Multi-omics-based intelligent domestication system and method for Maotai-flavor liquor-producing saccharomycetes flora

Abstract

The invention belongs to the technical field of intelligent brewing of baijiu, and discloses a multi-omics-based intelligent domestication system and method for a Maotai-flavor baijiu producing yeast group, and the method comprises the following steps: collecting multi-source data, constructing a multi-dimensional data framework, correcting space-time delay of a sensor, and generating an environment-metabolism associated data set; defining a response threshold value of each partition to perform partition environment risk assessment, and constructing a three-dimensional responsibility library; judging the risk state of the responsibility partition by combining the symbiotic efficiency index, and generating a flora binding rule table; dividing conflict interception levels, and generating a three-dimensional topology; four-stage structured gate interception is carried out, and a dynamic domestication instruction set is generated; analyzing flora response deviation by simulating a layered stress scene, and positioning a deviation source in combination with transcription-metabonomics traceability; the efficiency base line and layering strategy container parameters are dynamically updated, the problems of low yeast fermentation efficiency, poor stress resistance, extensive flavor regulation and control and the like in a traditional process are solved, and dynamic optimization of microbial communities and system self-adaptive capacity are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent liquor brewing, and more specifically, to a multi-omics-based intelligent domestication system and method for a yeast flora producing sauce-flavor liquor. Background Art

[0002] The brewing of Maotai-flavor liquor relies on a complex microbial community, particularly the metabolic activity of yeasts, whose performance directly determines the flavor and quality stability of the liquor. Traditionally, yeast screening and optimization have relied on empirical procedures and single-omics analyses (such as metagenomic or metabolomics). This has led to problems such as low fermentation efficiency, poor stress tolerance (such as sensitivity to high temperatures and high acidity), and unstable synthesis of flavor compounds. Furthermore, the dynamic microbial community in open fermentation systems makes it difficult to target core bacterial communities with traditional methods, resulting in frequent abnormal fermentations (such as rancidity and moldy odors) and significant fluctuations in liquor quality.

[0003] To address these issues, existing technologies still face significant shortcomings in yeast screening and process control. First, screening methods (such as WL agar medium) rely on phenotypic observations, resulting in low efficiency and poor specificity, making it difficult to quickly identify high-performing strains. Second, the control of fermentation parameters (such as temperature and pH) lacks a theoretical basis, which can easily lead to metabolic pathway blockage or accumulation of off-flavor substances (such as acetic acid and isovaleric acid). Third, flavor control methods are crude, with insufficient research on the synthesis pathways of key flavor substances (such as esters and alcohols) and systematic modeling of the microbial-metabolism-environmental relationship, making precise intervention difficult. These limitations have severely hampered improvements in the production efficiency and quality stability of Maotai-flavor liquor. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art and achieve the above-mentioned objectives, the present invention provides the following technical solution: a multi-omics-based intelligent domestication system for Maotai-flavor liquor-producing yeast flora, comprising: Environmental-Metabolism Fusion Unit: Deploys a multi-source biosensor network to collect multi-source data on cellar environment and metabolism, accesses historical multi-omics databases to build a multi-dimensional data framework, corrects for sensor spatiotemporal delays, and generates a spatiotemporally aligned environmental-metabolism correlation dataset. Response thresholds are defined for each partition to conduct regional environmental risk assessments and construct a three-dimensional responsibility database. The microbial topology binding unit determines the risk status of the responsibility partition based on the three-dimensional responsibility library and the yeast-lactic acid bacteria symbiotic performance index. It dynamically loads the hierarchical policy container and injects real-time parameters to generate a microbial binding rule table. Based on the microbial binding rule table, it divides the conflict interception level through priority arbitration, generates a three-dimensional topology of time, space, function, and performance, and tabulates it into a policy container scheduling instruction set. Structured gate interception unit: Based on the policy container scheduling instruction set, it performs four-level structured gate interception based on gene compatibility, metabolic pathway saturation rate, attenuation risk value, and dynamic arbitration, thereby generating a dynamic domestication instruction set and simultaneously recording the responsibility transfer path; Multi-omics closed-loop evolution unit: Based on the dynamic domestication instruction set and responsibility transfer path, the microbial response deviation is analyzed by simulating hierarchical stress scenarios, and the root cause of the deviation is located by combining transcriptomics-metabolomics tracing; the performance baseline and hierarchical strategy container parameters are dynamically updated to generate the responsibility system evolution package, and the three-dimensional responsibility library and hierarchical strategy container are reversely updated.

