal samples were cultured respectively. The experimental parameters included total nitrogen concentration, total phosphorus concentration, and initial algal cell density. Within a preset culture period, the first indicator data of the growth process of single algae samples and mixed algae samples were collected, and the volatile organic compounds released during the growth process of single algae samples and mixed algae samples were collected; the collected volatile organic compounds were measured by mass spectrometry to obtain mass spectrometry data; the obtained mass spectrometry data were preprocessed to obtain mass spectrometry characteristic data.

[0044] For example, taking a pre-set single algae sample including *Microcystis aeruginosa* single algae sample and *Chlorella vulgaris* single algae sample, and a mixed algae sample including *Microcystis aeruginosa* and *Chlorella vulgaris* co-culture samples with different initial experimental parameter combinations as an example, in a certain experimental design, 20 treatment groups were set up, representing a combination of 5 culture modes and 4 nutrient levels. The culture modes were set as follows: (1) *Microcystis aeruginosa* single culture (SM); (2) *Chlorella vulgaris* single culture (SC); and (3-5) co-culture of *Microcystis aeruginosa* and *Chlorella vulgaris*, with cell density ratios of 1:1 (1M1C), 1:2 (1M2C), and 2:1 (2M1C), respectively. The cell density of the single cultured *Microcystis aeruginosa* and ordinary *Chlorella vulgaris* was 1×10⁻⁶. 6 cells·mL -1 In all co-cultures, the numbers 1 and 2 in the proportions represent an initial cell density of 1 × 10⁻⁶ cells. 6 cells·mL -1 and 2×10 6 cells·mL -1All treatments were cultured for 15 days. To ensure that the results reflected the effects of interspecific interactions rather than differences in initial inoculation density, we maintained consistent initial densities for each algae species. The inoculation ratios were chosen based on previous studies and reflected ecologically relevant cell densities observed in freshwater lakes. To better simulate the environment of real-world water bodies, we simulated four nutrient levels (defined by total nitrogen (TN) and total phosphorus (TP) concentrations) in the laboratory to assess the effects of eutrophication. Cell densities of both algae species were monitored during culture.

[0045] The experimental conditions for each group are shown in Table 1. Each algal sample was categorized into four nutrient levels—Super, High, Moderate, and Low—based on total nitrogen (TN) and total phosphorus (TP) concentrations. During the 15-day culture period, samples were collected 1-2 times daily. For the monoculture groups (SM and SC), the absorbance (OD680) of the algal solution at 680 nm was measured to represent algal biomass. For the co-culture groups, algal cell density was measured using an independent detection channel of a flow cytometer. Specifically, 1 mL of culture was collected from each flask and filtered through a 300-mesh nylon sieve. A total of 200 μL of filtrate was mixed with 10 μL of standard microbeads for analysis. The main pigment of *Microcystis aeruginosa* is phycocyanin, detected using the FL4 channel (the channel in flow cytometry used to detect specific fluorescence signals); *Chlorella vulgaris* mainly contains chlorophyll a. During co-culture, the two species could be effectively distinguished in a two-dimensional plot using the FL4 channel, such as... Figure 2 As shown, APC_A represents the fluorescence channel, SSC_A represents the amount of intracellular refractive substances, H represents counting microspheres (in purple), F represents Microcystis aeruginosa (in red), and G represents Chlorella vulgaris (in blue). Table 1 In some exemplary embodiments, the mass spectrometry feature data includes the mass spectrometry signal intensity at multiple mass-to-charge ratio positions within a preset mass-to-charge ratio range.

[0046] For example, the preset mass-to-charge ratio range can be between 15 and 249; however, this disclosure does not limit it. The preset mass-to-charge ratio range can be set as needed.

[0047] In this embodiment of the disclosure, the multiple mass-to-charge ratio positions can be any integer mass-to-charge ratio position within the range of 15 to 249; however, this disclosure does not limit this. For example, in other examples, the multiple mass-to-charge ratio positions can be multiple mass-to-charge ratio positions spaced apart by a preset step size within a preset mass-to-charge ratio range, wherein the preset step size can be any real number such as 1 or 2. When the initial mass-to-charge ratio position is 15, the ending mass-to-charge ratio position is 249, and the preset step size is 1, the multiple mass-to-charge ratio positions are all integer mass-to-charge ratio positions within the range of 15 to 249.

[0048] The AVOCs mass spectrometry feature data disclosed in this embodiment include feature numbers across the entire mass spectrometry range. These feature combinations fully reflect the growth status of algae, improving the stability and accuracy of algal bloom prediction methods.

[0049] For example, during the algae cultivation process of the different groups mentioned above, at each sampling and measurement of algae density, the headspace gas of the culture flask was simultaneously collected to detect AVOCs. The specific steps are as follows: online measurement was performed using a mass spectrometer, with the mass resolution set to 2500 m / Δm (full width at half maximum), using H3O + As the reactant ion, the sample introduction system uses a PEEK capillary, heated to 70°C to prevent sample condensation, and headspace gas is directly introduced into the instrument through the injection port; the mass-to-charge ratio (m / z) range is set to 15-249, and the operating parameters include: ion source pressure 600V, drift tube voltage 800V, ion source current 3.00mA, RF voltage 150V, and signal accumulation time 20s.

[0050] The initial spectra obtained by the mass spectrometer were preprocessed (including peak identification, integration, and calibration) to finally obtain characteristic signal intensity data of AVOCs in the range of m / z 15-249 under different processing.

[0051] In this embodiment, Figure 3 Growth curves of algal samples grown under different conditions for 15 days are provided. Figure 3 In the table, a contains the growth curves of Microcystis aeruginosa from group SM and Chlorella from group SC; b contains the growth curves of Microcystis aeruginosa and Chlorella from group 1M1C; c contains the growth curves of Microcystis aeruginosa and Chlorella from group 2M1C; and d contains the growth curves of Microcystis aeruginosa and Chlorella from group 1M2C.

