A method and system for coordinated control of multiple parameters of an environment for rearing immunodeficient animals

Abstract

This invention relates to the field of laboratory animal environmental control technology, and particularly to a multi-parameter environmental collaborative control method and system for raising immunodeficient animals. A multi-parameter environmental collaborative control method for raising immunodeficient animals includes the following steps: S1: Real-time acquisition of preliminary multi-source data during the rearing process of immunodeficient animals, wherein the preliminary multi-source data includes animal gut microbiome data, opportunistic pathogen monitoring data, and environmental microbiome data; S2: Constructing a microbial warfare map based on the preliminary multi-source data, assessing the opportunistic infection risk based on the preliminary multi-source data, tracing the environmental transmission path based on the preliminary multi-source data, and generating a comprehensive risk assessment report. This invention constructs a microbial warfare map and calculates a deviation index by real-time acquisition of gut microbiome data, opportunistic pathogen monitoring data, and environmental microbiome data.

Description

Technical Field

[0001] This invention relates to the field of laboratory animal environmental control technology, and in particular to a multi-parameter environmental collaborative control method and system for raising immunodeficient animals. Background Technology

[0002] Immunodeficient animals, lacking T cells, B cells, and / or NK cells, are extremely sensitive to various opportunistic pathogens in the environment, and their rearing typically relies on strict barrier facilities. Current technologies primarily focus on physical environmental control, microbial cleanliness control, and standardized personnel procedures to build protective systems, aiming to achieve SPF-level or higher barrier standards. However, this control model has revealed several limitations in long-term practice: the control logic aims to maintain predetermined physical parameters, failing to incorporate the dynamic changes in the animal's gut microbiome into the feedback loop. The root cause of opportunistic infections in immunodeficient animals often lies in microecological imbalance rather than a single environmental indicator exceeding the standard. Furthermore, temperature, humidity, ventilation, and disinfection parameters operate independently, lacking a quantitative correlation between environmental changes, microecological response, and animal health, making proactive regulation based on microecological status impossible. In addition, facing the stress of long-distance transportation, humanized cell transplantation, and special experimental scenarios, existing static control schemes struggle to personalize adjustments based on actual microecological fluctuations. Moreover, when infection outbreaks occur, there is a lack of precise means to distinguish the source of contamination and the transmission route, often resulting in large-scale disinfection or batch culling, leading to wasted experimental costs and resources.

[0003] Therefore, there is an urgent need to develop a multi-parameter environmental collaborative control method and system for feeding immunodeficient animals, in order to overcome the bottlenecks of existing technologies such as the separation of physical environment and microbial state, isolated control parameters, poor scene adaptability, and lack of infection tracing capabilities. Summary of the Invention

[0004] To overcome the shortcomings of existing technologies, such as the disconnect between environmental control parameters and microbiome state, the lack of dynamic synergistic regulation based on microecological feedback, and insufficient adaptive adjustment and precise tracing of infection transmission pathways in special scenarios, this invention provides a multi-parameter environmental synergistic control method and system for raising immunodeficient animals.

[0005] The technical implementation scheme of the present invention is: a multi-parameter environmental synergistic control method for feeding immunodeficient animals, comprising the following steps: S1: Real-time collection of preliminary multi-source data during the feeding of immunodeficient animals, including animal gut microbiome data, opportunistic pathogen monitoring data, and environmental microbiome data; S2: Construct a microbial warfare map based on the preliminary multi-source data, conduct an opportunistic risk infection assessment on the preliminary multi-source data, trace the environmental transmission path based on the preliminary multi-source data, and generate a comprehensive risk assessment report; S3: Based on the comprehensive risk assessment report, perform multi-parameter collaborative decision-making to obtain the optimal environmental control instruction set; S4: Coordinate and control the execution equipment according to the optimal environment control instruction set, and record complete execution process data; S5: Form closed-loop feedback and self-optimization based on the complete execution process data.

[0006] Preferably, the preliminary multi-source data collected in real time during the rearing of immunodeficient animals includes animal gut microbiome data, opportunistic pathogen monitoring data, and environmental microbiome data. Specifically, the animal gut microbiome data includes species composition and relative abundance data of fecal samples; the opportunistic pathogen monitoring data includes quantitative data on nucleic acid copy numbers and single nucleotide polymorphism (SNP) sites for each pathogen; and the environmental microbiome data includes species composition and relative abundance data of environmental samples. Simultaneously, individual animal information, real-time monitoring data of environmental physical parameters, and rearing operation logs are collected as supplementary data.

[0007] Preferably, the step of constructing a microbial warfare map based on the preliminary multi-source data, assessing the opportunistic risk of infection based on the preliminary multi-source data, tracing environmental transmission pathways based on the preliminary multi-source data, and generating a comprehensive risk assessment report includes: the comprehensive risk assessment report containing a deviation index, a step-by-step intervention plan, a risk index, a risk transmission chain, and a microbial transmission network; and inferring inter-species interactions based on the species composition and relative abundance data of the fecal samples to construct a microbial warfare map. ,in For a set of nodes, For the directed edge set of war; extract the species composition and relative abundance data of the fecal samples to obtain steady-state fingerprints. fingerprint of the current bacterial community status Based on the steady-state fingerprint and the current microbial community state fingerprint, a deviation index D is obtained; if the deviation index exceeds a preset deviation threshold, it is determined to be a state of microbial imbalance, and a set of key imbalanced microbial communities is obtained using a backpropagation algorithm based on the deviation index; in the microbial war map, the shortest adjustment path is searched with the nodes in the set of key imbalanced microbial communities as the starting node and the region where the microbial community structure corresponding to the steady-state fingerprint is located as the ending node region; a step-by-step intervention plan is generated based on the shortest adjustment path.

[0008] Preferably, the opportunistic risk assessment of the preliminary multi-source data includes: obtaining an α diversity index based on the species composition and relative abundance data of the fecal samples, and extracting the short-chain fatty acid content; using the α diversity index, short-chain fatty acid content, quantitative data on the nucleic acid copy number of each pathogen, and auxiliary data as variable nodes, determining the directed dependencies between the variable nodes, and constructing a Bayesian belief network. ,in For a set of variable nodes, Let H be the set of directed edges of the belief network; define a risk index H based on the joint probability distribution of the Bayesian belief network. If the risk index exceeds a preset risk threshold, it is determined to be a high-risk state. When it is determined to be a high-risk state, extract each directed path of the Bayesian belief network and calculate its corresponding risk propagation probability. If the risk propagation probability exceeds a preset risk propagation threshold, the corresponding directed path is taken as a risk propagation chain.

