Method and device for mixture risk assessment based on environmental virtual cells

Abstract

The present application relates to the field of artificial intelligence risk assessment and bioinformatics technology, and discloses a mixture risk assessment method and device based on environmental virtual cells, wherein the present application maps the characterization of the pollutant mixture to be evaluated, the exposure condition characterization and the baseline cell characterization into the cell state of the target cell by using the environmental virtual cell model, and then determines the biological effect from the cell state and generates the risk assessment result. The overall biological effect of the mixture is reflected based on the actual state change of the cell, and the defect that the traditional technology is difficult to predict the real biological effect of the cell caused by the pollutant mixture is overcome. Meanwhile, for mixtures with different components, proportions and exposure conditions, the characterization data thereof is obtained to complete the evaluation, the cell state and the comprehensive risk assessment of the pollutant mixture under the real exposure scene are accurately characterized, and the rapid and reproducible evaluation of the risk of the pollutant mixture is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of artificial intelligence risk assessment and bioinformatics, and particularly relates to a mixture risk assessment method and device based on environmental virtual cells. BACKGROUND

[0002] With the mass production and wide application of chemicals, there are many coexisting pollutants in media such as water, soil, air, food and human biological samples, forming chemical mixtures with complex composition and dynamic concentration ratio over time. At present, the risk assessment of pollutant mixtures mainly uses dose addition model, independent action model or risk quotient addition method to quantitatively estimate the mixed risk. At the same time, computational toxicology and cell perturbation prediction methods are used to predict the potential biological effects of pollutants.

[0003] However, the above prior art still has the following defects: first, the existing prediction method lacks comprehensive characterization ability of mixture component composition, proportion change, exposure time and target cell baseline state difference, and it is difficult to realize the prediction of the overall cell state change under mixture exposure. Second, it is difficult to accurately reflect the cell response and comprehensive risk caused by chemical mixtures under real exposure scenarios, because the intracellular fate process, organelle-distinguished perturbation effect, cell state prediction and mixture risk assessment are not integrated into a unified framework. SUMMARY

[0004] The present application provides a mixture risk assessment method and device based on environmental virtual cells to solve the problems of predicting the overall cell state change under mixture exposure and accurately reflecting the cell response and comprehensive risk caused by chemical mixtures under real exposure scenarios.

[0005] In a first aspect, the present application provides a mixture risk assessment method based on environmental virtual cells, the method comprising: obtaining an environmental virtual cell model; wherein the environmental virtual cell model is trained based on pollutant exposure scenario data and cell state labels corresponding to the pollutant exposure scenario; obtaining target cell corresponding to the pollutant exposure scenario data to be evaluated; wherein the pollutant exposure scenario data to be evaluated includes pollutant mixture to be evaluated characterization, exposure condition to be evaluated characterization and baseline cell to be evaluated characterization; inputting the pollutant mixture to be evaluated characterization, the exposure condition to be evaluated characterization and the baseline cell to be evaluated characterization into the environmental virtual cell model to obtain the first cell state of the target cell; determining the first biological effect caused by the pollutant mixture to be evaluated according to the first cell state, and generating the risk assessment result of the pollutant mixture to be evaluated according to the first biological effect.

[0006] This invention utilizes an environmental virtual cell model to map the characterization of the pollutant mixture to be evaluated, the characterization of exposure conditions, and the baseline cell characterization to the cell state of target cells. The biological effects are then determined from the cell state, generating a risk assessment result. This allows the overall biological effects of the mixture to be reflected based on actual changes in cell state, overcoming the limitation of traditional techniques in predicting the real biological effects of pollutant mixtures on cells. Furthermore, for mixtures with different components, proportions, and exposure conditions, characterization data is acquired to complete the assessment, accurately characterizing the cell state of pollutant mixtures under real-world exposure scenarios and providing a comprehensive risk assessment. This enables rapid and reproducible risk assessment of pollutant mixtures.

[0007] In one alternative implementation, prior to obtaining the pre-trained environment virtual cell model, the method further includes: Acquire environmental virtual cell construction data and pollutant exposure scenario data. The environmental virtual cell construction data includes baseline state data of the target cell when it is not in a pollutant exposure scenario and cell perturbation response data of the target cell when it is in a pollutant exposure scenario. The pollutant exposure scenario data is encoded and characterized to obtain pollutant exposure scenario characterization data. Based on the baseline state data and the cell perturbation response data, cell state labels corresponding to the pollutant exposure scenario are determined. The pollutant exposure scenario characterization data includes pollutant mixture characterization, exposure condition characterization, and baseline cell characterization. Based on the pollutant exposure scenario characterization data and the cell state labels, a preset network model is trained to obtain an environmental virtual cell model.

[0008] This invention uses baseline state data and cell perturbation response data to determine cell state labels, and the model training data comes from actual cell responses, thus inputting data on real cell-level perturbation patterns into the model; at the same time, it uses an environmental virtual cell model for model construction and risk assessment, which can be repeatedly used to assess different mixtures without having to remodel for each assessment, thereby improving the biological reliability of the assessment results.

[0009] In one optional implementation, the pollutant exposure scenario data includes pollutant mixture data, exposure condition data, and baseline state data of target cells; the encoding and characterization processing of the pollutant exposure scenario data to obtain pollutant exposure scenario characterization data includes: Based on prior information about pollutants, each pollutant component in the pollutant mixture data is encoded to obtain the component characterization corresponding to each pollutant component, and the pollutant mixture characterization is determined based on the component characterization. The exposure condition data is encoded to obtain an exposure condition characterization; The baseline state data of the target cells are encoded to obtain baseline cell characterization.

[0010] This invention encodes each pollutant component separately based on prior information about the pollutants before determining the mixture characterization. This allows the mixture characterization to retain both the chemical properties of each component and the overall composition of the mixture, thereby supporting the evaluation of new component combinations not encountered during the training phase and expanding the model's applicability. By encoding and characterizing three types of information—pollutant mixture, exposure conditions, and baseline cells—separately, each characterization dimension is independent of the others. Selecting appropriate encoding methods for different data types helps improve the accuracy and stability of the model's predictions.

[0011] In one optional implementation, the step of training a preset network model based on the pollutant exposure scenario characterization data and the cell state labels to obtain an environmental virtual cell model includes: The pollutant exposure scenario characterization data are input into a preset network model to obtain the predicted cell state; The model training error is determined based on the predicted cell state and the cell state label. The parameters of the preset network model are updated based on the model training error. The steps of inputting the pollutant exposure scenario characterization data into the preset network model to obtain the predicted cell state are repeated until the model training error meets the preset convergence condition, thus obtaining the environmental virtual cell model.

[0012] This invention iteratively updates model parameters by predicting the error between cell state and cell state label, and controls training termination with preset convergence conditions, so that the model can converge stably with limited training data, avoiding overfitting or underfitting, and providing accuracy assurance for subsequent risk assessment of pollutant mixtures.

[0013] In one optional implementation, the step of determining the first biological effect caused by the pollutant mixture to be evaluated based on the first cell state, and generating a risk assessment result of the pollutant mixture to be evaluated based on the first biological effect, includes: The first cell state is compared with the baseline state of the target cell to determine the cell state change; The first biological effect caused by the pollutant mixture to be evaluated is determined based on the changes in cell state, and the type and intensity of the biological effect corresponding to the first biological effect are determined. The risk assessment results of the pollutant mixture to be evaluated are generated based on the type and intensity of the biological effects. The risk assessment results include risk screening results, risk treatment priorities, and risk levels.