[0052] In this embodiment of the disclosure, Figure 4A The data provided are a graph showing the changes in mass spectrometry characteristics of AVOCs in the SM, SC, 1M1C, 2M1C and 1M2C groups under hypertrophic conditions as they grew. Figure 4BThe graphs provide mass spectrometry characteristics of AVOCs in the SM, SC, 1M1C, 2M1C, and 1M2C groups under various different nutritional conditions, showing the changes in these characteristics over time. Figure 4B In the diagram, 'a' corresponds to eutrophic conditions, 'b' to mesotrophic conditions, and 'c' to oligotrophic conditions. The sample dataset was obtained based on the M / Z signal intensity data and algal density data of samples under different ecological conditions.

[0053] In some exemplary embodiments, training and testing machine learning models using single algae sample datasets or mixed algae sample datasets includes: Divide the single algae sample dataset or the mixed algae sample dataset into a training set, a validation set, and a test set; The machine learning model is trained using the training set; the hyperparameters of the machine learning model are tuned using the validation set, where the tuned hyperparameters include: the number of base estimators, the learning rate, and the maximum depth, and the optimal parameter model is saved. Use the test set to evaluate the performance of the optimal parameter model.

[0054] In this embodiment of the disclosure, the machine learning model can be an XGBoost regression model, or other regression models, such as k-Nearest Neighbor (KNN), Support Vector Machine (SVM), etc., and this disclosure does not limit it.

[0055] For example, during the 15-day culture period, multiple groups were treated separately, with each group having three replicates, resulting in a total of n=1268 samples. The mass spectrometry data for each processed sample were presented as mass spectrum signal intensities (m / z 15-249) at mass-to-charge ratios of 235, and algal cell density was presented as OD680 data (or other indicators in other examples). This formed a data matrix (1268×236) with AVOCs concentration data as input features and algal cell density as the output target. This data was used to construct an XGBoost model, and the model's predictive performance was finally evaluated. XGBoost achieved R² in both single-culture and co-culture studies. 2 The values ​​reached 0.94 and 0.96 respectively, such as Figure 5 As shown, Figure 5 In the diagram, a is a schematic diagram of the prediction performance of the XGBoost model trained using a single algae sample dataset, and b is a schematic diagram of the prediction performance of the XGBoost model trained using a mixed algae sample dataset.

[0056] In some exemplary embodiments, the importance of the input feature data of the first machine learning model and the second machine learning model is evaluated, including at least one of the following: The contribution of the mass spectrum signal intensity at each mass-to-charge ratio position in the input feature data to the output target of the first and second machine learning models is calculated using the feature importance algorithm of the regression model. The first and second machine learning models are explained using Shapley Additive Explanations (SHAP). The marginal contribution and direction of the mass spectrum signal intensity at each mass-to-charge ratio position in the input feature data to the output target of the first and second machine learning models are calculated.

[0057] This disclosure utilizes interpretability techniques of machine learning models to assess the importance of algal volatile organic compounds (AVOCs). Specifically, feature importance analysis and / or SHAP analysis methods can be used to evaluate the importance of input feature data. This disclosure quantifies the contribution and direction of influence of AVOCs on algal density prediction by separately assessing the importance of input feature data for a first machine learning model and a second machine learning model, providing interpretable evidence for biomarker screening.

[0058] For example, embodiments of this disclosure can calculate the contribution of input feature data (i.e., the mass spectrum signal intensity at each mass-to-charge ratio position) to algal density prediction using a feature importance algorithm of a regression model, and rank them according to importance indices. In this example, the SHAP method is used to interpret the regression model globally and locally, quantifying the marginal contribution of each AVOC to the prediction result and its direction of influence (positive or negative). This method can effectively reveal the importance and influence direction of each feature in the "black box model". Through SHAP analysis, AVOCs closely related to algal density are further confirmed. In embodiments of this disclosure, the AVOCs corresponding to the top 10 mass-to-charge ratio positions obtained by feature importance analysis of an XGBoost model trained using a single algal sample dataset are shown in Table 2, and the AVOCs corresponding to the top 10 mass-to-charge ratio positions obtained by feature importance analysis of an XGBoost model trained using a mixed algal sample dataset are shown in Table 3. The last column, the mean SHAP value, is the importance index of the SHAP analysis method. Table 2 Table 3 In some exemplary embodiments, the algal bloom biomarkers include: basal metabolic algal bloom biomarkers and interacting algal bloom biomarkers. The basal metabolic algal bloom biomarkers include volatile organic compounds corresponding to the top M1 important mass-to-charge ratio positions extracted for the first machine learning model, where M1 is a natural number greater than 1. The biomarkers of the interacting algal blooms include volatile organic compounds corresponding to M2 important mass-to-charge ratio positions. The M2 important mass-to-charge ratio positions are within the range of the first M3 important mass-to-charge ratio positions extracted for the second machine learning model, and do not overlap with the first M1 important mass-to-charge ratio positions extracted for the first machine learning model. M3 is a natural number greater than 1, and M2 is a natural number greater than or equal to 1.

[0059] Taking M1=M3=10 as an example, such as Figure 6 As shown, comparing the top 10 most important AVOCs in single-culture and co-culture systems ( Figure 6 In this study, 'a' corresponds to monoculture and 'b' corresponds to coculture. Based on the correlation analysis between each substance and algal physiological function, algal bloom biomarkers are categorized into two main types according to the following criteria: basal metabolic bloom biomarkers and interspecific bloom biomarkers. Basal metabolic bloom biomarkers (FMAs) are defined as the top 10 substances in monoculture systems that are correlated with algal density prediction, and these substances are also related to basic algal carbon, nitrogen, and sulfur metabolism. Interspecific bloom biomarkers (IAAs) are defined as AVOCs that appear only in the top 10 of coculture systems, not in the top 10 of monoculture systems, and whose metabolic mechanisms are directly related to interalgal chemical communication and interspecies interactions.