[0009] Preferably, the environmental transmission path tracing based on the preliminary multi-source data includes: constructing a three-dimensional spatial coordinate system model of the breeding area facility, mapping the three-dimensional spatial coordinates of each environmental sampling point with the species composition and relative abundance data of the environmental samples to form a three-dimensional pollution map; simulating the potential diffusion path of microorganisms within the facility based on the three-dimensional pollution map, and when a high-risk pathogen is detected at any environmental sampling point, searching for pathogen transmission paths and candidate pollution source locations with that environmental sampling point as the endpoint and other environmental sampling points within the facility as the starting point; and performing strain homology detection based on the single nucleotide polymorphism site information, using all environmental sampling points and animal individual sampling points that are detected as homologous strains as nodes, and the association relationship between the homologous strains at the sampling points as directed edges, the direction of which is determined by the sampling time sequence, to construct a microbial transmission network.

[0010] Preferably, the step of performing multi-parameter collaborative decision-making based on the comprehensive risk assessment report to obtain the optimal environmental control instruction set includes: applying constraints based on the step-by-step intervention plan, risk transmission chain, and microbial transmission network; and constructing a multi-objective optimization function based on the deviation index D and the risk index H. Where J represents the total cost; C represents the total operating cost; The initial optimization weight coefficients are dynamically adjusted according to the animal's physiological stage; a global candidate decision variable set u is defined; and the multi-objective optimization function is solved based on the global candidate decision variable set to obtain the optimal global decision vector. and local environmental control correction amount The optimal environmental control instruction set is obtained based on the optimal global decision vector and the local environmental control correction amount.

[0011] Preferably, the multi-parameter collaborative decision-making based on the comprehensive risk assessment report includes: the multi-parameter collaborative decision-making further includes special scenario adaptive decision-making; the special scenarios include transportation arrival events, experimental operation events, and cell transplantation events recorded in the feeding operation log; the special scenario adaptive decision-making is as follows: when a transportation arrival event is recorded in the feeding operation log, a microenvironment recovery mode is activated; when an experimental operation event is recorded in the feeding operation log, a rapid recovery mode is activated; and when a cell transplantation event is recorded in the feeding operation log, a humanized microecological support mode is activated.

[0012] Preferably, the step of coordinating and controlling the execution devices according to the optimal environmental control instruction set and recording complete execution process data includes: coordinating and controlling the temperature and humidity regulation device, ventilation system, disinfection device, automatic drinking system, automatic feeding system, and lighting system to work together according to the optimal environmental control instruction set; continuously collecting the operating status parameters of each execution device and the actual physical parameters of the animal's environment during execution; automatically pausing the execution of the device and issuing an alarm signal when the operating status parameters of any device exceed the preset safe operating range; automatically triggering correction instructions to readjust the corresponding execution device when the deviation between the actual physical parameters and the target value exceeds the preset execution deviation threshold; and confirming the completion of this round of execution and recording the complete execution process data of this round after all execution devices have reached the target state and operated stably for more than the preset stable time threshold.

[0013] Preferably, the step of forming closed-loop feedback and self-optimization based on the complete execution process data includes: re-collecting preliminary multi-source data within a preset feedback monitoring time window after execution, comparing the re-collected data with the complete execution process data item by item to obtain the difference values ​​of each indicator; performing a comprehensive evaluation of the control effect based on the difference values, performing parameter self-optimization based on the results of the comprehensive evaluation of the control effect, and recording the closed-loop feedback data of this round; and performing long-term trend analysis based on the continuously accumulated closed-loop feedback data.

[0014] Preferably, the multi-parameter environmental collaborative control system for raising immunodeficient animals further includes: The multi-source data acquisition module collects preliminary multi-source data in real time during the feeding of immunodeficient animals. The preliminary multi-source data includes animal gut microbiome data, opportunistic pathogen monitoring data, and environmental microbiome data. The data processing and analysis module constructs a microbial warfare map based on the preliminary multi-source data, conducts an opportunistic risk infection assessment on the preliminary multi-source data, traces the environmental transmission path based on the preliminary multi-source data, and generates a comprehensive risk assessment report. The multi-parameter collaborative decision-making module performs multi-parameter collaborative decision-making based on the comprehensive risk assessment report to obtain the optimal environmental control instruction set. The collaborative execution module coordinates and controls the execution device according to the optimal environment control instruction set and records complete execution process data; The closed-loop feedback self-optimization module forms closed-loop feedback and self-optimization based on the complete execution process data.

[0015] This invention has the following advantages: By collecting intestinal microbiome data, opportunistic pathogen monitoring data, and environmental microbiome data in real time, this invention constructs a microbial warfare map and calculates the deviation index, achieving a leap from maintaining physical parameters to actively maintaining microecological balance, enabling the control target to directly address the root cause of opportunistic infections; by constructing a quantitative correlation model between environmental parameters, microbial indicators, and animal health status through a Bayesian belief network, it achieves multi-parameter collaborative decision-making; it possesses adaptive decision-making capabilities for special scenarios such as transportation arrival, experimental operations, and cell transplantation, and can automatically trigger corresponding control modes based on feeding operation logs, achieving personalized and dynamic environmental regulation; through three-dimensional pollution map modeling, strain homology detection, and microbial transmission network construction, it achieves precise tracing of pathogen transmission paths and location of risk nodes, providing a basis for targeted eradication and precise intervention; through a closed-loop feedback self-optimization mechanism, it continuously performs long-term trend analysis and parameter self-optimization based on historical execution data, enabling the control system to have self-evolution capabilities. Attached Figure Description

[0016] Figure 1 This is a flowchart of a multi-parameter environmental collaborative control method for feeding immunodeficient animals.

[0017] Figure 2 This is a schematic diagram of a multi-parameter environmental collaborative control system for raising immunodeficient animals. Detailed Implementation

[0018] The present invention will now be described in detail with reference to the accompanying drawings.

[0019] Example 1: A multi-parameter environmental synergistic control method for feeding immunodeficient animals, such as... Figure 1 As shown, it includes the following steps: S1: Real-time collection of preliminary multi-source data during the feeding of immunodeficient animals, including animal gut microbiome data, opportunistic pathogen monitoring data, and environmental microbiome data; The animal gut microbiome data includes species composition and relative abundance data of fecal samples; the opportunistic pathogen monitoring data includes quantitative data on nucleic acid copy number and single nucleotide polymorphism site information for each pathogen; the environmental microbiome data includes species composition and relative abundance data of environmental samples; and animal individual information, real-time monitoring data of environmental physical parameters, and feeding operation logs are collected simultaneously as auxiliary data.