[0005] Furthermore, the generation method of the environment-metabolism association dataset includes: Multi-source data include physical environment data, biological metabolism data and chemical parameter data of the upper, middle and lower layers of the pit; Perform data cleaning and standardization on the collected multi-source data; Access historical multi-omics databases to extract metagenomic functional annotation libraries and metabolomics flavor association profiles; The standardized multi-source data are mapped with the functional gene annotations of the metagenomic functional annotation library and the metabolic pathway information of the metabolomics flavor association profile to form a multi-dimensional data fusion framework; Calculate the data transmission delay time of the lower-level sensor, and then calculate the compensation coefficient based on the delay time, and use the compensation coefficient as the weight of the corresponding sensor; Unify the multi-source data collected by different sensors into standard units, correct the timestamps, and generate spatiotemporally aligned multi-source data; The cellars were divided into different zones. Based on the historical multi-omics database, a correlation model between the dissolved oxygen concentration in each zone and acetic acid accumulation accidents was established, and the response threshold of each zone was defined. Based on the multi-source data aligned in time and space, the environmental risk assessment of each sub-region is obtained; The timestamps, coordinates of each partition, spatiotemporally aligned multi-source data and environmental risk assessment are integrated to generate a spatiotemporally aligned environmental-metabolic association dataset; at the same time, the response thresholds of each partition are integrated to generate an environmental risk threshold table.

[0006] Furthermore, the three-dimensional responsibility library is constructed in the following manner: Based on the spatiotemporally aligned multi-source data and environmental risk threshold table in the environmental-metabolic association dataset, the responsibility type of each partition is defined and labeled; The performance baseline of each partition was set using historical metabolomic flavor association profiles; At the same time, the maximum metabolic load of each partition is annotated by combining the environmental risk assessment in the environment-metabolism association dataset; The spatial coordinates, responsibility type, response threshold, performance baseline and maximum metabolic load of the partition are integrated to construct a three-dimensional responsibility library.

[0007] Furthermore, the generation method of the bacterial community binding rule table includes: According to the three-dimensional responsibility library and based on the yeast-lactic acid bacteria symbiosis model, the symbiotic effectiveness index was obtained; Based on the three-dimensional responsibility database, each zone is treated as a responsibility zone. When the symbiotic performance index does not meet expectations, a high-risk warning is triggered and the corresponding responsibility zone is marked as a high-risk area. Record the responsibility zones and symbiotic performance indexes that trigger alarms, and generate high-risk responsibility zone logs; When the symbiotic effectiveness index reaches the expected level, the responsibility partition is judged to be normal; For the responsibility partitions judged to be normal, according to the responsibility type in the three-dimensional responsibility library, functional strains are matched for each responsibility partition from the functional gene annotations in the multi-dimensional data fusion framework to obtain the corresponding binding strains; At the same time, based on the metabolic pathway information and functional gene annotations in the multi-dimensional data fusion framework, the metabolic targets of each responsibility type are defined; According to the responsibility type and environmental risk assessment, the hierarchical strategy container is dynamically loaded through the preset loading strategy, and the real-time parameters are extracted from the environmental-metabolism association dataset and injected into the hierarchical strategy container; Integrate partition coordinates, responsibility types, binding strains, metabolic targets and hierarchical strategy containers to establish a bacterial community binding rule table.

[0008] Furthermore, the three-dimensional topology is generated in the following manner: Extracting the flavor association profile of the partition from the historical metabolome flavor association profile, and extracting the flavor enhancement potential of the current flavor association profile as the flavor weight of the partition; According to the delay time and response threshold of real-time data transmission of sensors in the partition, the timeliness is obtained as a timeliness evaluation; The flora binding rule corresponding to each partition in the flora binding rule table and the corresponding hierarchical policy container are used as the current policy. Based on the resource usage of the current policy and the historical conflict records, the conflict risk is obtained. Combining flavor weight, timeliness assessment, and conflict risk through priority arbitration to obtain the strategy priority; Based on policy priorities, hierarchical policy containers are divided into different conflict interception levels; Then, partition coordinates, hierarchical strategy container IDs, and metabolic targets are extracted from the bacterial community binding rule table, and conflict interception levels are assigned according to strategy priorities to generate a three-dimensional topology of time, space, function, and effectiveness. The three-dimensional topology is constructed into a table form to obtain a policy container scheduling instruction set.