[0060] Based on the above criteria, a total of 10 FMAs (Methanal Acids) were identified in M1, including dimethylamine (m / z 46), methanethiol (m / z 49), 2-methyl-1-butanol (m / z 88), and acetone (m / z 58) (i.e., all substances in Table 2). Dimethylamine exhibited negative SHAP dependence (high concentration inhibits algal density) in both systems, originating from downstream polyamine demethylation in glutamate metabolism, reflecting nitrogen flux disturbance and growth inhibition. Methanethiol exhibited positive SHAP dependence (high concentration promotes algal density), participating in sulfur-containing amino acid metabolism and cell growth-related methylation reactions, reflecting enhanced biosynthetic activity. These AVOCs all embed in the basic carbon, nitrogen, and sulfur metabolic pathways, capturing the basic physiological state of algae. A total of 4 IAAs (Intra-Acids) were identified in M2: phenylethanol (m / z 123), chloromethane (m / z 51), isophorone (m / z 139), and DMNT (m / z 151). These substances only entered the top 10 in the co-culture system, and their metabolic mechanisms are closely related to interspecies interactions: phenethyl alcohol is derived from phenylpropanol metabolism, isophorone and DMNT are terpene-derived compounds, both of which are recognized interalgal chemical communication signaling molecules; chloromethane participates in metabolic regulation related to interalgal nutrient competition, further verifying its correlation with interalgal interactions.

[0061] In some exemplary embodiments, the method further includes: Transcriptomic analysis was performed on algal samples from the growth of at least one type of algae alone and on algal samples from the mixed growth of multiple types of algae. Using algal samples from the growth of at least one type of algae as a reference, differentially expressed genes were screened from algal samples from the mixed growth of multiple algae, and genes related to algal metabolic pathways were selected from the differentially expressed genes. Correlation analysis was performed on the mass spectrometry signal intensity of interacting algal bloom biomarkers and the expression levels of genes related to algal metabolic pathways to determine the correlation between the mass spectrometry signal intensity of interacting algal bloom biomarkers and genes related to algal metabolic pathways. Biomarkers of algal blooms with a correlation lower than a preset correlation threshold were removed.

[0062] For example, to further clarify the association between phenylethanol (m / z 123), chloromethane (m / z 51), isophorone (m / z 139), DMNT (m / z 151) and algae, this embodiment combines transcriptome sequencing technology to verify the relationship at the gene expression level. The specific steps and results are as follows: (1) Experimental design and sample collection Three culture treatments were set up: a single culture group of *Microcystis aeruginosa*, a single culture group of *Chlorella vulgaris*, and a co-culture group of *Microcystis aeruginosa* and *Chlorella vulgaris* (inoculation ratio 1:1). On day 3 of culture, algal samples were collected from the three groups for transcriptomics testing.

[0063] (2) Differential gene screening and metabolic pathway localization Using the single-culture group as a control, differentially expressed genes under co-culture treatment were screened; based on the carotenoid synthesis pathway, MAPK signaling pathway and fatty acid metabolism pathway closely related to algal interactions in literature research, key genes related to the above metabolic pathways were selected from the differentially expressed genes.

[0064] (3) Correlation analysis Spearman correlation analysis was performed on the concentration data of phenylethanol, chloromethane, isophorone, and DMNT with the expression levels of key genes involved in carotenoid synthesis, MAPK signaling, and fatty acid metabolism pathways. The results are as follows: Figure 7 As shown, the four interacting AVOCs (i.e., interacting algal bloom biomarkers) were significantly correlated with the expression of key genes in multiple metabolic pathways (p < 0.05), and all were positively correlated. Figure 7 In the diagram, red circles indicate positive correlations, blue circles indicate negative correlations, the size of the circles represents the magnitude of the correlation, and the asterisk (*) inside the circles indicates the significance level. These results validate the association between phenylethanol, chloromethane, isophorone, DMNT, and algal interactions at the gene expression level, clarifying that the release of these substances is directly related to key algal interaction pathways such as carotenoid synthesis, signal transduction, and fatty acid metabolism, further supporting their scientific validity as interacting AVOCs.

[0065] To validate the role of IAAs in predicting algal blooms based on AVOCs, we conducted the following analysis. Given that the interaction between *Microcystis aeruginosa* and *Chlorella vulgaris* occurs in two distinct phases, we assessed the role of these volatile organic compound (AVOC) groups at specific stages. FMAs maintained high predictive accuracy in the early stages (R0.05). 2 = 0.92), but it decreased in the later stages (R = 0.92). 2 = 0.83), such as Figure 8A As shown. In comparison, combining FMAs and IAAs in both stages achieved better accuracy, especially in the later stages of algal blooms (early stage: R...). 2 = 0.94; Later stage: R 2 = 0.88), such as Figure 8B As shown in the figure. These results indicate that FMAs provide a general metabolic baseline for prediction, while IAAs capture dynamic interaction signals that become increasingly critical during algal bloom development. This underscores the need to consider both primary metabolites and interaction-related terpenoids when selecting biomarkers for algal density prediction.

[0066] This disclosure, through parallel comparison of single-culture and co-culture systems, combined with SHAP analysis and transcriptome verification, for the first time screened out four specific AVOCs (phenylethanol, chloromethane, isophorone, and DMNT) that only respond to the interaction between *Microcystis aeruginosa* and *Chlorella vulgaris*. These substances are directly related to key interaction pathways such as interalgal carotenoid synthesis, MAPK signaling, and fatty acid metabolism, solving the problem of insufficient indicator specificity caused by neglecting interalgal interactions in existing technologies, and making algal bloom monitoring more consistent with the actual scenario of multiple algae coexisting in natural water bodies.