[0020] It should also be noted that the species composition and relative abundance data of the fecal samples were obtained by collecting fresh fecal samples from immunodeficient animals, extracting total DNA, and performing 16S rRNA gene sequencing. The sequencing results were compared with a microbial reference database and species classification annotation was performed. The sequence proportion of each species was then calculated as the relative abundance data. The microbial reference database integrates standard reference genomes and marker gene sequences of bacteria, fungi, and viruses from international public nucleic acid sequence databases. On this basis, it is supplemented with lineage-specific sequences obtained by whole-genome resequencing of local epidemic opportunistic pathogen isolates isolated and preserved in the target breeding facility over the years. After sequence redundancy removal, phylogenetic tree consistency verification, and functional annotation quality control, a comprehensive reference sequence library suitable for microbial classification, identification, SNP typing, and homology analysis in immunodeficient animal facilities was constructed. The quantification of nucleic acid copy numbers for each pathogen was achieved by extracting nucleic acids from animal feces or related tissue samples, designing specific primers and probes targeting conserved genes of specific opportunistic pathogens, and performing absolute quantification using real-time quantitative PCR. The single nucleotide polymorphism (SNP) site information was obtained by performing whole-genome sequencing on the detected opportunistic pathogens, aligning the obtained sequences to the pathogen's reference genome, and utilizing variation... The detection algorithm identifies single base variation sites that differ from the reference sequence, thereby obtaining SNP site information. The species composition and relative abundance data of the environmental samples are obtained by collecting microbial samples from the rearing environment through sterile swab wiping of cages, walls, or air sedimentation. Environmental DNA is extracted and subjected to metagenomic sequencing or 16S rRNA sequencing. The sequence abundance percentage of each microbial species is obtained through comparison with a microbial reference database and taxonomic analysis. Animal individual information is obtained simultaneously by integrating electronic ear tag RFID radio frequency identification records from the husbandry management system, cage card entry information, and quality inspection reports provided by animal suppliers. The basic data include lineage, age, sex, weight, and immunodeficiency genotype; real-time monitoring data of environmental physical parameters are continuously collected and uploaded through IoT monitoring devices such as temperature and humidity sensors, differential pressure gauges, anemometers, and particle counters deployed in the rearing isolators or barrier facilities, including temperature, relative humidity, differential pressure, air exchange rate, and air cleanliness; the rearing operation log is automatically generated by rearing personnel manually entering or scanning QR codes when operating management software on mobile terminals or PCs, recording the time, content, and operator of each entry / exit from the barrier, cage replacement, addition of feed and water, drug treatment, and experimental operations, and automatically generating a structured text log with timestamps.

[0021] S2: Construct a microbial warfare map based on the preliminary multi-source data, conduct an opportunistic risk infection assessment on the preliminary multi-source data, trace the environmental transmission path based on the preliminary multi-source data, and generate a comprehensive risk assessment report; The comprehensive risk assessment report includes deviation index, step-by-step intervention plan, risk index, risk transmission chain, and microbial transmission network; based on the species composition and relative abundance data of the fecal samples, inter-species interactions are inferred, and a microbial warfare map is constructed. ),in For a set of nodes, For the directed edge set of war; extract the species composition and relative abundance data of the fecal samples to obtain steady-state fingerprints. fingerprint of the current bacterial community status Based on the steady-state fingerprint and the current microbial community state fingerprint, a deviation index D is obtained; if the deviation index exceeds a preset deviation threshold, it is determined to be a state of microbial imbalance, and a set of key imbalanced microbial communities is obtained using a backpropagation algorithm based on the deviation index; in the microbial war map, the shortest adjustment path is searched with the nodes in the set of key imbalanced microbial communities as the starting node and the region where the microbial community structure corresponding to the steady-state fingerprint is located as the ending node region; a step-by-step intervention plan is generated based on the shortest adjustment path.

[0022] Based on the species composition and relative abundance data of the fecal samples, the α diversity index was obtained, and the short-chain fatty acid content was extracted. Using the α diversity index, short-chain fatty acid content, quantitative data on the nucleic acid copy number of each pathogen, and auxiliary data as variable nodes, the directed dependencies between the variable nodes were determined, and a Bayesian belief network was constructed. ,in For a set of variable nodes, Let H be the set of directed edges of the belief network; define a risk index H based on the joint probability distribution of the Bayesian belief network. If the risk index exceeds a preset risk threshold, it is determined to be a high-risk state. When it is determined to be a high-risk state, extract each directed path of the Bayesian belief network and calculate its corresponding risk propagation probability. If the risk propagation probability exceeds a preset risk propagation threshold, the corresponding directed path is taken as a risk propagation chain.

[0023] A three-dimensional spatial coordinate system model of the breeding area facility is constructed, mapping the three-dimensional spatial coordinates of each environmental sampling point with the species composition and relative abundance data of the environmental samples to form a three-dimensional pollution map. Based on the three-dimensional pollution map, the potential diffusion paths of microorganisms within the facility are simulated. When a high-risk pathogen is detected at any environmental sampling point, the pathogen transmission path and candidate pollution source locations are searched, with that environmental sampling point as the endpoint and other environmental sampling points within the facility as the starting points. Based on the single nucleotide polymorphism site information, strain homology detection is performed. All environmental sampling points and individual animal sampling points that are detected as homologous strains are used as nodes, and the association relationship between the homologous strains at the sampling points is used as directed edges. The direction of the directed edges is determined by the sampling time sequence to construct a microbial transmission network.