[0014] This invention compares the predicted first cell state with the baseline state of the target cell itself, and uses changes in cell state as a criterion to determine the biological effects, their types, and intensities. This achieves a quantitative reflection of the nature and extent of the biological effects of the mixture, avoiding the bias caused by directly inferring the effects from external concentrations. The risk assessment results also include risk screening results, risk treatment priorities, and risk levels, supporting rapid screening and sorting of large-scale mixtures and providing a direct basis for hierarchical management and disposal decisions.

[0015] In one optional implementation, the characterization of the pollutant mixture to be evaluated includes single-component pollutant characterization corresponding to each pollutant component; the method further includes: The environmental virtual cell model is used to predict the second cell state of the target cells corresponding to each single-component pollutant characterization. The single-component biological effects corresponding to each pollutant component are determined based on the second cell state. The individual biological effects of each pollutant component were mixed to determine the second biological effect; The second biological effect is compared with the first biological effect, and the interaction relationship between each pollutant component is determined based on the comparison results.

[0016] This invention utilizes an environmental virtual cell model to predict the individual biological effects of each pollutant component upon exposure. The resulting second biological effect is then compared with the first biological effect to narrow the scope of subsequent experiments in determining whether the interactions between components are additive, synergistic, or antagonistic. This overcomes the limitations of traditional methods in analyzing the nature of mixed effects and identifying synergistic risk combinations, providing a direct basis for the formation analysis and targeted control of mixed-effect risks. Based on the prediction results, in vitro cell experiments are conducted for validation. The validation results provide updates and iterations for the model, gradually increasing its accuracy.

[0017] In one optional implementation, the method further includes: The pollutant components in the pollutant mixture to be evaluated are adjusted to obtain the adjusted characterization of the pollutant mixture to be evaluated; The environmental virtual cell model is used to predict the third cell state of the target cells corresponding to the adjusted pollutant mixture to be evaluated, and the third biological effect is determined based on the third cell state. Based on the difference between the third biological effect and the first biological effect, the risk contribution value corresponding to the pollutant component is determined; The risk assessment results are updated based on the risk contribution value.

[0018] This invention determines the risk contribution value of each component by extrapolating the deletion or adjustment of pollutant components in a mixture and comparing the differences in biological effects before and after the adjustment, thus achieving quantitative identification of risk-driven components. Based on this, updating the risk assessment results can clarify the main sources of risk in the mixture and avoid the waste of resources caused by indiscriminate control of all components.

[0019] Secondly, the present invention provides a mixture risk assessment device based on environmental virtual cells, the device comprising: An acquisition module is used to acquire an environmental virtual cell model; wherein, the environmental virtual cell model is trained based on pollutant exposure scenario data and cell state labels corresponding to the pollutant exposure scenario; The characterization module is used to acquire data on the pollutant exposure scenario to be evaluated corresponding to the target cells; wherein, the pollutant exposure scenario data to be evaluated includes characterization of the pollutant mixture to be evaluated, characterization of the exposure conditions to be evaluated, and characterization of the baseline cells to be evaluated; The prediction module is used to input the characterization of the pollutant mixture to be evaluated, the characterization of the exposure conditions to be evaluated, and the characterization of the baseline cells to be evaluated into the environmental virtual cell model to obtain the first cell state of the target cells; A generation module is used to determine the first biological effect caused by the pollutant mixture to be evaluated based on the first cell state, and to generate a risk assessment result of the pollutant mixture to be evaluated based on the first biological effect.

[0020] Thirdly, the present invention provides an electronic device comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the computer instructions to perform the mixture risk assessment method based on environmental virtual cells described in the first aspect or any corresponding embodiment thereof.

[0021] Fourthly, the present invention provides a computer-readable storage medium storing computer instructions for causing a computer to perform the mixture risk assessment method based on environmental virtual cells described in the first aspect or any corresponding embodiment thereof. Attached Figure Description

[0022] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0023] Figure 1This is a schematic diagram of an application scenario according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the first process of the mixture risk assessment method based on environmental virtual cells according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the second process of the mixture risk assessment method based on environmental virtual cells according to an embodiment of the present invention; Figure 4 This is a schematic diagram of the process for determining intracellular exposure characterization according to an embodiment of the present invention; Figure 5 This is a schematic diagram of the multi-route risk assessment process according to an embodiment of the present invention; Figure 6 This is a schematic flowchart illustrating the cellular response of the pollutant mixture as a whole and each pollutant component according to an embodiment of the present invention. Figure 7 This is a structural block diagram of a mixture risk assessment device based on environmental virtual cells according to an embodiment of the present invention; Figure 8 This is a schematic diagram of the hardware structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0025] It is understood that before using the technical solutions disclosed in the various embodiments of the present invention, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in the present invention and their authorization should be obtained in accordance with relevant laws and regulations through appropriate means.

[0026] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0027] As an optional application scenario of this invention, such as Figure 1 As shown, the mixture risk assessment system based on environmental virtual cells may include at least one terminal device and at least one server. Figure 1The system is illustrated in the example, which includes a computer 101, a mobile terminal 102, and a server 103, and the terminal devices such as the computer 101 and the mobile terminal 102 are connected to the server 103 through a network 110.

[0028] Specifically, the terminal device can be a smartphone, tablet, laptop, PDA, desktop computer, game console, smart TV, smart wearable device, in-vehicle terminal, VR (Virtual Reality) device, AR (Augmented Reality) device, etc. Server 103 can be a standalone physical server, a server cluster, a distributed system, or a cloud server providing cloud services. Network 110 can be a wired or wireless network, examples of which include, but are not limited to, the Internet, corporate intranet, local area network, wide area network, mobile communication network, and combinations thereof.

[0029] With the large-scale production and widespread application of chemicals, multiple pollutants coexist in media such as water, soil, air, food, and human biological samples, forming complex chemical mixtures with dynamically changing concentration ratios over time. Different components may produce additive, synergistic, or antagonistic mixed effects, making it difficult to accurately reflect the comprehensive biological effects and risks of the mixture in actual exposure scenarios based solely on the concentration or toxicity data of a single pollutant.

[0030] Currently, conventional technical approaches to risk assessment of chemical mixtures primarily rely on external exposure concentrations of pollutants, known toxic endpoints, and effect thresholds, employing dose-addition models, independent-action models, or risk quotient summation methods to quantitatively estimate the risk of mixtures. At the experimental level, in vitro cell assays can be used to determine response indicators such as changes in gene expression, biological pathway perturbations, and alterations in cell phenotypes induced by pollutant exposure. Simultaneously, computational toxicology and cell perturbation prediction methods utilize the chemical structure, biological activity information, target sites, baseline cell states, and existing perturbation response data of pollutants to model and predict the potential biological effects of pollutants.