[0067] In some exemplary embodiments, the method further includes: Calculate the SHAP values ​​of interacting algal bloom biomarkers in a mixed algal sample dataset; The relationship between the content of interacting algal bloom biomarkers and their SHAP values ​​was fitted using a generalized additive model (GAM). The content of substances at the critical point where the SHAP value is 0 in the fitted curve was taken as the threshold content of the interacting algal bloom biomarkers.

[0068] For example, based on the four interacting AVOCs (phenylethanol, m / z 123; chloromethane, m / z 51; isophorone, m / z 139; DMNT, m / z 151) obtained from the above screening, this embodiment further quantifies the correlation between the content of each substance and algal density through generalized additive model (GAM) combined with SHAP value analysis, and determines the reference threshold point for algal bloom early warning. The generalized additive model (GAM) method is used to fit the SHAP value to the important AVOCs. The Intergovernmental Panel on Climate Change (IPCC) defines a critical point as "a point at which small changes become significant enough to trigger larger, more critical changes that may be sudden, irreversible, and have cascading effects." In this embodiment, the point where the SHAP value changes from positive to negative or from negative to positive (SHAP = 0) is defined as the critical point. The threshold concentration at the critical point can be used as an auxiliary and supplementary indicator of an impending increase in algal density.

[0069] Figure 9 Fitting curves for the SHAP values ​​of the four interacting AVOCs, as shown below. Figure 9 As shown, for chloromethane (m / z51): the coefficient of determination of the SHAP value fitting curve. = 0.818, the SHAP value changes from negative to approaching 0, and the inflection point corresponds to a substance content of 13431.54 (unit: signal strength). This critical point can be used as an auxiliary indicator of algal density changes. When the chloromethane content is below this threshold, the algal density shows a rapid increasing trend.

[0070] For phenylethanol (m / z 123): Fitted curve = 0.829, the SHAP value changes from negative to positive, and the inflection point corresponds to a substance content of 1449.53 (unit: signal strength). When the phenylethanol content is higher than this threshold, the growth trend of algal density is significantly enhanced.

[0071] For isophorone (m / z 139): Fitted curve == 0.697, the SHAP value changes from close to 0 to a positive value, and the turning point corresponds to a substance content of 2802.28 (unit: signal strength). This threshold can be used as a basis for judging whether algal density is about to enter a rapid growth phase.

[0072] For DMNT (m / z151): Fitted curve = 0.649, the SHAP value changes from close to 0 to a negative value, and the inflection point corresponds to a substance content of 10649.5 (unit: signal intensity). When the DMNT content is higher than this threshold, the growth trend of algal density will be inhibited.

[0073] In this embodiment, the critical point at which the SHAP value changes from negative to positive or from positive to negative is defined as the reference threshold point for algal bloom early warning. The threshold concentration (signal intensity) corresponding to each substance can serve as an auxiliary indicator that a significant change in algal density is imminent: when the content of a certain substance approaches or exceeds the corresponding threshold in actual monitoring, the trend of algal density change can be predicted, providing a quantitative basis for early warning and management decisions of algal blooms.

[0074] To gain a deeper understanding of the synergistic effect of the four interacting AVOCs identified in the screening on algal density prediction, this embodiment further employs partial dependency graph (PDP) analysis to clarify the interaction between substances, identify the changing trend of algal density under multi-indicator synergistic scenarios, and provide multi-dimensional reference for algal bloom management decisions. Figure 10 The figure shows the PDP analysis results of pairwise interactions of the four types of interacting AVOCs.

[0075] See Figure 10 For substance combination 1 (such as M / Z_51 and M / Z_123): As can be seen from the color partitions and numerical gradients in the PDP diagram, when M / Z_51 is in the low concentration range and the concentration of M / Z_123 gradually increases, the algal density shows a significant increasing trend; however, when the concentration of M / Z_51 exceeds a certain threshold, even if the concentration of M / Z_123 continues to increase, the growth rate of algal density will slow down significantly.

[0076] For material combination 2 (such as M / Z_139 and M / Z_151): The PDP plot shows that the algal density is at a relatively high level in the range of low concentration of M / Z_139 and high concentration of M / Z_151; when the concentration of M / Z_139 increases and the concentration of M / Z_151 decreases, the algal density shows a gradient decreasing trend.

[0077] For other substance combinations: the remaining PDP diagrams of the pairwise interactions clearly show the variation of algal density under different concentration combinations. For example, some combinations maintain stable algal density in a specific concentration range, while in another range, they show rapid growth or decline.

[0078] In this embodiment, the partial dependency graph (PDP) visually illustrates the marginal effect of two interacting AVOCs on algal density under different concentration combinations. By analyzing these interaction patterns, the changing trend of algal density under the synergistic effect of multiple substances can be clarified, providing decision-makers with a visual reference tool for "multi-index concentration combination - algal density trend": when the concentrations of multiple interacting AVOCs are within a certain combination range in actual monitoring, decision-makers can predict the development direction of algal density and then formulate targeted algal bloom control strategies, improving the scientific nature and accuracy of management decisions.

[0079] After identifying biomarkers for algal blooms, related technologies only establish a simple correspondence between AVOCs and algal density, without quantifying the dose-response relationship between key indicators and algal biomass. Therefore, they lack quantitative management capabilities, cannot provide actionable early warning thresholds, and do not consider the interaction between indicators, resulting in a significant decrease in prediction accuracy under complex environments.