[0024] It should also be noted that the process of inferring interspecies interactions based on the species composition and relative abundance data of the fecal samples is as follows: Based on the obtained species composition and relative abundance fluctuation data of the fecal samples over time, a sparsely regularized regression model is constructed using the generalized Lotka-Volterra kinetic equation. By solving the partial correlation and linear interaction coefficient matrix between the abundance changes of each species and the abundance of other species, the sign and significance level of the regression coefficients are used to determine whether there is a promoting, inhibiting, or neutral relationship between species. The quantified directed interaction coefficients are then assigned to the weights and directions of the directed edges between nodes in the microbial warfare map. The microbial warfare map is constructed by using all the microbial species detected in the fecal samples as nodes and the promoting, inhibiting, or neutral relationships between species inferred based on the generalized Lotka-Volterra kinetic model as directed weighted edges. A topological network reflecting the competition and synergy dynamics within the gut microbiota was constructed. The node set is the set of all bacterial, fungal, and opportunistic pathogen species identified by 16S rRNA sequencing in fecal samples. The war-oriented edge set is the set of all directed interaction relationships quantified by the interaction inference method, where each directed edge points from the acting species node to the acted species node. Based on the inter-species interaction coefficients estimated by the generalized Lotka-Volterra dynamic regression model, non-zero regression coefficients with p-values ​​less than 0.05 were normalized by absolute value min-max, while retaining their positive or negative signs to characterize promoting / inhibiting effects. The normalized values ​​were multiplied by a significance correction factor to obtain the corresponding edge weight coefficients. The significance correction factor is obtained by hypothesis testing of the p-values ​​of the interaction coefficients, calculating their negative logarithms and truncating the upper limit, and used as the significance correction factor. The steady-state fingerprint... To construct a standardized baseline abundance vector by performing species-by-species arithmetic averaging of the relative abundance data of fecal samples collected multiple times from immunodeficient animals of the same strain within the facility during a stable period free from infection and with normal physiological indicators, the initial value was taken as the mean of the relative abundance of species from fecal samples of at least three immunodeficient animals of the same strain collected initially as the initial steady-state fingerprint; the current microbial community state fingerprint... To sum and normalize the relative abundance data of each species obtained from sequencing of the fecal sample collected at the current moment, directly using it as the current microbial community abundance distribution vector; the deviation index D is used to quantify the overall deviation of the current intestinal microecological structure from the healthy homeostasis level, through... The preset deviation threshold is obtained by collecting historical deviation index data from multiple measurements of immunodeficient animals of the same strain within the facility during a clearly defined period of health stability, and taking the 95th percentile as the threshold; the key dysregulated microbial community set is calculated by gradient backpropagation of the current microbial community abundance vector based on the deviation index, and the species are accumulated in descending order of their contribution to the deviation index, and species with a cumulative contribution of more than 80% are listed as key dysregulated microbial communities; the search for the shortest regulation path uses the reciprocal of the directed edge weight in the microbial warfare map as the movement cost, and adopts Dijkstra's algorithm to find the key dysregulated microbial community set. Starting from the swarm node, the shortest node sequence path to the region where the microbial community corresponding to the steady-state fingerprint is located is searched. The relative abundance in the steady-state fingerprint is accumulated from high to low, and the core species nodes with the first cumulative proportion exceeding 90% are selected. The set formed by these nodes is defined as the termination node region of the steady-state microbial community structure. The step-by-step intervention scheme is based on the order of the nodes in the shortest regulation path and the inhibition / promotion requirements of each node. The path is decomposed into a phased intervention operation sequence that combines the targeted elimination of key dysregulated bacteria by narrow-spectrum antibacterial agents / specific bacteriophages with the directional amplification of beneficial bacteria by prebiotics / soluble dietary fiber.

[0025] Based on the species composition and relative abundance data of fecal samples, the corresponding Shannon diversity index, Simpson diversity index, and Chao1 richness estimates were obtained, and the average of the three was taken as the α diversity index of the sample. Fresh fecal samples from immunodeficient animals were centrifuged and filtered, and the content of short-chain fatty acids (SCFAs), including acetic acid, propionic acid, and butyric acid, was detected using gas chromatography-mass spectrometry (GC-MS) to determine the SCFA content. The process of determining the directed dependencies between variable nodes involved initially establishing a candidate directed edge skeleton based on expert prior knowledge, then optimizing the candidate skeleton through a PC algorithm, and finally determining the final directed dependencies between variable nodes based on the network structure with the highest BDeu score. The α diversity index, SCFA content, quantitative data on nucleic acid copy numbers of each pathogen, and auxiliary data were defined as nodes, and the determined directed dependencies between variables were set as directed edges. A conditional probability table for each node under the conditions of its parent node's value was obtained through maximum likelihood estimation, thereby constructing a complete Bayesian belief network. For each node in the Bayesian belief network... The conditional probability distribution of a node given its parent node state is obtained by counting the frequency of each value of the node in the training data given all possible combinations of its parent node values, and then dividing that frequency by the total number of samples in that parent node combination. The conditional probability distributions of all nodes constitute the joint probability distribution of the Bayesian belief network. The risk index H is used to quantify the current overall infection risk level of the system, through... The preset risk threshold is obtained based on the risk index distribution preceding a clearly occurring opportunistic infection event in the historical data of the facility, with the 90th percentile of this distribution taken as the preset risk threshold; for each directed path The corresponding risk transmission probability pass Before calculation, a first-order Markov property determination is performed on each directed path. This verifies that the risk state of the successor node on the path is independent of the conditions of earlier historical nodes, given the direct predecessor node. If the determination passes, the nodes are multiplied together using chain multiplication. If the determination fails, the fully joint probability distribution of the Bayesian belief network must be used instead. The preset risk propagation threshold is obtained by statistically analyzing the risk propagation probability of confirmed pathogen propagation events in the historical records of the facility and taking the lower quartile of these probability values ​​as the threshold.

[0026] A three-dimensional spatial coordinate system for the rearing area was constructed, and the three-dimensional coordinates of all environmental sampling points were labeled. Then, the species composition and relative abundance data obtained from sequencing at each sampling point were mapped to the corresponding coordinate points as attribute values. Finally, the abundance in unsampled areas was estimated using the Kriging interpolation algorithm, generating a three-dimensional continuous pollution distribution map covering the entire facility, i.e., a three-dimensional pollution map. Based on the three-dimensional pollution map and the facility ventilation layout and personnel movement model, a path simulation algorithm based on aerodynamic diffusion and contact propagation was used, driven by pollution concentration gradient, to obtain the temporal propagation trajectory of microbial aerosols and surface sediments in space, generating a potential diffusion path network from high-concentration areas to low-concentration areas. The network and directional arrows; the high-risk pathogens are environmental sampling points that meet the criteria of having a nucleic acid copy number exceeding the historical safety reference upper limit for that pathogen in the same strain of immunodeficient animals, and having a risk transmission probability exceeding a preset risk transmission threshold in the Bayesian belief network, while also being located as the source or key intermediate node of at least one risk transmission chain in the three-dimensional contamination map and microbial transmission network; the historical safety reference upper limit is calculated by collecting annual nucleic acid copy number quantitative detection data of healthy immunodeficient animals of the same strain in the facility during periods when no infection occurred and physiological indicators were stable, calculating its mean and standard deviation, and taking the mean plus 3 times the standard deviation as the historical safety reference upper limit; the pathogens detected at each sampling point Sequencing data were aligned to the same reference genome, and high-confidence core genome single nucleotide polymorphism (SNP) sites were extracted. The number of SNP differences between each pair of samples was calculated. If the number of differences was less than a preset threshold, the two samples were determined to be homologous strains. The preset threshold was based on the distribution of core genome SNP differences between historically confirmed homologous strain pairs within the facility, with the 95th percentile as the preset threshold, and an initial value of 5. Using the environmental sampling point where high-risk pathogens were detected as the endpoint, reverse aerodynamic diffusion trajectory simulation was performed by combining data on facility airflow direction, ventilation pressure difference distribution, and personnel operation flow to track multiple possible routes of pathogen transmission from the potential origin to the endpoint. The path is analyzed, and cross-validation is performed based on the detection time sequence of homologous strains at each associated sampling point and the decay gradient law of nucleic acid abundance along the path. The sampling point that meets the condition of the earliest time sequence and is located upstream of the reverse trajectory is selected as the candidate location of the source of pollution, and the diffusion node sequence of each path is output simultaneously as the pathogen transmission path. The environmental sampling points and animal individual sampling points that detect homologous strains are used as network nodes. If there are homologous strains between two nodes and the detection time of one sampling point is earlier than that of the other, a directed edge is established from the earlier detected node to the later detected node. By combining all nodes and directed edges, a microbial transmission network reflecting the direction of strain transmission in the spatial and temporal dimensions is formed.