[0009] Furthermore, the four-level structured gate interception is defined as: comparing the matching degree between the functional genes of the exogenous bacteria and the receptors of the native bacteria through the gene compatibility gate; scanning the bottleneck of target metabolite synthesis through the metabolic pathway saturation rate gate; Integrate gene compatibility gate information and metabolic pathway saturation gate information to generate a gene-metabolism conflict detection report; According to the gene-metabolism conflict detection report, the strain environmental adaptability is calculated through the attenuation risk value gate; the conflict level is determined through the dynamic arbitration gate; Methods for comparing the degree of compatibility between exogenous bacterial functional genes and native bacterial receptors through gene compatibility gates include: Extracting foreign bacterial functional genes and native bacterial receptor proteins from the strategic container scheduling instruction set; Extract sequence similarity between functional genes of foreign bacteria and receptor proteins of indigenous bacteria; Predict the three-dimensional structure of the protein of the foreign bacterial functional gene and the native bacterial receptor protein, and analyze the structural compatibility of the binding sites between the two; Combined with the metabolome flavor association spectrum, the functional adaptability of the exogenous bacterial functional genes and the metabolic requirements of the indigenous bacterial receptor proteins was verified; Combining sequence similarity, structural fitness and functional fitness, the matching evaluation of functional genes was obtained; When the partition matching evaluation does not meet expectations, the gene compatibility gate alarm is activated; Methods for scanning the target metabolite synthesis bottleneck through metabolic pathway saturation rate gates include: Extract target metabolites from the environmental-metabolism association dataset; Then, the target metabolites were detected and the pathway saturation rate was obtained by combining the metabolic flux model; If the pathway saturation rate does not meet expectations, it is determined that a metabolic bottleneck exists in the partition, and the bottleneck type is then identified; based on the bottleneck type, the preset caching strategy is activated.

[0010] Furthermore, the generation method of the dynamic domestication instruction set includes: According to the gene-metabolism conflict detection report, the ways to obtain the strain's environmental adaptability through the attenuation risk value gate include: Obtain temperature fluctuations and pH changes in the partitions, and combine the matching degree to assess the risk of strain attenuation; If the strain attenuation risk does not meet expectations, a high-risk warning for the partition is triggered and a cooling sub-strategy is added; The methods for determining conflict levels through dynamic arbitration gates include: Based on the matching evaluation and metabolic bottleneck information, the conflict level is determined and classified into different conflict levels; Taking the conflict interception level as the core and the conflict level as the policy detail adjustment factor, the conflict interception level and the conflict level are linked in rules to activate the corresponding responsibility transfer strategies of different partitions and simultaneously generate the responsibility transfer path; Based on the results of the four-level gate conflict interception, a dynamic domestication instruction set is generated.

[0011] Furthermore, the method of analyzing microbial response deviations by simulating layered stress scenarios and tracing the root causes of the deviations by combining transcriptomics and metabolomics includes: According to the dynamic domestication instruction set, the environmental parameters of the simulated layered stress scenario are set, and then the intervention effect of the dynamic domestication instruction set on the bacterial metabolic pathway is verified by simulating the layered stress scenario. The deviation of the bacterial response under environmental stress is simulated to obtain the deviation performance; Combining the transcriptome and metabolome dimensions of transcriptomics-metabolomics, we can trace the root causes of microbial response deviations and attribute them to specific responsibility partitions to obtain attribution partitions. Taking the transcriptome and metabolome as analysis dimensions, the deviation performance, attribution partitioning and corresponding correction suggestions of the corresponding analysis dimensions are integrated to generate a deviation root cause analysis report, synchronously update the responsibility transfer path and correct the hierarchical strategy container.