[0031] However, the aforementioned technologies still have the following limitations: First, traditional prediction methods are mostly targeted at single pollutants, fixed toxicity endpoints, or specific exposure conditions, lacking the comprehensive characterization ability of mixture composition, ratio changes, exposure time, and differences in target cell baseline states, making it difficult to predict overall cell state changes under mixture exposure. Second, the external nominal exposure concentration often does not represent the actual effective dose inside the cell. After entering the cell, pollutants undergo processes such as uptake, distribution, transformation, efflux, and clearance, and differences in load within different subcellular organelles can trigger varying cell functional perturbations. Related technologies have failed to integrate intracellular fate processes, organelle-distinguished perturbation effects, cell state prediction, and mixture risk assessment into a unified framework, making it difficult to accurately reflect the cell response and overall risk induced by chemical mixtures in real exposure scenarios.

[0032] This invention provides a mixture risk assessment method based on environmental virtual cells. Utilizing an environmental virtual cell model, the characterization of the pollutant mixture to be assessed, the characterization of exposure conditions, and the baseline cell characterization are mapped to the cell state of target cells. The biological effects are then determined from the cell state, generating a risk assessment result. This method reflects the overall biological effects of the mixture based on actual changes in cell state, overcoming the limitation of traditional techniques that struggle to predict the real biological effects of pollutant mixtures on cells. Furthermore, for mixtures with different components, proportions, and exposure conditions, characterization data is acquired to complete the assessment, accurately representing the cell state of pollutant mixtures under real-world exposure scenarios and providing a comprehensive risk assessment. This enables rapid and reproducible risk assessment of pollutant mixtures.

[0033] According to an embodiment of the present invention, a mixture risk assessment method based on environmental virtual cells is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0034] This embodiment provides a mixture risk assessment method based on virtual environmental cells, which can be used on the aforementioned mobile terminals, such as mobile phones, tablets, etc. (the executing entity is described in conjunction with the actual situation). Figure 2 This is a flowchart of a mixture risk assessment method based on environmental virtual cells according to an embodiment of the present invention, such as... Figure 2 As shown, the process includes the following steps: S201, Obtain the virtual cell model of the environment.

[0035] In this embodiment, the environmental virtual cell model is trained based on pollutant exposure scenario data and cell state labels corresponding to the pollutant exposure scenarios. The environmental virtual cell model can be a model for predicting cell state transitions. The aforementioned environmental virtual cell model outputs cell states, and the model parameters learn the mapping relationship between pollutant exposure scenarios and cell state changes, enabling it to predict the corresponding cell states given a new exposure scenario.

[0036] Pollutant exposure scenario data can include pollutant mixture data, exposure condition data, and baseline state data of target cells. It can be used to describe "what pollutant acts on cells under what conditions", for example, "10 μmol / L bisphenol A acts on human liver cells and is continuously exposed for 24 hours", which constitutes a pollutant exposure scenario.

[0037] Cell state labels, which are the measured post-exposure cell states corresponding to a specific scenario, can be derived from actual measurements of exposed cells. Cell state can be represented by one or more of the following: gene expression, protein expression, metabolite levels, biological pathway activity, physiological state, or cell phenotype. For example, measuring the gene expression profile of hepatocytes 24 hours after exposure can serve as a cell state label for the corresponding scenario.

[0038] In this embodiment, the trained virtual cell model of the environment is acquired to execute the evaluation process. When initiating the risk assessment process, the network structure and parameters of the previously saved virtual cell model of the environment, along with the corresponding data processing rules, can be read from memory.

[0039] S202, Obtain data on the pollutant exposure scenario to be evaluated for the target cells.

[0040] In this embodiment, the target cell can refer to the cell targeted in this risk assessment, which is the object predicted by the model. For example, it can be primary cells or passaged cell lines such as animal-derived nerve cells or lung epithelial cells, or other cells selected according to the purpose of the assessment. The same pollutant mixture may cause different biological effects when applied to different cells.

[0041] The data on the exposure scenarios of the pollutants to be assessed include the characterization of the pollutant mixture, the characterization of the exposure conditions, and the characterization of the baseline cells to be assessed. These data can be used to describe: what the exposed pollutant is, under what conditions it is exposed, and in what state the cells are affected.

[0042] The characterization of the pollutant mixture to be evaluated can reflect which pollutant components are contained in the mixture and the relative composition of each component. For example, if an analysis of a pollutant mixture detects three pollutants, A, B, and C, the characterization of the pollutant mixture to be evaluated is obtained by pre-processing the identity and composition information of the three components.

[0043] The characterization of the exposure conditions to be evaluated can be the data representation obtained after pre-processing the exposure conditions set for this evaluation. Exposure conditions may include exposure dose or concentration, exposure duration, exposure mode, etc. For example, setting the exposure conditions as "10 μmol / L, continuous exposure for 24 hours" and obtaining the characterization of the exposure conditions to be evaluated after pre-processing.

[0044] The baseline cell characterization to be evaluated refers to the data representation obtained after pre-processing the target cells in the state they were not exposed to in this pollutant scenario. The baseline state can be represented by measurements such as gene expression, metabolite levels, or cell morphology of unexposed control cells, and the baseline cell characterization to be evaluated can be obtained after pre-processing.

[0045] Data on the pollutant exposure scenarios to be assessed corresponding to the target cells are obtained, and the exposure scenarios to be predicted are determined using the aforementioned data on pollutant exposure scenarios to be assessed.

[0046] S203 inputs the characterization of the pollutant mixture to be evaluated, the characterization of the exposure conditions to be evaluated, and the characterization of the baseline cells to be evaluated into the environmental virtual cell model to obtain the first cell state of the target cells.

[0047] In this embodiment, the first cell state can be the post-exposure cell state of the target cell predicted by the environmental virtual cell model for the pollutant exposure scenario to be evaluated. The content of the first cell state can correspond to the content of the cell state label, such as one or more of the predicted post-exposure gene expression profile, protein expression, metabolite level, biological pathway activity, organelle function, or cell phenotype changes.

[0048] The aforementioned data on the pollutant exposure scenario to be evaluated is input into an environmental virtual cell model, which then outputs a predicted first cell state. For example, the pollutant mixture characterization, exposure condition characterization, and baseline cell characterization corresponding to the scenario "a mixture of A and B, 10 μmol / L, continuous exposure for 24 hours, target cells are nerve cells" are input into the environmental virtual cell model. The model outputs the predicted post-exposure cell state of nerve cells under this scenario, resulting in a first cell state, such as predicted changes in gene expression or key biological pathway activity.

[0049] S204, determine the first biological effect caused by the pollutant mixture to be evaluated based on the first cell state, and generate the risk assessment result of the pollutant mixture to be evaluated based on the first biological effect.

[0050] In this embodiment, the first biological effect can refer to the biological changes caused by exposure of the pollutant mixture to be evaluated to the target cells, which is a biological interpretation of the changes reflected by the first cellular state. Biological effects may include changes in gene expression, protein expression, metabolite levels, disturbances in biological pathways (such as oxidative stress, activation of inflammatory response-related pathways), changes in organelle function, changes in physiological state, and changes in cell phenotype (such as decreased cell viability, morphological changes), etc.

[0051] Risk assessment results can be used to characterize the risk status of the pollutant mixture being assessed. These results may include risk screening results (such as whether there are any noteworthy biological effects), risk treatment priorities (the treatment order of the mixture), and risk levels (such as high, medium, and low).