[0080] like Figure 11 As shown in the embodiments of this disclosure, an algal bloom early warning method is also provided, including: Step 1101: Based on the pre-determined algal bloom biomarkers, construct the biomarker smoothing contribution function, whereby the biomarker smoothing contribution function is used to represent the relationship between the mass spectrometry signal intensity of the algal bloom biomarker and the algal bloom risk contribution, and the algal bloom risk contribution is used to represent the impact on the fluctuation of the first indicator data of the water body, whereby the first indicator data is algal density data or can be used to calculate algal density data. Step 1102: Obtain mass spectrometry characteristic data of volatile organic compounds released by algae in the target water body, and extract the mass spectrometry signal intensity of algal bloom biomarkers from the mass spectrometry characteristic data; Step 1103: Substitute the mass spectrometry signal intensity of the extracted algal bloom biomarker into the biomarker smoothing contribution function to obtain the algal bloom risk contribution corresponding to the algal bloom biomarker, and calculate the algal bloom risk probability of the target water body based on the algal bloom risk contribution.

[0081] The algal bloom early warning method of this disclosure constructs a biomarker smoothing contribution function to fit the quantitative relationship between algal bloom biomarkers and algal biomass, thereby providing a multi-indicator comprehensive decision-making basis for precise algal bloom management.

[0082] In this embodiment of the disclosure, the pre-determined algal bloom biomarkers can be determined using the aforementioned algal bloom biomarker determination method. When the aforementioned algal bloom biomarker determination method is used, the pre-determined algal bloom biomarkers may include the aforementioned basal metabolic algal bloom biomarkers and / or interacting algal bloom biomarkers; however, this disclosure does not limit this.

[0083] In some exemplary embodiments, the number of algal bloom biomarkers includes multiple ones, and the biomarker smoothing contribution function includes multiple single-substance smoothing contribution functions and at least one multi-substance joint smoothing contribution function. Each single-substance smoothing contribution function represents the relationship between the mass spectrometry signal intensity of a single algal bloom biomarker and its contribution to algal bloom risk, while each multi-substance joint smoothing contribution function represents the relationship between the mass spectrometry signal intensity of at least two algal bloom biomarkers and their contribution to algal bloom risk. This disclosure, by constructing multiple single-substance smoothing contribution functions and at least one multi-substance joint smoothing contribution function, can analyze the synergistic or antagonistic effects between different algal bloom biomarkers, thereby improving the accuracy of algal bloom prediction in complex environments.

[0084] In this embodiment of the disclosure, a single-substance smoothing contribution function can be constructed for each algal bloom biomarker, or a single-substance smoothing contribution function can be constructed only for the aforementioned basic metabolic algal bloom biomarkers. This disclosure does not impose any restrictions on this.

[0085] In this embodiment of the disclosure, a multi-substance joint smoothing contribution function can be constructed for each pair of interacting algal bloom biomarkers, or a multi-substance joint smoothing contribution function can be constructed for each pair of arbitrary algal bloom biomarkers. This disclosure does not limit this.

[0086] In some exemplary embodiments, when the contribution of algal bloom risk is greater than 0, it indicates that its "promoting contribution" to algal bloom is positive; when the contribution of algal bloom risk is less than 0, it indicates that its "promoting contribution" to algal bloom is negative and its "inhibiting contribution" to algal bloom is positive.

[0087] In some exemplary embodiments, calculating the probability of algal bloom risk in a target water body based on the contribution of algal bloom risk includes: The comprehensive algal bloom risk index is obtained by weighting the algal bloom risk contribution output by multiple single-substance smoothing contribution functions with the algal bloom risk contribution output by at least one multi-substance joint smoothing contribution function. The comprehensive index of algal bloom risk is standardized to obtain the probability of algal bloom risk.

[0088] To achieve quantitative prediction of algal bloom risk based on interaction metabolic indicators, this disclosure proposes a method for calculating the probability of algal bloom risk that combines the nonlinear response of a single substance with the interaction effect between substances. The method specifically includes the following steps: (1) Single-substance effect modeling For the key metabolite indicators obtained from the screening ( The nonlinear response of algae to algal density was fitted using a generalized additive model (GAM), and the following functional form was established: ; in, Metabolites The smooth contribution function of algal density change (i.e., the single-substance smooth contribution function) reflects the smooth influence of its content change on algal growth rate or biomass fluctuation. This is a nonparametric smoothing function estimated from sample data, used to describe the nonlinear relationship between metabolite concentration and algal density response.

[0089] Determine the threshold concentration based on the changing trend of the function curve. This threshold represents the critical point at which algal density changes from a low-risk to a high-risk state. It can serve as a reference indicator for early warning of algal blooms.

[0090] (2) Modeling of interactions between matter For each of the two key metabolites The combined marginal effects function is extracted using partial dependency plot (PDP). ; in, This represents the nonparametric joint smoothing function (i.e., the multi-substance joint smoothing contribution function) obtained by fitting sample data, which describes the combined effect of the synergistic changes in the concentrations of two metabolites on algal density. The corresponding smoothing contribution term is used as the input model for the comprehensive effect of interactions on algal density changes. Through analysis... The contour distribution and numerical gradient can identify key concentration ranges for interactive enhancement or inhibition, providing a quantitative basis for the algal density variation pattern under multi-substance synergy.

[0091] (3) Calculation of the comprehensive risk index of algal bloom A weighted combination of single-substance response functions and interaction functions is used to define a comprehensive algal bloom risk index. : ; in, The baseline risk term (intercept) represents the basic risk of algal bloom when metabolite concentrations are at equilibrium or at a reference background level. The weighting coefficients contributing to single-substance smoothing reflect the metabolites The relative importance of changes in the overall risk of algal blooms; The weighting coefficients represent the interactions between substances, reflecting the strength of the synergistic or antagonistic effects of metabolites on changes in algal density.

[0092] In actual use, The value can be set according to the parameters of the actual water body, for example, It can be 0.5.

[0093] In some exemplary embodiments, to avoid the influence of differences in material scale on the results, the weighting coefficients are standardized using variance normalization, specifically as follows: .