[0012] Furthermore, the generation method of the responsibility system evolution package includes: Based on the modified effect of the layered policy container, the performance baseline of the corresponding partition in the three-dimensional responsibility library is dynamically adjusted to form a new performance baseline; At the same time, based on the deviation root cause analysis report, the layered policy container parameters are optimized and high-conflict responsibility partitions are split; Integrate the split high-conflict responsibility partition coordinates, new performance baselines and hierarchical strategy container parameters to generate a responsibility system evolution package, and then reversely update the three-dimensional responsibility library and hierarchical strategy container.

[0013] Furthermore, the multi-omics-based intelligent domestication method for the yeast flora of Maotai-flavor liquor production includes: S1: Collect multi-source data, correct the temporal and spatial delay of sensors to generate an environmental-metabolic association dataset; define the response threshold of each partition to calculate the partition environmental risk assessment and build a three-dimensional responsibility library; S2: Based on the three-dimensional responsibility library and combined with the symbiotic performance index, the risk status of the responsibility partition is determined, the functional strains are matched and the hierarchical strategy container is dynamically loaded to generate a bacterial community binding rule table, and a three-dimensional topology and strategy container scheduling instruction set; S3: Based on the policy container scheduling instruction set, it executes four-level structured gate interception, generates a dynamic domestication instruction set, and simultaneously records the responsibility transfer path; S4: Based on a dynamic domestication instruction set and responsibility transfer path, the microbial response deviation is analyzed by simulating hierarchical stress scenarios, and the root cause of the deviation is located in combination with transcriptomics-metabolomics; the performance baseline and hierarchical strategy container parameters are dynamically updated to generate a closed-loop iterative responsibility system evolution package.

[0014] The technical effects and advantages of the multi-omics-based intelligent domestication system and method for Maotai-flavor liquor-producing yeast flora of the present invention are as follows: The present invention deploys a multi-source biosensor network to collect multi-source data in real time and build a three-dimensional responsibility database, which significantly improves data integration efficiency and risk assessment accuracy, providing a high-precision decision-making basis for subsequent microbial population optimization. Secondly, the risk status of each partition is determined based on the synergistic performance index. A priority arbitration formula is used to calculate strategy priorities based on flavor weight, timeliness assessment, and conflict risk. This prioritizes strategies and ensures that critical tasks (such as acetate synthesis) are prioritized. This achieves a precise match between bacterial flora and responsible partitions, optimizes conflict management efficiency, and significantly improves the synergy and stability of the fermentation process. Then, through the four-level gate (gene compatibility, metabolic bottleneck, attenuation risk, dynamic arbitration), targeted domestication instructions (such as buffer protocol and cooling strategy) were generated to accurately solve the problem of bacterial conflicts, repair the gap in ethyl acetate synthesis, reduce the risk of odor and ensure the timeliness of the task; Finally, by simulating stress scenarios and transcriptomic-metabolomic tracing, the performance baseline is dynamically updated, strategy parameters are optimized, and high-conflict responsibility units are split, forming a closed-loop iteration and achieving continuous evolution of the system.

[0015] The present invention combines the genome, transcriptome and metabolome with a three-dimensional responsibility library to construct an "environment-metabolism-microbial community" dynamic model, realizing a full-chain analysis of yeast functions. Through the full-process design of "environmental fusion → microbial community binding → gate interception → closed-loop evolution", it solves the problems of low yeast fermentation efficiency, poor stress resistance, and extensive flavor regulation in traditional processes, significantly improves the survival rate, flavor stability, and production efficiency, and realizes the dynamic optimization of microbial communities and the system's adaptability. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 Schematic diagram of the multi-omics-based intelligent domestication system for Maotai-flavor liquor-producing yeast flora of the present invention; Figure 2 This is a schematic diagram of the process of the structured gate interception unit in the multi-omics-based intelligent domestication system of the Maotai-flavor liquor-producing yeast community of the present invention; Figure 3 Schematic diagram of the multi-omics-based intelligent domestication method for sauce-flavor liquor-producing yeast flora of the present invention. DETAILED DESCRIPTION