[0052] This embodiment analyzes the first cell state to identify the biological changes caused by the pollutant mixture to be evaluated, thereby determining the biological effects caused by the mixture. For example, if the first cell state shows a significant upregulation of oxidative stress-related genes and a marked change in the activity of mitochondrial-related pathways, it can be determined that the pollutant mixture causes oxidative stress-related biological effects.

[0053] Based on this, the type and intensity of the corresponding biological effects are determined according to the identified biological effects, and corresponding risk conclusions are given according to preset judgment rules. For example, when the intensity of a biological effect exceeds a preset concern level, a "high" risk level is output, and risk treatment priorities are generated among multiple mixtures to be evaluated according to the intensity of the effect.

[0054] This embodiment provides a mixture risk assessment method based on environmental virtual cells. Using an environmental virtual cell model, it maps the characterization of the pollutant mixture to be assessed, the characterization of exposure conditions, and the baseline cell characterization to the cell state of target cells. The biological effects are then determined from the cell state, generating a risk assessment result. This method reflects the overall biological effects of the mixture based on actual changes in cell state, overcoming the limitation of traditional techniques in predicting the real biological effects of pollutant mixtures on cells. Furthermore, for mixtures with different components, proportions, and exposure conditions, it acquires characterization data to complete the assessment, accurately characterizing the cell state of pollutant mixtures under real exposure scenarios and providing a comprehensive risk assessment. This enables rapid and reproducible risk assessment of pollutant mixtures.

[0055] This embodiment provides a mixture risk assessment method based on virtual environmental cells, which can be used on the aforementioned mobile terminals, such as mobile phones and tablets.Figure 3 This is a flowchart of a mixture risk assessment method based on environmental virtual cells according to an embodiment of the present invention, such as... Figure 3 As shown, the process includes the following steps: S301, acquire environmental virtual cell construction data and pollutant exposure scenario data.

[0056] In this embodiment, the environmental virtual cell construction data includes baseline state data of the target cells when they are not in a pollutant exposure scenario and cell perturbation response data of the target cells when they are in a pollutant exposure scenario. The baseline state data can be used to characterize the initial state of the target cells under conditions without pollutant perturbation, and can come from measurement results of unexposed cells, solvent control cells, or other control cells matched with the exposed sample. The cell perturbation response data can be used to characterize the measured state of the target cells after pollutant exposure. Each perturbation record can include pollutant identification information, exposure level, exposure time, target cell information, and the measured cell state after exposure. The pollutant exposure scenario data can include pollutant mixture data (such as information on each pollutant component in the mixture and its corresponding exposure amount and component ratio), exposure condition data (such as exposure dose, exposure duration, exposure frequency, exposure order, etc.), and baseline state data of the target cells.

[0057] In this embodiment, pollutant mixture data can be obtained from pollutant databases or publicly available literature, and recorded and archived cell baseline state data and cell perturbation response data can be read from experimental data management systems. This invention does not impose any specific limitations.

[0058] S302, the pollutant exposure scenario data is encoded and characterized to obtain pollutant exposure scenario characterization data, and based on baseline state data and cell perturbation response data, the cell state label corresponding to the pollutant exposure scenario is determined.

[0059] In this embodiment, the pollutant exposure scenario characterization data includes pollutant mixture characterization, exposure condition characterization, and baseline cell characterization. The pollutant exposure scenario data undergoes structured analysis and feature extraction. The extracted information is then encoded and characterized to obtain the pollutant exposure scenario characterization data. Subsequently, using the baseline state data before pollutant exposure as a reference for cell state changes, and combining it with cell perturbation response data collected under the corresponding exposure scenarios, the state changes of cells under different pollutant exposure conditions are analyzed. This determines the cell state labels corresponding to each pollutant exposure scenario, establishing the input-output correspondence for subsequent training of the virtual cell model of the environment.

[0060] Specifically, S302 includes: S3021, based on prior information about pollutants, each pollutant component in the pollutant mixture data is encoded to obtain the component characterization corresponding to each pollutant component, and the pollutant mixture characterization is determined based on the component characterization.

[0061] In this embodiment, the pollutant exposure scenario data includes pollutant mixture data, exposure condition data, and baseline state data of target cells. Prior pollutant information can refer to relevant background information about known pollutants, and may include one or more of the following: standardized pollutant identification, chemical structure, physicochemical properties, target sites, binding information, pathway activity information, and biological activity information, derived from pollutant databases, publicly available literature, existing experimental data, or computational prediction results. Pollutant mixture data can refer to the identity information of each pollutant component in the mixture and its corresponding exposure amount (e.g., concentration, dose, mole fraction, mass fraction) or component dose ratio. Component characterization can be a representation obtained by encoding individual pollutant components, such as a fixed-dimensional molecular fingerprint vector.

[0062] This embodiment extracts and encodes the chemical structure, physicochemical properties, biological activity, and target sites of each pollutant component based on prior information about the pollutants, obtaining the component characterization corresponding to each pollutant component. Then, based on the characterization of each component and its compositional relationship in the pollutant mixture, the characteristics of each pollutant component are combined or aggregated to characterize the composition and characteristics of the pollutant mixture as a whole, obtaining the pollutant mixture characterization and providing input for subsequent model predictions.

[0063] In one example, the pollutant mixture characterization can be obtained by weighted aggregation of component characterization based on the exposure level or dose ratio of each pollutant component. When pollutant components have different mechanisms of action, each component can also be divided into one or more effect groups according to the mechanism of action, and then the effect group characterizations can be formed separately and integrated to obtain the pollutant mixture characterization.

[0064] S3022, Encode the exposure condition data to obtain the exposure condition characterization.

[0065] In this embodiment, exposure condition data can characterize data used to describe various elements of the exposure process, and may include one or more of the following: exposure dose, exposure concentration, component ratio, exposure duration, exposure time point, exposure frequency, exposure sequence, exposure interval, and exposure mode. The exposure mode can include single exposure, repeated exposure, continuous exposure, intermittent exposure, and exposure that varies over time. For example, in a scenario of intermittent exposure "3 times a week, 6 hours each time, for 4 weeks," the number of exposures, the duration of each exposure, and the exposure interval all fall within the scope of exposure condition data.

[0066] By encoding the exposure conditions, the aforementioned exposure condition elements are converted into numerical features or embedding vectors to obtain an exposure sequence representation as the exposure condition representation. For example, for a single exposure, the exposure level, start time, duration, and sampling time are represented; for repeated or intermittent exposures, the level, interval, and number of each exposure are further represented; for exposures where composition, dose, or dose ratio changes over time, the exposure process is divided into multiple time periods, and the pollutant composition, component ratio, and duration in each time period are recorded, forming a dynamic exposure sequence representation in chronological order. It should be noted that exposing pollutant A before pollutant B, and exposing pollutant B before pollutant A, results in two different dynamic exposure sequence representations.

[0067] S3023 encodes the baseline state data of the target cells to obtain baseline cell characterization.

[0068] In this embodiment, baseline state data can be state measurement data of target cells when they are not exposed to pollutants, and can be represented by one or more of the following: genomics, transcriptomics, epigenomics, proteomics, metabolomics, spatial omics, cell morphology, and cell imaging data. For example, transcriptome sequencing results, cell morphology images, or metabolite profiles of nerve cells under solvent control conditions can all be used as baseline state data.