[0094] (4) Quantification of Algal Bloom Risk Probability Comprehensive risk index The probability of algal bloom risk is obtained by standardization mapping using a sigmoid function. : .

[0095] in, This indicates the probability of an algal bloom occurring; the closer the value is to 1, the higher the risk. Low-risk, warning, and high-risk levels can be categorized based on empirical thresholds (e.g., 0.2, 0.6).

[0096] (5) Output Results and Applications Obtained through calculation This model can provide aquatic ecosystem management departments with quantitative algal bloom risk assessment results, enabling dynamic monitoring and risk classification management of algal blooms based on metabolic interaction indicators. The model combines data-driven characteristics with biological interpretability, and can be extended to the prediction and early warning of algal blooms in multi-algal species and multi-aquatic environments.

[0097] To quantitatively characterize the risk level of algal blooms, we propose an "algal bloom risk composite index" model. The core idea of ​​this model is to transform the variation pattern of metabolite concentration into a contribution to the risk of algal blooms, and then combine these contributions to form a risk probability in the range of 0–1.

[0098] Specifically, the formation of algal blooms is often related to changes in the concentration of various metabolites. Increased concentrations of certain metabolites can promote algal growth and increase the risk of algal blooms, while others may have an inhibitory effect. There may also be synergistic or antagonistic relationships between different substances. For example, two metabolites may have no significant effect when they are alone, but when they are both elevated, they can significantly accelerate algal proliferation.

[0099] Therefore, we decompose the risk of algal blooms into three parts: (1) Base risk: represents the inherent risk of algal bloom in water bodies under normal water quality conditions and without obvious metabolite abnormalities.

[0100] (2) Contribution of a single metabolite: describes the individual effect of changes in the concentration of each metabolite on the increase in algal density. For example, when the concentration of a certain volatile organic compound (VOC) exceeds a threshold, the model will determine that its "promoting contribution" to algal bloom is positive.

[0101] (3) Interactions between metabolites: reflecting the synergistic or inhibitory effects on the risk of algal bloom when two substances coexist.

[0102] The model calculates a risk contribution value for each metabolite and interaction term based on measured data (such as the historical relationship between algal density and metabolite concentration). These values ​​reflect the strength of their impact on algal bloom trends. Subsequently, the model weights and sums all contributions to obtain a comprehensive risk index. This index itself can be positive (indicating an increased risk of algal blooms) or negative (indicating a lower risk of algal blooms).

[0103] To make this index more intuitively reflect the risk level, we further use a logical mapping function to transform the comprehensive risk index... Convert to algal bloom risk probability in the 0–1 interval .when When the value is close to 0, it indicates that the system is relatively stable and unlikely to experience algal blooms in the short term; when... When the value approaches 1, it indicates that multiple metabolite states have entered a high-risk zone, and the likelihood of algal blooms increases significantly. In other words, the model does not directly predict the "numerical value of algal density," but rather the degree of risk of "whether algal density may rise rapidly under the current metabolite characteristics." In this way, the model achieves a continuous mapping from changes in biochemical signals → trends in algal density changes → risk probability, and can dynamically and quantitatively reflect the algal bloom risk level of water bodies.

[0104] This disclosure quantifies the dose-response relationship between the interaction of AVOCs and algal density using a generalized additive model (GAM), clarifying the algal bloom warning thresholds for each substance (e.g., signal intensity values ​​of chloromethane 13431.54, phenylethanol 1449.53, etc.). Simultaneously, it utilizes partial dependency graphs (PDPs) to analyze inter-substance interactions, forming a comprehensive judgment rule of "single-index threshold + multi-index synergy." This quantitative system provides managers with directly actionable warning criteria, enabling earlier algal bloom warning response times and significantly improving prediction accuracy in complex environments. Furthermore, by optimizing experimental design and analysis processes, this disclosure enables end-to-end application from AVOCs detection to algal bloom warning decision-making, and is widely adaptable to monitoring scenarios of different water bodies such as drinking water sources, lakes, and reservoirs, demonstrating strong feasibility for widespread adoption.

[0105] This disclosed embodiment breaks through the limitations of the traditional "passive treatment after algal bloom" approach. It achieves "active early warning before the bloom" based on the metabolic signals of algal interactions, avoiding the secondary pollution risks of physical dredging, the residual hazards of chemical algaecides, and the lag in the effect of biological algae control. It reduces the environmental and economic costs of algal bloom control from the source and helps the sustainable management of aquatic ecosystems.

[0106] This disclosure also provides an apparatus for determining algal bloom biomarkers, including a memory; and a processor connected to the memory, the memory being used to store instructions, the processor being configured to perform the steps of the algal bloom biomarker determination method as described in any embodiment of this disclosure based on the instructions stored in the memory.

[0107] like Figure 12As shown, in one example, the algal bloom biomarker identification device may include: a first processor 1210, a first memory 1220, a first bus system 1230, and a first transceiver 1240, wherein the first processor 1210, the first memory 1220, and the first transceiver 1240 are connected via the first bus system 1230, the first memory 1220 is used to store instructions, and the first processor 1210 is used to execute the instructions stored in the first memory 1220 to control the first transceiver 1240 to transmit and receive signals. Specifically, the first transceiver 1240, under the control of the first processor 1210, can acquire single algae sample datasets and mixed algae sample datasets. The single algae sample dataset includes mass spectrometry data of volatile organic compounds released during the individual growth of at least one algae, along with first index data. The mixed algae sample dataset includes mass spectrometry data of volatile organic compounds released during the mixed growth of multiple algae, along with first index data. The first index data is algal density data or can be used to calculate the algal density. The first processor 1210 uses the mass spectrometry data as input feature data and the first index data as the output target to construct a machine learning model. The machine learning model is trained and tested using the single algae sample dataset to obtain a trained first machine learning model. The machine learning model is then trained and tested using the mixed algae sample dataset to obtain a trained second machine learning model. The importance of the input feature data of the first and second machine learning models is evaluated to extract one or more volatile organic compounds corresponding to important mass-to-charge ratio positions as biomarkers for algal blooms.