[0017] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0018] Example 1 See also Figure 1 and 2 As shown, this embodiment is a multi-omics-based intelligent domestication system for Maotai-flavor liquor-producing yeast flora, including: Environmental-Metabolism Fusion Unit: Deploys a multi-source biosensor network to collect multi-source data on cellar environment and metabolism, accesses historical multi-omics databases to build a multi-dimensional data framework, corrects for sensor spatiotemporal delays, and generates a spatiotemporally aligned environmental-metabolism correlation dataset. Response thresholds are defined for each partition to conduct regional environmental risk assessments and construct a three-dimensional responsibility database. The microbial topology binding unit determines the risk status of the responsibility partition based on the three-dimensional responsibility library and the yeast-lactic acid bacteria symbiotic performance index. It dynamically loads the hierarchical policy container and injects real-time parameters to generate a microbial binding rule table. Based on the microbial binding rule table, it divides the conflict interception level through priority arbitration, generates a three-dimensional topology of time, space, function, and performance, and tabulates it into a policy container scheduling instruction set. Structured gate interception unit: Based on the policy container scheduling instruction set, it performs four-level structured gate interception based on gene compatibility, metabolic pathway saturation rate, attenuation risk value, and dynamic arbitration, thereby generating a dynamic domestication instruction set and simultaneously recording the responsibility transfer path; Multi-omics closed-loop evolution unit: Based on the dynamic domestication instruction set and responsibility transfer path, the microbial response deviation is analyzed by simulating hierarchical stress scenarios, and the root cause of the deviation is located by combining transcriptomics-metabolomics tracing; the performance baseline and hierarchical strategy container parameters are dynamically updated to generate the responsibility system evolution package, and the three-dimensional responsibility library and hierarchical strategy container are reversely updated.