[0027] S3: Based on the comprehensive risk assessment report, perform multi-parameter collaborative decision-making to obtain the optimal environmental control instruction set; Constraints are applied based on the step-by-step intervention plan, risk transmission chain, and microbial transmission network; a multi-objective optimization function is constructed based on the deviation index D and the risk index H. Where J is the total cost; The maximum observed deviation index; C represents the maximum risk index; C represents the overall operating cost. The initial optimization weight coefficients are dynamically adjusted according to the animal's physiological stage; a global candidate decision variable set u is defined; and the multi-objective optimization function is solved based on the global candidate decision variable set to obtain the optimal global decision vector. and local environmental control correction amount The optimal environmental control instruction set is obtained based on the optimal global decision vector and the local environmental control correction amount.

[0028] The multi-parameter collaborative decision-making also includes special scenario adaptive decision-making; the special scenarios include transportation arrival events, experimental operation events, and cell transplantation events recorded in the feeding operation log; the special scenario adaptive decision-making is as follows: when a transportation arrival event is recorded in the feeding operation log, the microenvironment recovery mode is activated; when an experimental operation event is recorded in the feeding operation log, the rapid recovery mode is activated; when a cell transplantation event is recorded in the feeding operation log, the humanized microecological support mode is activated.

[0029] It should also be noted that the constraint processing includes step-by-step intervention scheme constraints, risk propagation chain constraints, and microbial transmission network constraints. The step-by-step intervention scheme constraint stipulates that the execution order and intensity of environmental control commands during multi-objective optimization must strictly follow the phased operation sequence decomposed from the shortest adjustment path, and cannot be executed ahead of schedule or in reverse order. The risk propagation chain constraint requires that multi-objective optimization decisions must prioritize the variable nodes traversed by each risk propagation chain in the Bayesian belief network, forcibly reducing the probability of propagation along the chain to below a preset threshold; otherwise, the decision scheme is considered invalid. The microbial transmission network constraint stipulates that environmental disinfection and ventilation control measures must prioritize high-risk pathogen nodes and their upstream source nodes identified in the microbial transmission network. If the homologous strain association edge of any high-risk pathogen node is not severed, the enhanced disinfection command for that node is prohibited from being cancelled. The maximum observed deviation index is collected from the historical data of immunodeficient animals of the same strain within the facility. The deviation index of all valid monitoring moments in the cycle is used, and the maximum value is taken as the maximum observed deviation index. If there is no historical data, the square root of the total number of stable detectable species in the gut microbiome of this strain is used as the initial estimate. The maximum risk index is collected from the risk indices of all valid monitoring moments in the historical operating cycle of the facility, and the maximum value is taken as the maximum risk index. If there is no historical data, the sum of the marginal entropies of all variable nodes in the Bayesian belief network under their uniform state distribution is used as the initial estimate. The comprehensive operating cost is calculated by converting the electricity, consumables, equipment depreciation, and manpower costs consumed by temperature and humidity control, ventilation, disinfection, automatic drinking / feeding, and lighting systems within the preset accounting cycle into monetary values ​​at market prices and summing them up. Then, it is normalized to min-max based on the maximum allowable operating cost under the historical or design conditions of the facility to obtain the dimensionless comprehensive operating cost. The initial optimization weight coefficient is dynamically adjusted according to the physiological stage of the animal: during the adaptation period, it is set to... Set up during the steady-state feeding period During the experimental treatment period, set A global candidate decision variable set u is defined by aggregating factors such as temperature setpoint, relative humidity setpoint, air exchange rate, light cycle, UV disinfection duration, probiotic / prebiotic addition scheme, and feed type adjustments. The optimal global decision vector is obtained by solving the multi-objective optimization function based on this global candidate decision variable set. and local environmental control correction amount The process is as follows: Based on the global candidate decision variable set, a two-layer optimization method combining genetic algorithm and reinforcement learning is used to solve the multi-objective optimization function. The genetic algorithm layer performs a global search with a population of 50-200 individuals, employing tournament selection, crossover, and mutation operations. After 50-200 generations of evolution, it outputs the globally optimal candidate decision vector. During the initialization of the genetic algorithm, the parameter combination corresponding to the step-by-step intervention scheme is used as the high-probability seed individuals in the first-generation population. These high-probability seed individuals account for 5%-15% of the total number of individuals in the first-generation population. The reinforcement learning layer uses the globally optimal candidate decision vector as the initial state and employs a deep deterministic policy gradient algorithm for local fine-tuning. Under the premise of satisfying the constraints set by the constraint processing, it outputs the optimal global decision vector. Local environmental control corrections for areas where high-risk pathogen nodes are located The optimal global decision vector and the local environmental control correction are merged and parsed into an executable set of environmental control instructions as the optimal environmental control instruction set. The optimal environmental control instruction set includes global temperature adjustment instructions, global humidity adjustment instructions, global ventilation adjustment instructions, light adjustment instructions, disinfection control instructions, nutrient supply instructions, as well as local air exchange rate adjustment instructions and local pressure difference adjustment instructions for the area where high-risk pathogen nodes are located.