[0108] It should be understood that the first processor 1210 can be a central processing unit (CPU), or it can be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), off-the-shelf programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor, etc.

[0109] The first memory 1220 may include read-only memory and random access memory, and provides instructions and data to the first processor 1210. A portion of the first memory 1220 may also include non-volatile random access memory. For example, the first memory 1220 may also store device type information.

[0110] The first bus system 1230 may include a power bus, control bus, and status signal bus, in addition to a data bus. However, for clarity, in... Figure 12The general labeled all buses as the first bus system 1230.

[0111] In implementation, the processing performed by the processing device can be accomplished through integrated logic circuits in the hardware of the first processor 1210 or through software instructions. That is, the method steps of this embodiment can be executed by the hardware processor, or by a combination of hardware and software modules within the processor. The software modules can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other storage media. This storage medium is located in the first memory 1220. The first processor 1210 reads information from the first memory 1220 and, in conjunction with its hardware, completes the steps of the above method. To avoid repetition, further details are omitted here.

[0112] This disclosure also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the method for determining algal bloom biomarkers as described in any embodiment of this disclosure. The method for determining algal bloom biomarkers driven by executing executable instructions is essentially the same as the method provided in the above embodiments of this disclosure, and will not be described in detail here.

[0113] In some possible implementations, various aspects of the method for determining algal bloom biomarkers provided in this disclosure can also be implemented as a program product comprising program code that, when run on a computer device, causes the computer device to perform the steps in the method for determining algal bloom biomarkers according to various exemplary embodiments of this disclosure as described above. For example, the computer device can execute the method for determining algal bloom biomarkers as described in the embodiments of this disclosure.

[0114] This disclosure also provides an algal bloom early warning device, including a memory; and a processor connected to the memory, the memory being used to store instructions, the processor being configured to execute the steps of the algal bloom early warning method as described in any embodiment of this disclosure based on the instructions stored in the memory.

[0115] like Figure 13As shown, in one example, the algal bloom early warning device may include: a second processor 1310, a second memory 1320, a second bus system 1330, and a second transceiver 1340. The second processor 1310, the second memory 1320, and the second transceiver 1340 are connected through the second bus system 1330. The second memory 1320 is used to store instructions, and the second processor 1310 is used to execute the instructions stored in the second memory 1320 to control the second transceiver 1340 to transmit and receive signals. Specifically, the second transceiver 1340, under the control of the second processor 1310, can acquire mass spectrometry characteristic data of volatile organic compounds released by algae in the target water body. The second processor 1310 constructs a marker smoothing contribution function based on a pre-determined algal bloom biomarker, wherein the marker smoothing contribution function is used to represent the relationship between the mass spectrometry signal intensity of the algal bloom biomarker and the algal bloom risk contribution, and the algal bloom risk contribution is used to represent the impact on the fluctuation of the first indicator data, which is algal density data or data that can be used to calculate the algal density. The processor extracts the mass spectrometry signal intensity of the algal bloom biomarker from the mass spectrometry characteristic data; substitutes the extracted mass spectrometry signal intensity of the algal bloom biomarker into the marker smoothing contribution function to obtain the algal bloom risk contribution corresponding to the algal bloom biomarker; and calculates the algal bloom risk probability of the target water body based on the algal bloom risk contribution.

[0116] It should be understood that the second processor 1310 can be a central processing unit (CPU), or it can be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), off-the-shelf programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor, etc.

[0117] The second memory 1320 may include read-only memory and random access memory, and provides instructions and data to the second processor 1310. A portion of the second memory 1320 may also include non-volatile random access memory. For example, the second memory 1320 may also store device type information.

[0118] In addition to the data bus, the second bus system 1330 may also include a power bus, a control bus, and a status signal bus. However, for clarity, in... Figure 13 The general designated all buses as the second bus system 1330.

[0119] In implementation, the processing performed by the processing device can be accomplished through integrated logic circuits in the hardware of the second processor 1310 or through software instructions. That is, the method steps of this embodiment can be executed by a hardware processor, or by a combination of hardware and software modules within the processor. The software modules can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other storage media. This storage medium is located in the second memory 1320. The second processor 1310 reads information from the second memory 1320 and, in conjunction with its hardware, completes the steps of the above method. To avoid repetition, further details are omitted here.

[0120] This disclosure also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the algal bloom warning method as described in any embodiment of this disclosure. The algal bloom warning method driven by executing executable instructions is essentially the same as the algal bloom warning method provided in the above embodiments of this disclosure, and will not be described in detail here.

[0121] In some possible implementations, various aspects of the algal bloom warning method provided in this disclosure can also be implemented as a program product comprising program code that, when run on a computer device, causes the computer device to perform the steps in the algal bloom warning method according to various exemplary embodiments of this disclosure as described above. For example, the computer device can execute the algal bloom warning method described in the embodiments of this disclosure.

[0122] The program product may employ any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0123] It will be understood by those skilled in the art that all or some of the steps, systems, or apparatuses disclosed above, and their functional modules / units, can be implemented as software, firmware, hardware, or suitable combinations thereof. In hardware implementations, the division between functional modules / units mentioned above does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed collaboratively by several physical components. Some or all components may be implemented as software executed by a processor, such as a digital signal processor or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit (ASIC). Such software may be distributed on a computer-readable medium, which may include computer storage media (or non-transitory media) and communication media (or transient media). As is known to those skilled in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and can be accessed by a computer. Furthermore, it is well known to those skilled in the art that communication media typically contain computer-readable instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0124] It should be noted that the above embodiments or implementation methods are merely exemplary and not restrictive. Therefore, this disclosure is not limited to the content specifically shown and described herein. Various modifications, substitutions, or omissions can be made to the form and details of the implementations without departing from the scope of this disclosure.