[0069] Specifically, the baseline state data of target cells is subjected to dimensionality reduction, feature selection, encoding, or representation learning, and one or more of the target cell type, species, tissue origin, culture conditions, and physiological state are added as auxiliary information to the baseline cell characterization. For multimodal baseline data, these can be encoded separately and then fused in a unified representation space.

[0070] S303, based on pollutant exposure scenario characterization data and cell state labels, trains a preset network model to obtain an environmental virtual cell model.

[0071] Specifically, S303 includes: S3031, input the pollutant exposure scenario characterization data into the preset network model to obtain the predicted cell state.

[0072] In this embodiment, the preset network model can refer to a pre-built network model type, which can be a machine learning model capable of learning cell state transitions, such as an autoencoder model, graph neural network model, Transformer model, diffusion model, or flow model. The environmental virtual cell model can be a model capable of predicting the transition of target cells from a baseline state to a post-exposure state under a given exposure scenario.

[0073] Specifically, the pollutant mixture characterization, exposure condition characterization, and baseline cell characterization are input into the preset network model according to the input format specified by the model. Each type of characterization can be processed by its corresponding processing unit to output the predicted cell state. In the early stages of training, the model parameters have not yet been properly adjusted, and there may be a significant deviation between the predicted cell state and the actual state.

[0074] S3032, determine the model training error based on the predicted cell state and cell state label.

[0075] In this embodiment, the model training error can be used to quantify the degree to which the predicted cell state deviates from the cell state label, and can be calculated using loss functions such as mean squared error and cross-entropy. For example, when the cell state label is the post-exposure state itself, the model training error can be the difference between the predicted cell state and the post-exposure state; when the label is a state change, the model training error can be the difference between the predicted change and the measured change. With the available data, regularization terms that utilize dose-time relationships, pollutant-biomolecule interaction information, or biological pathway relationships to assist in model constraint can be added to the training error, ensuring that the training process conforms to known biological laws.

[0076] S3033, update the parameters of the preset network model according to the model training error, return the step of inputting the pollutant exposure scenario characterization data into the preset network model to obtain the predicted cell state, until the model training error meets the preset convergence condition, and obtain the environmental virtual cell model.

[0077] In this embodiment, the preset convergence condition may include the model training error being lower than a preset error threshold; however, this invention does not impose any specific limitations.

[0078] Specifically, the direction and amount of adjustment for the model parameters are determined based on the training error, and the parameters of the network model are updated. Then, the updated network model is used to re-predict and recalculate the training error. Through multiple iterations, the model parameters are continuously optimized until the training error meets the preset convergence condition, thereby obtaining an environmental virtual cell model that can characterize the relationship between pollutant exposure and changes in cell state.

[0079] During the training process, validation data can be used to continuously evaluate prediction errors, the correlation between prediction results and measured results, and the consistency of perturbations in key biological pathways, in order to monitor training quality.

[0080] S304, Obtain the virtual cell model of the environment.

[0081] Please see details Figure 1 S201 of the illustrated embodiment will not be described again here.

[0082] S305, acquire data on the pollutant exposure scenario to be evaluated for the target cells.

[0083] Please see details Figure 1 S202 of the illustrated embodiment will not be described again here.

[0084] In an optional embodiment of this example, the data on the aforementioned pollutant exposure scenario to be evaluated may further include the external nominal exposure level of the pollutant mixture and exposure information related to the pollutant entering the target cell and undergoing changes within the cell. For example... Figure 4 As shown, Figure 4 This is a flowchart for determining intracellular exposure characterization according to an embodiment of the present invention.

[0085] Specifically, the free extracellular concentration of pollutants is determined based on the aforementioned nominal external exposure level to characterize the actual exposure level at which pollutants can participate in cellular uptake before entering the target cell. After determining the free extracellular concentration, the intracellular concentration of pollutants at different time points is determined based on the cellular uptake, efflux, and clearance processes of pollutants. Furthermore, based on the intracellular and subcellular distribution of pollutants, the distribution of pollutants in subcellular compartments such as the cytoplasm, nucleus, mitochondria, lysosomes, endoplasmic reticulum, Golgi apparatus, lipid droplets, and cell membrane is determined.

[0086] Furthermore, based on the pollutant's molecular structure, the expression levels of metabolic enzymes and transport proteins in the target cells, and pre-defined biotransformation rules, the transformation products corresponding to the parent pollutant and their generation status can be determined. The composition of the parent pollutant and transformation products at different time points can be updated to obtain an intracellular mixture. Based on the updated parent pollutant and transformation products, one or more of the following can be obtained at different time points: free intracellular concentration, total intracellular concentration, subcellular compartment load, and target site dose. The acquired data can then be integrated to form an intracellular exposure characterization corresponding to the aforementioned pollutant exposure scenario to be evaluated.

[0087] S306. Input the characterization of the pollutant mixture to be evaluated, the characterization of the exposure conditions to be evaluated, and the characterization of the baseline cells to be evaluated into the environmental virtual cell model to obtain the first cell state of the target cells.

[0088] Please see details Figure 1 S203 of the illustrated embodiment will not be described again here.

[0089] S307, determine the first biological effect caused by the pollutant mixture to be evaluated based on the first cell state, and generate the risk assessment result of the pollutant mixture to be evaluated based on the first biological effect.

[0090] In the optional scheme of this embodiment, based on the research objectives, existing data, and prior understanding of pollutants and biological effects, pre-defined environmental virtual cells, non-target analysis methods, and computational toxicology methods can be combined in different application sequences to obtain risk assessment results for the pollutant mixture to be evaluated. Different combinations can be implemented separately, or the pollutant identification results, cell state prediction results, and risk assessment results obtained from different routes can be compared or integrated. For example... Figure 5 As shown, Figure 5 This is a schematic diagram of the multi-route risk assessment process according to an embodiment of the present invention.

[0091] Specifically, in Route 1, a top-down assessment approach can be adopted, which starts from pre-determined toxic effects, toxic endpoints, biological pathways, or cell states, uses computational toxicology methods to determine the scope of pollutant analysis related to the effects of concern, then uses non-target analysis methods to identify the corresponding pollutants and their exposure levels from actual environmental samples, and then constructs a pollutant mixed exposure scenario based on the actual identified pollutant composition, exposure level, exposure duration, and target cell baseline state, and uses environmental virtual cells to evaluate the cellular response and risk of the pollutant mixture.

[0092] In Route 2, one or more methods, such as non-target analysis, suspected screening, or targeted analysis, can be used to identify pollutants and their exposure levels from actual environmental samples. Then, computational toxicology methods are used to determine the toxic endpoints, effect thresholds, reference doses, or other risk benchmarks corresponding to the identified pollutants, and to form initial risk results. Afterward, the composition of pollutants, exposure levels, exposure times, and target cell baseline states in the actual environmental samples are input into virtual environmental cells to predict changes in cell state when the pollutant mixture as a whole and when the preferred pollutants are exposed individually or in combination. The predicted cell responses are then compared with the initial risk results to verify the initial risk results and analyze possible mechanisms of action.