[0030] The microenvironment recovery mode automatically adjusts the temperature and humidity linearly to the optimal recovery range for the immunodeficient animal strain within a preset time period, according to a preset rate of change, upon recording the arrival of transport. Simultaneously, it activates low-speed, gentle ventilation and a remotely controlled automatic feeding system to provide easily digestible, high-energy feed. Regular cage changing operations are paused to reduce human stress interference, and the preset recovery duration is maintained until the system determines recovery is complete. The preset rate of change is based on the 25th percentile of the critical rate of change at which the immunodeficient animal strain first exhibited a stress response in historical temperature and humidity step tests; its reciprocal is used as the safety upper limit, and the preset rate of change is set accordingly. The preset time period is adjusted from the abnormal state to the current state based on the facility's historical data. The 90th percentile of the shortest adjustment time for animals in a stable condition with no recorded stress is used as the preset adjustment duration. The preset optimal recovery interval is obtained by summarizing environmental temperature and humidity monitoring data of healthy animals of the same strain in the facility during the stable period, and taking the 95% confidence interval of the mean temperature and the 95% confidence interval of the mean humidity as the optimal recovery interval. The preset recovery duration is based on the median of the distribution of time required for the animal deviation index to recover to below the threshold after historical transportation arrival, plus the standard deviation of the distribution. The rapid recovery mode automatically compensates the temperature and humidity to a mild upward level from the baseline before the operation when an experimental operation event is recorded, and simultaneously activates the independent air supply unit to increase the air exchange rate by a preset ratio to accelerate the removal of the operation zone. The introduced particulate matter gradually recovers to the baseline setting within a set recovery time; the pre-operation baseline mild rise level is based on the statistical upper limit of the acceptable fluctuation range of animal physiological indicators after historical experimental operations, with the temperature rise set at 1 / 3 of this upper limit and the humidity rise set at 1 / 2 of this upper limit as the preset mild rise level; the preset ratio is selected from the measured relationship curve between particulate matter removal efficiency and air exchange rate in the facility, using the minimum air exchange rate multiple at which 90% maximum removal efficiency is achieved as the preset ratio; the set recovery time is based on the actual time distribution of environmental parameters after historical operations, which have been actively adjusted to return to the baseline setting value and in which there is no record of animal stress, and the 75th percentile of this distribution is taken as the preset recovery time; the humanized microecological support module The formula is as follows: upon recording a cell transplantation event, the temperature and humidity are automatically adjusted to the target temperature and humidity range that simulates the human gut microenvironment, and the automatic drinking water system is triggered to switch to warm water with added specific prebiotics. At the same time, the low-intensity circadian rhythm mode of the light system is activated to reduce stress hormone levels, and the system continues to run for a preset support period until the system determines that the microecology is stable. The target temperature and humidity range is obtained by referring to the internationally recognized recommended values ​​of temperature and humidity in the standard for simulating the human gut microenvironment and scaling it down according to 80% of the upper limit of the heat and humidity tolerance of the animal strain. The preset support period is based on the 90th percentile of the time distribution required for the gut microecological diversity index of animals to first reach a stable plateau period after historical humanized cell transplantation.

[0031] S4: Coordinate and control the execution equipment according to the optimal environment control instruction set, and record complete execution process data; The optimal environmental control instruction set coordinates the operation of temperature and humidity control equipment, ventilation system, disinfection equipment, automatic drinking system, automatic feeding system, and lighting system. During execution, the operating status parameters of each device and the actual physical parameters of the animal's environment are continuously collected. When the operating status parameters of any device exceed the preset safe operating range, the execution of that device is automatically suspended and an alarm signal is issued. When the deviation between the actual physical parameters and the target value exceeds the preset execution deviation threshold, a correction instruction is automatically triggered to readjust the corresponding device. When all devices have reached the target state and operated stably for more than the preset stable time threshold, the execution of this round is confirmed to be complete, and the complete execution process data of this round is recorded.

[0032] It should also be noted that the process of coordinating the temperature and humidity control equipment, ventilation system, disinfection equipment, automatic drinking water system, automatic feeding system, and lighting system according to the optimal environmental control instruction set is as follows: The global temperature and humidity control instructions are sent to the temperature and humidity control equipment, controlling it to adjust according to the set temperature and humidity change rate limits. Under normal conditions, the temperature change rate does not exceed 2℃ / hour and the humidity change rate does not exceed 5% / hour; under microenvironment recovery mode, the temperature change rate does not exceed 1℃ / hour and the humidity change rate does not exceed 3% / hour. The global ventilation control instruction is then sent to the ventilation system, controlling its… The air exchange rate is adjusted to the target value. The local air exchange rate adjustment command and the local pressure difference adjustment command are sent to the ventilation subsystem of the high-risk transmission node area, increasing the local air exchange rate in that area to 1.5 to 2.0 times the global air exchange rate setting. The pressure difference between that area and adjacent areas is adjusted to a negative pressure state. The disinfection control command is sent to the disinfection equipment to control its start-up, stop, and operating time. During the operation of the disinfection equipment, the ventilation system supply to that area is automatically suspended through a safety interlock mechanism and automatically resumed after disinfection. The nutrient supply command is sent to the automatic drinking water system and the automatic feeding system to control the drinking water pH value and probiotics. The prebiotic addition type and dosage are specified, and feed type switching is controlled. The feed type switching adopts a gradual mixing switching method, mixing new and old feeds proportionally to gradually switch to the target feed. The light adjustment command is sent to the light system to control it to operate according to the target light cycle and light intensity. The preset safe operating range is determined by combining the nominal safe operating limits provided by each equipment manufacturer, the actual distribution range of operating status parameters of each device during the facility's historical normal operating cycle, and the equipment redundancy protection boundary and animal safety redundancy requirements, after joint review by equipment engineers and veterinary experts. The preset execution deviation threshold is obtained by collecting data within the facility. The absolute deviation data between the actual measured values ​​and target set values ​​of various environmental physical parameters during the historical normal operation cycle are used as the 95th percentile of the deviation sequence of each parameter as the threshold. The preset stabilization time threshold is based on the time distribution of the time taken from equipment startup to the continuous stabilization of all environmental parameters within the target value ± execution deviation threshold range during multiple normal execution processes in the history of the facility. The 90th percentile of this distribution is used as the preset stabilization time threshold. The complete execution process data includes the instruction reception time, start execution time, target attainment time, time series of operating status parameters, time series of actual environmental physical parameters, deviation index, and risk index of each execution device.

[0033] S5: Form closed-loop feedback and self-optimization based on the complete execution process data.

[0034] After execution, preliminary multi-source data is re-collected within a preset feedback monitoring time window. The re-collected data is compared item by item with the complete execution process data to obtain the difference values ​​of each indicator. Based on the difference values, a comprehensive evaluation of the control effect is performed. Based on the results of the comprehensive evaluation of the control effect, the parameters are self-optimized, and the closed-loop feedback data of this round is recorded. Based on the continuously accumulated closed-loop feedback data, a long-term trend analysis is performed.