Claims

1. A method for determining biomarkers of algal blooms, characterized in that, include: Obtain a single algae sample dataset and a mixed algae sample dataset. The single algae sample dataset includes mass spectrometry feature data and first index data of volatile organic compounds released during the growth of at least one algae. The mixed algae sample dataset includes mass spectrometry feature data and first index data of volatile organic compounds released during the mixed growth of multiple algae. The first index data is algal density data or can be used to calculate the algal density data. Using the mass spectrometry feature data as input feature data and the first index data as output target, a machine learning model is constructed; the machine learning model is trained and tested using the single algae sample dataset to obtain a trained first machine learning model; the machine learning model is trained and tested using the mixed algae sample dataset to obtain a trained second machine learning model. The importance of the input feature data of the first machine learning model and the second machine learning model is evaluated respectively, so as to extract one or more volatile organic compounds corresponding to important mass-to-charge ratio positions as biomarkers of algal blooms.

2. The method for determining algal bloom biomarkers according to claim 1, characterized in that, The algal bloom biomarkers include: basal metabolic algal bloom biomarkers and interaction algal bloom biomarkers. The basal metabolic algal bloom biomarkers include volatile organic compounds corresponding to the top M1 important mass-to-charge ratio positions extracted by the first machine learning model, where M1 is a natural number greater than 1. The biomarkers of the interacting algal blooms include volatile organic compounds corresponding to M2 important mass-to-charge ratio positions. The M2 important mass-to-charge ratio positions are within the range of the first M3 important mass-to-charge ratio positions extracted for the second machine learning model, and do not overlap with the range of the first M1 important mass-to-charge ratio positions extracted for the first machine learning model. M3 is a natural number greater than 1, and M2 is a natural number greater than or equal to 1.

3. The method for determining algal bloom biomarkers according to claim 1, characterized in that, The method further includes: Transcriptomic analysis was performed on algal samples from the growth of at least one type of algae alone and on algal samples from the mixed growth of multiple types of algae. Using algal samples from the growth of at least one type of algae alone as a reference, differentially expressed genes in algal samples from the mixed growth of multiple types of algae are screened, and genes related to algal metabolic pathways are selected from the differentially expressed genes. A correlation analysis was performed between the mass spectrometry signal intensity of the algal bloom biomarker and the expression level of genes related to the algal metabolic pathway to determine the correlation between the mass spectrometry signal intensity of the algal bloom biomarker and the genes related to the algal metabolic pathway. Algal bloom biomarkers with a correlation lower than a preset correlation threshold were removed.

4. A device for identifying biomarkers of algal blooms, characterized in that, The method includes a memory; and a processor connected to the memory, the memory being used to store instructions, the processor being configured to perform the steps of the method for determining algal bloom biomarkers as described in any one of claims 1 to 3 based on the instructions stored in the memory.

5. A method for early warning of algal blooms, characterized in that, include: Based on predetermined algal bloom biomarkers, a biomarker smoothing contribution function is constructed, wherein the biomarker smoothing contribution function is used to represent the relationship between the mass spectrometry signal intensity of the algal bloom biomarker and the algal bloom risk contribution, wherein the algal bloom risk contribution is used to represent the impact on the fluctuation of the first indicator data of the water body, wherein the first indicator data is algal density data or can be used to calculate the algal density data; Mass spectrometry feature data of volatile organic compounds released by algae in the target water body are obtained, and the mass spectrometry signal intensity of the algal bloom biomarker in the mass spectrometry feature data is extracted. Substituting the mass spectrometry signal intensity of the algal bloom biomarker into the smoothing contribution function of the biomarker, the algal bloom risk contribution amount corresponding to the algal bloom biomarker is obtained, and the algal bloom risk probability of the target water body is calculated based on the algal bloom risk contribution amount.

6. The algal bloom early warning method according to claim 5, characterized in that, The number of algal bloom biomarkers includes multiple ones, and the biomarker smoothing contribution function includes multiple single-substance smoothing contribution functions and at least one multi-substance joint smoothing contribution function. Each single-substance smoothing contribution function is used to represent the relationship between the mass spectrometry signal intensity of a single algal bloom biomarker and the algal bloom risk contribution. Each multi-substance joint smoothing contribution function is used to represent the relationship between the mass spectrometry signal intensity of at least two algal bloom biomarkers and the algal bloom risk contribution.

7. The algal bloom early warning method according to claim 6, characterized in that, The step of calculating the algal bloom risk probability of the target water body based on the algal bloom risk contribution includes: The algal bloom risk contribution output by multiple single-substance smoothing contribution functions is weighted and calculated with the algal bloom risk contribution output by at least one multi-substance joint smoothing contribution function to obtain the comprehensive algal bloom risk index. The comprehensive index of algal bloom risk is standardized to obtain the probability of algal bloom risk.

8. An algal bloom early warning device, characterized in that, The method includes a memory; and a processor connected to the memory, the memory being used to store instructions, the processor being configured to perform the steps of the algal bloom early warning method as described in any one of claims 5 to 7 based on the instructions stored in the memory.

9. A computer-readable storage medium, characterized in that, It stores a computer program that, when executed by a processor, implements the method for determining biomarkers of algal blooms as described in any one of claims 1 to 3, or the method for early warning of algal blooms as described in any one of claims 5 to 7.

10. A computer program product, characterized in that, The instruction includes, when the computer program product is executed by a computer, the instruction performing the method for determining algal bloom biomarkers as described in any one of claims 1 to 3, or the method for early warning of algal blooms as described in any one of claims 5 to 7.

Images