[0093] In Route 3, a bottom-up assessment approach can be adopted. First, virtual lifecycle screening is conducted using environmental virtual cells from a pre-defined pollutant space. Candidate pollutants are screened based on their cellular availability, intracellular fate, and predicted responses to parent pollutants and their transformation products. Then, computational toxicology methods are used to determine the corresponding toxic endpoints, effect thresholds, or risk benchmarks for the candidate pollutants, and the risk or priority of the candidate pollutants is determined based on the prediction results. Finally, non-target analysis methods are used to confirm the presence of candidate pollutants in actual environmental samples, and a risk assessment is completed in conjunction with actual exposure levels. For pollutants that may undergo biotransformation, the predicted transformation products can also be included in the candidate pollutant pool.

[0094] In Route 4, non-target analysis methods can be used to obtain the composition and exposure levels of pollutants in actual environmental samples. Then, environmental virtual cells can be used to conduct comprehensive risk screening and multi-endpoint assessment of single pollutants or pollutant mixtures. Based on the predicted changes in cell state, the toxic endpoints that need to be prioritized for evaluation can be determined. Finally, computational toxicology methods can be used to quantitatively analyze the toxic endpoints and, in combination with the contribution of pollutant components, determine the risk-driving components that need to be prioritized for attention.

[0095] Furthermore, in any of the above assessment routes, if there is significant uncertainty in pollutant structure annotation, exposure level, or model prediction, the corresponding uncertainty information can be retained and used to determine the priority order of subsequent standard validation, targeted quantification, or experimental verification. The risk assessment results may also include risk screening results, risk priority, risk value, risk level, key biological pathways, key organelle responses, mixed effect types, contributions of each pollutant component, candidate risk-driving components, and the confidence level of the results combined with the model's applicable domain and prediction uncertainty.

[0096] Specifically, the above S307 includes: S3071 compares the first cell state with the baseline state of the target cell to determine the cell state change.

[0097] In this embodiment, changes in cell state can be used to characterize the differences in cell state before and after exposure. These changes can manifest as one or more of the following: changes in gene expression, protein expression, metabolites, biological pathway activity, physiological state, organelle function, or cell phenotype. For example, the differences can be quantified using methods such as difference, ratio, logarithmic fold change, or standardized change. For instance, comparing the predicted post-exposure transcriptome with the baseline transcriptome gene by gene yields a differential expression vector, which is used as the cell state change.

[0098] Specifically, using the baseline state of target cells before pollutant exposure as a reference, the first cell state is compared with the baseline state in corresponding dimensions or characteristics. Changes in cell indicators across different states are analyzed, and the cell state changes are determined based on the comparison results. These cell state changes can be used to characterize alterations in cell function, molecular characteristics, or phenotypes caused by pollutant exposure, providing a basis for subsequent biological effect assessment and risk evaluation.

[0099] S3072, determine the first biological effect caused by the pollutant mixture to be evaluated based on changes in cell state, and determine the type and intensity of the biological effect corresponding to the first biological effect.

[0100] In this embodiment, the first biological effect can be used to characterize the biological interpretation of changes in cellular state, such as oxidative stress, activation of inflammatory pathways, metabolic disorders, and organelle dysfunction, and is a general designation of the overall exposure effect of the mixture. The type of biological effect can be used to characterize the biological category to which the effect belongs; the intensity of the biological effect can be used to characterize the severity or magnitude of the effect.

[0101] Specifically, characteristic analysis can be performed on changes in cell state to determine the direction, degree, and duration of changes in different cell state indicators relative to the baseline state. Based on these changes, the primary biological effect caused by the pollutant mixture being evaluated can be determined. Subsequently, the primary biological effect is classified according to the biological significance of the changes in cell state, determining its corresponding biological effect type. Simultaneously, the primary biological effect is quantified based on the degree, range, or deviation from the baseline state of the cell state changes, determining the corresponding biological effect intensity.

[0102] S3073 generates risk assessment results for the mixture of pollutants to be assessed based on the type and intensity of biological effects. The risk assessment results include risk screening results, risk treatment priorities, and risk levels.

[0103] In this embodiment, it can be determined whether the pollutant mixture to be evaluated produces a biological effect requiring attention based on the type of biological effect, and a corresponding risk screening result can be generated. Then, based on the type and intensity of the biological effect, the risk of the pollutant mixture to be evaluated is ranked, and the corresponding risk treatment priority is determined. Specifically, a corresponding effect threshold or risk benchmark can be set, and the predicted biological effect is compared with the corresponding effect threshold or risk benchmark. Based on the comparison result, the risk level of the pollutant mixture to be evaluated is determined, thereby obtaining a risk assessment result including risk screening results, risk treatment priority, and risk level.

[0104] In an optional embodiment, the environmental virtual cell model is used to predict the second cell state of the target cells corresponding to each single-component pollutant characterization; the single-component biological effects corresponding to each pollutant component are determined based on the second cell state; the single-component biological effects of each pollutant component are mixed to determine the second biological effect; the second biological effect is compared with the first biological effect, and the interaction relationship between each pollutant component is determined based on the comparison result.

[0105] The characterization of the pollutant mixture to be evaluated includes the single-component pollutant characterization corresponding to each pollutant component. The single-component pollutant characterization can be the characterization of each pollutant component in the mixture as an individual exposure substance, and the characterization of each component already encoded during the mixture characterization process can be directly reused. The second cell state can be the predicted cell state output by the environmental virtual cell model under the individual exposure scenario of each pollutant component, and its exposure conditions (such as exposure duration) can be set to be consistent with the pollutant mixture scenario. The single-component biological effect can be the individual exposure effect of each component obtained by parsing the second cell state. The second biological effect can be a reference mixture effect obtained by mixing the single-component biological effects according to a preset reference model. For example, for pollutant components with the same mode of action or that can be converted according to equivalent doses, a dose-addition model can be used; for pollutant components with independent modes of action, an independent action model can be used. The interaction relationship can refer to the additive, synergistic, or antagonistic effects between the pollutant components. Figure 6 As shown, Figure 6 This is a schematic flowchart illustrating the cellular response of the pollutant mixture as a whole and each pollutant component according to an embodiment of the present invention.

[0106] Specifically, individual exposure scenarios are constructed for each pollutant component in the mixture, and the corresponding second cell states are predicted using an environmental virtual cell model. Then, based on the changes in each second cell state relative to the baseline state, the individual biological effects corresponding to each pollutant component are determined. Next, based on the individual biological effects, a pre-set reference model (such as a dose-addition model or an independent-action model) is used to determine the second biological effect, i.e., the reference mixed effect. Finally, the second biological effect is compared with the first biological effect corresponding to overall exposure to the pollutant mixture, and combined with a pre-set deviation range, the interaction relationship between the pollutant components is determined based on the difference between the two. When the difference is within the corresponding range, it can be determined as an additive effect consistent with the reference model; when it exceeds the corresponding range, a synergistic or antagonistic effect is determined based on the direction of the difference.

[0107] In another optional embodiment, the pollutant components in the pollutant mixture to be evaluated are adjusted to obtain an adjusted pollutant mixture characterization; the third cell state of the target cells corresponding to the adjusted pollutant mixture characterization is predicted using the environmental virtual cell model, and a third biological effect is determined based on the third cell state; the risk contribution value corresponding to the corresponding pollutant component is determined according to the difference between the third biological effect and the first biological effect; and the risk assessment result is updated according to the risk contribution value.