[0035] It should also be noted that the preset feedback monitoring time window is based on the shortest and longest time required for the deviation index to first show a clear improvement trend in the fecal samples collected after multiple historical executions within the facility. The 50th percentile of this time distribution is taken as the threshold, with an initial value of 24 hours. The difference values ​​of each indicator include the deviation index difference value and the risk index difference value. The deviation index difference value is calculated by substituting the deviation index recalculated from the fecal sample data collected after the execution within the feedback monitoring time window and comparing it with the deviation index recorded at the beginning of the current execution. The result is used as the deviation index difference value to quantify the degree of improvement or deterioration of the microecology before and after regulation. The risk index difference value is calculated by substituting the various variable data collected after the execution within the feedback monitoring time window into the Bayesian belief network and comparing it with the risk index recorded at the beginning of the current execution. The result is used as the risk index difference value to quantify the degree of decrease or increase in infection risk before and after regulation. The process of comprehensively evaluating the control effect based on the difference values ​​is as follows: if the deviation index difference value and the risk index difference value are different... If both the deviation index difference and the risk index difference are less than 0, it is considered an effective intervention. If only the deviation index difference or the risk index difference is less than 0, it is considered a partially effective intervention. If both the deviation index difference and the risk index difference are greater than 0, it is considered an ineffective intervention. If the overall evaluation of the control effect determines it to be an ineffective intervention, the weight coefficients of the deviation index and the risk index in the multi-objective optimization function are increased by a preset large step size. If the overall evaluation of the control effect determines it to be an effective intervention, it remains unchanged. If it is determined to be an effective intervention three times consecutively, the cost coefficient is gradually decreased to reduce redundant operation consumption. If the control effect... If the comprehensive evaluation determines that the intervention is partially effective, the weight coefficients of the deviation index and risk index in the multi-objective optimization function are fine-tuned according to a preset small step size. The preset large step size is based on the average value of the weight coefficient adjustments actually used when the facility was successfully adjusted from an ineffective intervention state to an effective intervention state through manual optimization in its historical operation, with an initial value of 0.1. The preset small step size is based on the average value of the weight coefficient adjustments actually used when the facility was adjusted from a partially effective intervention state to an effective intervention state through manual fine-tuning in its historical operation, with an initial value of 0.02; The initial state data before this round of execution, the complete control commands and equipment operating parameter time series during execution, the difference between the deviation index and risk index after execution, the comprehensive evaluation level of control effect, the new weight coefficients after parameter self-optimization adjustment, and the microbiome re-detection data within the feedback monitoring time window are combined to form the closed-loop feedback data for this round. This data is then encapsulated in a unified data structure and added to the facility's historical closed-loop feedback database. Based on the accumulated multi-round data in the historical closed-loop feedback database, statistical regression and rate of change calculations are performed on the deviation index change trend, risk index change trend, the correlation between the adjustment magnitude of each control parameter and the improvement effect, and the optimal weight coefficient distribution under different animal physiological stages, respectively, along the time dimension, to identify the systematic patterns of continuous decline or seasonal fluctuations in control effect.

[0036] Example 2: Based on Example 1, a multi-parameter environmental collaborative control system for raising immunodeficient animals, such as... Figure 2 As shown, it also includes: The multi-source data acquisition module collects preliminary multi-source data in real time during the feeding of immunodeficient animals. The preliminary multi-source data includes animal gut microbiome data, opportunistic pathogen monitoring data, and environmental microbiome data. The data processing and analysis module constructs a microbial warfare map based on the preliminary multi-source data, conducts an opportunistic risk infection assessment on the preliminary multi-source data, traces the environmental transmission path based on the preliminary multi-source data, and generates a comprehensive risk assessment report. The multi-parameter collaborative decision-making module performs multi-parameter collaborative decision-making based on the comprehensive risk assessment report to obtain the optimal environmental control instruction set. The collaborative execution module coordinates and controls the execution device according to the optimal environment control instruction set and records complete execution process data; The closed-loop feedback self-optimization module forms closed-loop feedback and self-optimization based on the complete execution process data.

[0037] Although the present invention has been described in detail with reference to the above embodiments, it will be apparent to those skilled in the art that various changes or modifications can be made to the invention without departing from the principles and spirit of the invention as defined by the claims. Therefore, the detailed description of the embodiments in this disclosure is for illustrative purposes only and is not intended to limit the invention; rather, the scope of protection is defined by the content of the claims.

Claims

1. A method for multi-parameter environmental synergistic control in the rearing of immunodeficient animals, characterized in that, Includes the following steps: S1: Real-time collection of preliminary multi-source data during the feeding of immunodeficient animals, including animal gut microbiome data, opportunistic pathogen monitoring data, and environmental microbiome data; S2: Construct a microbial warfare map based on the preliminary multi-source data, conduct an opportunistic risk infection assessment on the preliminary multi-source data, trace the environmental transmission path based on the preliminary multi-source data, and generate a comprehensive risk assessment report; S3: Based on the comprehensive risk assessment report, perform multi-parameter collaborative decision-making to obtain the optimal environmental control instruction set; S4: Coordinate and control the execution equipment according to the optimal environment control instruction set, and record complete execution process data; S5: Form closed-loop feedback and self-optimization based on the complete execution process data.

2. The multi-parameter environmental synergistic control method for feeding immunodeficient animals according to claim 1, characterized in that, The preliminary multi-source data collected in real time during the rearing of immunodeficient animals includes animal gut microbiome data, opportunistic pathogen monitoring data, and environmental microbiome data. Specifically, the animal gut microbiome data includes species composition and relative abundance data of fecal samples; the opportunistic pathogen monitoring data includes quantitative data on nucleic acid copy numbers and single nucleotide polymorphism (SNP) sites for each pathogen; and the environmental microbiome data includes species composition and relative abundance data of environmental samples. Simultaneously, individual animal information, real-time monitoring data of environmental physical parameters, and rearing operation logs are collected as supplementary data.

3. The multi-parameter environmental collaborative control method for feeding immunodeficient animals according to claim 1, characterized in that, The process involves constructing a microbial warfare map based on the preliminary multi-source data, assessing the opportunistic risk of infection from the preliminary multi-source data, tracing environmental transmission pathways based on the preliminary multi-source data, and generating a comprehensive risk assessment report. This comprehensive risk assessment report includes: a deviation index, a step-by-step intervention plan, a risk index, a risk transmission chain, and a microbial transmission network; and inferring inter-species interactions based on the species composition and relative abundance data of the fecal samples to construct the microbial warfare map. ,in For a set of nodes, For the directed edge set of war; extract the species composition and relative abundance data of the fecal samples to obtain steady-state fingerprints. fingerprint of the current bacterial community status Based on the steady-state fingerprint and the current microbial community state fingerprint, a deviation index D is obtained; if the deviation index exceeds a preset deviation threshold, it is determined to be a state of microbial imbalance, and a set of key imbalanced microbial communities is obtained using a backpropagation algorithm based on the deviation index; in the microbial war map, the shortest adjustment path is searched with the nodes in the set of key imbalanced microbial communities as the starting node and the region where the microbial community structure corresponding to the steady-state fingerprint is located as the ending node region; a step-by-step intervention plan is generated based on the shortest adjustment path.