[0108] The adjustment can involve deleting or altering the exposure levels of pollutant components in the pollutant mixture to be evaluated, including one or more of the following: deleting at least one pollutant component from the mixture, or altering the exposure level of at least one pollutant component. The adjusted characterization of the pollutant mixture to be evaluated can be a regenerated characterization of the adjusted mixture. The third cell state and the third biological effect are the predicted cell states and their corresponding biological effects under the adjusted pollutant mixture exposure scenario, respectively. The risk contribution value can be used to quantify the degree to which the corresponding pollutant component contributes to the overall risk of the mixture.

[0109] Specifically, such as Figure 6 As shown, for one or more pollutant components in a pollutant mixture to be evaluated, the original pollutant mixture can be adjusted by deleting the corresponding components or adjusting their exposure levels according to a preset ratio, and the adjusted pollutant mixture characterization can be obtained. Then, the adjusted pollutant mixture characterization is input into an environmental virtual cell model to predict the corresponding third cell state, and the adjusted third biological effect is determined based on the third cell state. The third biological effect is compared with the first biological effect corresponding to the original pollutant mixture, and the risk contribution value of the corresponding pollutant component is determined based on the degree of change in the biological effect before and after adjustment. Finally, based on the risk contribution values ​​corresponding to each pollutant component, the original risk assessment results are updated to determine the risk contribution of each pollutant component and candidate risk-driving components.

[0110] This embodiment provides a mixture risk assessment method based on environmental virtual cells. Using an environmental virtual cell model, it maps the characterization of the pollutant mixture to be assessed, the characterization of exposure conditions, and the baseline cell characterization to the cell state of target cells. The biological effects are then determined from the cell state, generating a risk assessment result. This method reflects the overall biological effects of the mixture based on actual changes in cell state, overcoming the limitation of traditional techniques in predicting the real biological effects of pollutant mixtures on cells. Furthermore, for mixtures with different components, proportions, and exposure conditions, it acquires characterization data to complete the assessment, accurately characterizing the cell state of pollutant mixtures under real exposure scenarios and providing a comprehensive risk assessment. This enables rapid and reproducible risk assessment of pollutant mixtures.

[0111] This embodiment also provides a mixture risk assessment device based on environmental virtual cells, which is used to implement the above embodiments and preferred embodiments, and will not be repeated as already described. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0112] This embodiment provides a mixture risk assessment device based on environmental virtual cells, such as... Figure 7 As shown, it includes: The acquisition module 701 is used to acquire an environmental virtual cell model; wherein the environmental virtual cell model is trained based on pollutant exposure scenario data and cell state labels corresponding to the pollutant exposure scenario; The characterization module 702 is used to acquire the pollutant exposure scenario data corresponding to the target cells; wherein, the pollutant exposure scenario data includes the characterization of the pollutant mixture to be evaluated, the characterization of the exposure conditions to be evaluated, and the characterization of the baseline cells to be evaluated. Prediction module 703 is used to input the characterization of the pollutant mixture to be evaluated, the characterization of the exposure conditions to be evaluated, and the characterization of the baseline cells to be evaluated into the environmental virtual cell model to obtain the first cell state of the target cells; The generation module 704 is used to determine the first biological effect caused by the pollutant mixture to be evaluated based on the first cell state, and to generate a risk assessment result of the pollutant mixture to be evaluated based on the first biological effect.

[0113] In some alternative embodiments, the apparatus further includes: Training module 705 is specifically used for: Acquire environmental virtual cell construction data and pollutant exposure scenario data. The environmental virtual cell construction data includes baseline state data of the target cell when it is not in a pollutant exposure scenario and cell perturbation response data of the target cell when it is in a pollutant exposure scenario. The pollutant exposure scenario data is encoded and characterized to obtain pollutant exposure scenario characterization data. Based on the baseline state data and the cell perturbation response data, cell state labels corresponding to the pollutant exposure scenario are determined. The pollutant exposure scenario characterization data includes pollutant mixture characterization, exposure condition characterization, and baseline cell characterization. Based on the pollutant exposure scenario characterization data and the cell state labels, a preset network model is trained to obtain an environmental virtual cell model.

[0114] In some alternative implementations, the training module 705 is further configured to: Based on prior information about pollutants, each pollutant component in the pollutant mixture data is encoded to obtain the component characterization corresponding to each pollutant component, and the pollutant mixture characterization is determined based on the component characterization. The exposure condition data is encoded to obtain an exposure condition characterization; The baseline state data of the target cells are encoded to obtain baseline cell characterization.

[0115] In some alternative implementations, the training module 705 is further configured to: The pollutant exposure scenario characterization data are input into a preset network model to obtain the predicted cell state; The model training error is determined based on the predicted cell state and the cell state label. The parameters of the preset network model are updated based on the model training error. The steps of inputting the pollutant exposure scenario characterization data into the preset network model to obtain the predicted cell state are repeated until the model training error meets the preset convergence condition, thus obtaining the environmental virtual cell model.

[0116] In some alternative implementations, the generation module 704 is specifically used for: The first cell state is compared with the baseline state of the target cell to determine the cell state change; The first biological effect caused by the pollutant mixture to be evaluated is determined based on the changes in cell state, and the type and intensity of the biological effect corresponding to the first biological effect are determined. The risk assessment results of the pollutant mixture to be evaluated are generated based on the type and intensity of the biological effects. The risk assessment results include risk screening results, risk treatment priorities, and risk levels.

[0117] In some alternative embodiments, the apparatus further includes: Comparison module 706 is specifically used for: The environmental virtual cell model is used to predict the second cell state of the target cells corresponding to each single-component pollutant characterization. The single-component biological effects corresponding to each pollutant component are determined based on the second cell state. The individual biological effects of each pollutant component were mixed to determine the second biological effect; The second biological effect is compared with the first biological effect, and the interaction relationship between each pollutant component is determined based on the comparison results.

[0118] In some alternative embodiments, the apparatus further includes: The 707 update module is specifically used for: The pollutant components in the pollutant mixture to be evaluated are adjusted to obtain the adjusted characterization of the pollutant mixture to be evaluated; The environmental virtual cell model is used to predict the third cell state of the target cells corresponding to the adjusted pollutant mixture to be evaluated, and the third biological effect is determined based on the third cell state. Based on the difference between the third biological effect and the first biological effect, the risk contribution value corresponding to the pollutant component is determined; The risk assessment results are updated based on the risk contribution value.

[0119] The mixture risk assessment device based on environmental virtual cells provided in this embodiment of the invention can execute the mixture risk assessment method based on environmental virtual cells provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects for executing the method. Further functional descriptions of the above modules and units are the same as in the corresponding embodiments described above, and will not be repeated here.

[0120] Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention.

[0121] The following is a detailed reference. Figure 8 This diagram illustrates a suitable structural schematic for implementing an electronic device according to embodiments of the present invention. The electronic device may include a processor (e.g., a central processing unit, graphics processor, etc.) 801, which can perform various appropriate actions and processes based on a program stored in read-only memory (ROM) 802 or a program loaded from memory 808 into random access memory (RAM) 803. The RAM 803 also stores various programs and data required for the operation of the electronic device. The processor 801, ROM 802, and RAM 803 are interconnected via a bus 804. An input / output (I / O) interface 805 is also connected to the bus 804.