4. A multi-parameter environmental synergistic control method for feeding immunodeficient animals according to claim 3, characterized in that, The opportunistic risk assessment of the preliminary multi-source data includes: obtaining the α diversity index based on the species composition and relative abundance data of the fecal samples, and extracting the short-chain fatty acid content; using the α diversity index, short-chain fatty acid content, quantitative data on the nucleic acid copy number of each pathogen, and auxiliary data as variable nodes, determining the directed dependencies between the variable nodes, and constructing a Bayesian belief network. ,in For a set of variable nodes, Let H be the set of directed edges of the belief network; define a risk index H based on the joint probability distribution of the Bayesian belief network. If the risk index exceeds a preset risk threshold, it is determined to be a high-risk state. When it is determined to be a high-risk state, extract each directed path of the Bayesian belief network and calculate its corresponding risk propagation probability. If the risk propagation probability exceeds a preset risk propagation threshold, the corresponding directed path is taken as a risk propagation chain.

5. A multi-parameter environmental synergistic control method for feeding immunodeficient animals according to claim 3, characterized in that, The environmental transmission path tracing based on the preliminary multi-source data includes: constructing a three-dimensional spatial coordinate system model of the breeding area facility, mapping the three-dimensional spatial coordinates of each environmental sampling point with the species composition and relative abundance data of the environmental samples to form a three-dimensional pollution map; simulating the potential diffusion path of microorganisms within the facility based on the three-dimensional pollution map, and when a high-risk pathogen is detected at any environmental sampling point, searching for pathogen transmission paths and candidate pollution source locations with that environmental sampling point as the endpoint and other environmental sampling points within the facility as the starting points; and performing strain homology detection based on the single nucleotide polymorphism site information, using all environmental sampling points and animal individual sampling points that are detected as homologous strains as nodes, and the association relationship between the homologous strains at the sampling points as directed edges, the direction of which is determined by the sampling time sequence, to construct a microbial transmission network.

6. A multi-parameter environmental collaborative control method for feeding immunodeficient animals according to claim 1, characterized in that, The step of making multi-parameter collaborative decisions based on the comprehensive risk assessment report to obtain the optimal environmental control instruction set includes: applying constraints based on the step-by-step intervention plan, risk transmission chain, and microbial transmission network; and constructing a multi-objective optimization function based on the deviation index D and the risk index H. Where J is the total cost; The maximum observed deviation index; C represents the maximum risk index; C represents the overall operating cost. The initial optimization weight coefficients are dynamically adjusted according to the animal's physiological stage; a global candidate decision variable set u is defined; and the multi-objective optimization function is solved based on the global candidate decision variable set to obtain the optimal global decision vector. and local environmental control correction amount The optimal environmental control instruction set is obtained based on the optimal global decision vector and the local environmental control correction amount.

7. A multi-parameter environmental collaborative control method for feeding immunodeficient animals according to claim 6, characterized in that, The multi-parameter collaborative decision-making based on the comprehensive risk assessment report includes: the multi-parameter collaborative decision-making further includes adaptive decision-making for special scenarios; the special scenarios include transportation arrival events, experimental operation events, and cell transplantation events recorded in the feeding operation log; the adaptive decision-making for special scenarios is as follows: when a transportation arrival event is recorded in the feeding operation log, the microenvironment recovery mode is activated; when an experimental operation event is recorded in the feeding operation log, the rapid recovery mode is activated; when a cell transplantation event is recorded in the feeding operation log, the humanized microecological support mode is activated.

8. A multi-parameter environmental collaborative control method for feeding immunodeficient animals according to claim 1, characterized in that, The process of coordinating and controlling the execution devices according to the optimal environmental control instruction set and recording complete execution process data includes: coordinating and controlling the temperature and humidity regulation device, ventilation system, disinfection device, automatic drinking system, automatic feeding system, and lighting system to work together according to the optimal environmental control instruction set; continuously collecting the operating status parameters of each execution device and the actual physical parameters of the animal's environment during execution; automatically pausing the execution of the device and issuing an alarm signal when the operating status parameters of any device exceed the preset safe operating range; automatically triggering correction instructions to readjust the corresponding execution device when the deviation between the actual physical parameters and the target value exceeds the preset execution deviation threshold; and confirming the completion of this round of execution and recording the complete execution process data of this round after all execution devices have reached the target state and operated stably for more than the preset stable time threshold.

9. A multi-parameter environmental collaborative control method for feeding immunodeficient animals according to claim 1, characterized in that, The process of forming closed-loop feedback and self-optimization based on the complete execution process data includes: re-collecting preliminary multi-source data within a preset feedback monitoring time window after execution, comparing the re-collected data with the complete execution process data item by item to obtain the difference values ​​of each indicator; performing a comprehensive evaluation of the control effect based on the difference values, performing parameter self-optimization based on the results of the comprehensive evaluation of the control effect, and recording the closed-loop feedback data of this round; and performing long-term trend analysis based on the continuously accumulated closed-loop feedback data.

10. A multi-parameter environmental collaborative control system for raising immunodeficient animals, used to implement the multi-parameter environmental collaborative control method for raising immunodeficient animals as described in any one of 1-9, characterized in that it further includes... include: The multi-source data acquisition module collects preliminary multi-source data in real time during the feeding of immunodeficient animals. The preliminary multi-source data includes animal gut microbiome data, opportunistic pathogen monitoring data, and environmental microbiome data. The data processing and analysis module constructs a microbial warfare map based on the preliminary multi-source data, conducts an opportunistic risk infection assessment on the preliminary multi-source data, traces the environmental transmission path based on the preliminary multi-source data, and generates a comprehensive risk assessment report. The multi-parameter collaborative decision-making module performs multi-parameter collaborative decision-making based on the comprehensive risk assessment report to obtain the optimal environmental control instruction set. The collaborative execution module coordinates and controls the execution device according to the optimal environment control instruction set and records complete execution process data; The closed-loop feedback self-optimization module forms closed-loop feedback and self-optimization based on the complete execution process data.

Images