[0122] Typically, the following devices can be connected to I / O interface 805: input devices 806 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 807 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; memory devices 808 including, for example, magnetic tapes, hard disks, etc.; and communication devices 809. Communication device 809 allows electronic devices to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 8 Electronic devices with various devices are shown, but it should be understood that it is not required to implement or have all of the devices shown, and more or fewer devices may be implemented or have instead.

[0123] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device 809, or installed from a memory 808, or installed from a ROM 802. When the computer program is executed by the processor 801, it performs the functions defined in the mixture risk assessment method based on environmental virtual cells according to embodiments of the present invention.

[0124] Figure 8 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments of the present invention.

[0125] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded over a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code. When the software or computer code is accessed and executed by the computer, processor, or hardware, the hybrid risk assessment method based on environmental virtual cells shown in the above embodiments is implemented.

[0126] A portion of this invention can be applied as a computer program product, such as computer program instructions, which, when executed by a computer, can invoke or provide the methods and / or technical solutions according to the invention through the operation of the computer. Those skilled in the art will understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, installation package files, etc. Correspondingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executing the instructions, or the computer compiling the instructions and then executing the corresponding compiled program, or the computer reading and executing the instructions, or the computer reading and installing the instructions and then executing the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to a computer.

[0127] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A mixture risk assessment method based on virtual environmental cells, characterized in that, The method includes: Obtain an environmental virtual cell model; wherein the environmental virtual cell model is trained based on pollutant exposure scenario data and cell state labels corresponding to the pollutant exposure scenarios; Acquire data on the pollutant exposure scenario corresponding to the target cells; wherein, the pollutant exposure scenario data includes the characterization of the pollutant mixture to be evaluated, the characterization of the exposure conditions to be evaluated, and the characterization of the baseline cells to be evaluated; The characterization of the pollutant mixture to be evaluated, the characterization of the exposure conditions to be evaluated, and the characterization of the baseline cells to be evaluated are input into the environmental virtual cell model to obtain the first cell state of the target cells; The first biological effect caused by the pollutant mixture to be evaluated is determined based on the first cell state, and a risk assessment result of the pollutant mixture to be evaluated is generated based on the first biological effect.

2. The mixture risk assessment method based on environmental virtual cells according to claim 1, characterized in that, Prior to acquiring the virtual cell model of the environment, the method further includes: Acquire environmental virtual cell construction data and pollutant exposure scenario data. The environmental virtual cell construction data includes baseline state data of the target cell when it is not in a pollutant exposure scenario and cell perturbation response data of the target cell when it is in a pollutant exposure scenario. The pollutant exposure scenario data is encoded and characterized to obtain pollutant exposure scenario characterization data. Based on the baseline state data and the cell perturbation response data, cell state labels corresponding to the pollutant exposure scenario are determined. The pollutant exposure scenario characterization data includes pollutant mixture characterization, exposure condition characterization, and baseline cell characterization. Based on the pollutant exposure scenario characterization data and the cell state labels, a preset network model is trained to obtain an environmental virtual cell model.

3. The mixture risk assessment method based on environmental virtual cells according to claim 2, characterized in that, The pollutant exposure scenario data includes pollutant mixture data, exposure condition data, and baseline state data of target cells; The encoding and characterization of the pollutant exposure scenario data to obtain pollutant exposure scenario characterization data includes: Based on prior information about pollutants, each pollutant component in the pollutant mixture data is encoded to obtain the component characterization corresponding to each pollutant component, and the pollutant mixture characterization is determined based on the component characterization. The exposure condition data is encoded to obtain an exposure condition characterization; The baseline state data of the target cells are encoded to obtain baseline cell characterization.

4. The mixture risk assessment method based on environmental virtual cells according to claim 2, characterized in that, The step of training a preset network model based on the pollutant exposure scenario characterization data and the cell state labels to obtain an environmental virtual cell model includes: The pollutant exposure scenario characterization data are input into a preset network model to obtain the predicted cell state; The model training error is determined based on the predicted cell state and the cell state label. The parameters of the preset network model are updated based on the model training error. The steps of inputting the pollutant exposure scenario characterization data into the preset network model to obtain the predicted cell state are repeated until the model training error meets the preset convergence condition, thus obtaining the environmental virtual cell model.

5. The mixture risk assessment method based on environmental virtual cells according to claim 1, characterized in that, The step of determining the first biological effect caused by the pollutant mixture to be evaluated based on the first cell state, and generating a risk assessment result for the pollutant mixture to be evaluated based on the first biological effect, includes: The first cell state is compared with the baseline state of the target cell to determine the cell state change; The first biological effect caused by the pollutant mixture to be evaluated is determined based on the changes in cell state, and the type and intensity of the biological effect corresponding to the first biological effect are determined. The risk assessment results of the pollutant mixture to be evaluated are generated based on the type and intensity of the biological effects. The risk assessment results include risk screening results, risk treatment priorities, and risk levels.

6. The mixture risk assessment method based on environmental virtual cells according to claim 5, characterized in that, The characterization of the pollutant mixture to be evaluated includes the characterization of individual pollutants corresponding to each pollutant component; the method further includes: The environmental virtual cell model is used to predict the second cell state of the target cells corresponding to each single-component pollutant characterization. The single-component biological effects corresponding to each pollutant component are determined based on the second cell state. The individual biological effects of each pollutant component were mixed to determine the second biological effect; The second biological effect is compared with the first biological effect, and the interaction relationship between each pollutant component is determined based on the comparison results.

7. The mixture risk assessment method based on environmental virtual cells according to claim 6, characterized in that, The method further includes: The pollutant components in the pollutant mixture to be evaluated are adjusted to obtain the adjusted characterization of the pollutant mixture to be evaluated; The environmental virtual cell model is used to predict the third cell state of the target cells corresponding to the adjusted pollutant mixture to be evaluated, and the third biological effect is determined based on the third cell state. Based on the difference between the third biological effect and the first biological effect, the risk contribution value corresponding to the pollutant component is determined; The risk assessment results are updated based on the risk contribution value.

8. A mixture risk assessment device based on environmental virtual cells, characterized in that, The device includes: An acquisition module is used to acquire an environmental virtual cell model; wherein, the environmental virtual cell model is trained based on pollutant exposure scenario data and cell state labels corresponding to the pollutant exposure scenario; The characterization module is used to acquire data on the pollutant exposure scenario to be evaluated corresponding to the target cells; wherein, the pollutant exposure scenario data to be evaluated includes characterization of the pollutant mixture to be evaluated, characterization of the exposure conditions to be evaluated, and characterization of the baseline cells to be evaluated; The prediction module is used to input the characterization of the pollutant mixture to be evaluated, the characterization of the exposure conditions to be evaluated, and the characterization of the baseline cells to be evaluated into the environmental virtual cell model to obtain the first cell state of the target cells; A generation module is used to determine the first biological effect caused by the pollutant mixture to be evaluated based on the first cell state, and to generate a risk assessment result of the pollutant mixture to be evaluated based on the first biological effect.

9. An electronic device, characterized in that, include: A memory and a processor are communicatively connected, the memory storing computer instructions, and the processor executing the computer instructions to perform the mixture risk assessment method based on any one of environmental virtual cells as described in claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to perform the mixture risk assessment method based on environmental virtual cells as described in any one of claims 1 to 7.

Images