[0019] Multi-source data include physical environment data, biological metabolism data and chemical parameter data of the upper, middle and lower layers of the pit; It should be noted that high-precision sensors can be buried at key coordinate points in the cellar to collect real-time data streams (including temperature, dissolved oxygen, pH value, metabolite concentration, etc.); The dynamic monitoring range needs to cover different depths in the upper, middle and lower layers of the pit to ensure the comprehensiveness of multimodal data (such as physical environmental parameters, biological metabolic signals, chemical reaction rates, etc.); Specifically, physical environment data is collected through physical environment sensors, including but not limited to temperature and humidity sensors and dissolved oxygen probes, which cover the key coordinates of the upper, middle and lower layers of the cellar. These sensors are installed at intervals (e.g., a temperature sensor is installed every 30 cm) to regularly collect physical environment data, including but not limited to cellar temperature, humidity and dissolved oxygen. Deploy biometabolism sensors (such as NADH fluorescent probes) in the upper, middle, and lower layers of the pit to monitor biometabolism activity and other biometabolism data, with a focus on areas with active ethanol synthesis (such as the middle layer of the pit); Collect chemical parameter data (such as pH value and concentration of metabolites (such as acetic acid)) of each layer in the pit through chemical parameter sensors (such as pH probes and acetic acid concentration detection modules); Through physical environment sensors, biological metabolism sensors and chemical parameter sensors, a multimodal data acquisition network is formed; Perform data cleaning and standardization on the collected multi-source data; Specifically, data cleaning includes missing value filling (using linear interpolation to supplement missing data), outlier removal (identifying and removing outliers through the Z-score method or IQR method) and noise filtering (using sliding average filtering for biological metabolic data (such as NADH activity)). Data standardization includes unification of spatiotemporal benchmarks (aligning timestamps to UTC time and converting the coordinate system to the WGS84 coordinate system) and unit normalization (converting different sensor outputs to a unified unit (such as temperature → °C, dissolved oxygen → mg / L, pH → pH value)). Access historical multi-omics databases to extract metagenomic functional annotation libraries (such as CRISPR acid-resistant Clostridium acs genes) and metabolomic flavor association profiles (such as ethyl acetate / ethyl lactate ratio); The standardized multi-source data are mapped with the functional gene annotations of the metagenomic functional annotation library and the metabolic pathway information of the metabolomics flavor association profile to form a multi-dimensional data fusion framework, providing data support for subsequent microbiome strategy binding and risk modeling. It should be noted that this multi-dimensional data fusion framework is used to provide data support for subsequent microbial community binding and metabolic association, while ensuring the consistency of global data; The data transmission delay of the lower-level sensor is calculated using the timestamp method (the difference between the reception time and the transmission time is taken as the delay time). A distinction is made between fixed delay (sampling period) and random delay (network jitter) to provide a basis for the compensation mechanism. A compensation coefficient is then calculated based on the delay time (multiplying the delay time (e.g., in seconds) by a preset scaling factor (e.g., 0.15, 0.2) to obtain the compensation coefficient. For example, if the lower-level sensor delay is 4 seconds, the compensation coefficient is 0.6. This compensation coefficient is used as the weight of the corresponding sensor to balance the temporal and spatial deviations. The sensor sampling frequency is also adjusted based on the real-time delay to minimize the delay time (e.g., if the delay is greater than 2 seconds, the sampling time is shortened to 5 seconds). Unify the multi-source data collected by different sensors into standard units, correct the timestamps, and generate spatiotemporally aligned multi-source data; Divide the cellar into zones (i.e., divide each cellar layer into grids of equal length (e.g., 30 cm x 30 cm), with each grid as a zone and assigned a unique ID, such as upper layer - XY01, upper layer - XY02, lower layer - XY-34). Based on a historical multi-omics database, establish a correlation model between dissolved oxygen concentration in each zone within the cellar and acetic acid accumulation accidents, and define response thresholds for each zone (e.g., dissolved oxygen < 18% in the upper layer - XY01 zone triggers a response). Exemplary correlation model: if dissolved oxygen < 5%: the probability of acetic acid exceeding the standard = 82%; if 5% ≤ dissolved oxygen < 10%: the probability of acetic acid exceeding the standard = 35%; if dissolved oxygen ≥ 10%, the probability of acetic acid exceeding the standard = 5%; Based on the spatiotemporally aligned multi-source data, the weights of the same type of sensors in each zone and the corresponding multi-source data are calculated and averaged. Then, the mean of all types of data is weightedly fused with the mean of the weights of the corresponding type of sensors to obtain the environmental risk score of each zone, which is used as the environmental risk assessment of each zone. The weighted fusion method is to multiply the mean of each type of data by the weight mean, and then add the multiplication results of all types; A dynamic early warning mechanism can be defined. When the environmental risk assessment exceeds the preset environmental risk warning threshold, the cellar environment control strategy will be automatically triggered. Integrate timestamps, coordinates of each partition, spatiotemporally aligned multi-source data, and environmental risk assessments to generate a spatiotemporally aligned environmental-metabolic association dataset. Simultaneously, integrate the response thresholds of each partition to generate an environmental risk threshold table. Based on the spatiotemporally aligned multi-source data and environmental risk threshold table in the environmental-metabolism association dataset, the metabolic responsibility type of each partition is defined and labeled, such as oxygenation to maintain fermentation and acid suppression to prevent spoilage. Exemplary: Oxygenation to maintain yeast activity: For areas where dissolved oxygen is below the threshold (such as the lower layer - XY32), oxygenation is used to maintain yeast activity; Acid inhibition and rancidity prevention: Targeting high-risk areas for acetic acid accumulation (such as the lower layer - XY32), we inhibit rancidity by regulating metabolic pathways; The performance baseline of each partition was set using historical metabolomic flavor association profiles (e.g., ethyl acetate / ethyl lactate ratio, acetate conversion rate) (e.g., bottom-XY32 acetate conversion rate baseline = 55%, temperature adaptation coefficient = 0.9 (when the temperature gradient is ≤ 2°C)); Specifically, the performance baseline is set by taking data of the same responsibility type from historical data. For example, for the "acid inhibition and anti-failure" partition, historical records (such as the ethyl acetate / ethyl lactate ratio or acetic acid conversion rate) are extracted, and the average or median of the historical data is taken as the performance limit value; At the same time, combined with the environmental risk assessment in the environment-metabolism association dataset, the maximum metabolic load of each partition was marked (e.g., upper limit of acetic acid conversion rate = 65%); Specifically, maximum metabolic load = historical average × (environmental risk assessment × attenuation coefficient); where the historical average is the average of all historical performance baseline values ​​set in historical data within a certain period of time, and the attenuation coefficient is the preset attenuation ratio, for example, 0.1%; It should be noted that in the dynamic baseline adjustment logic of the maximum metabolic load, environmental risk assessment is used as a dynamic variable to correct the metabolic load upper limit in real time, which can ensure the stability of the system when operating in different areas; Integrate the spatial coordinates, responsibility type, response threshold, performance baseline and maximum metabolic load of the partition to construct a three-dimensional responsibility library with performance baseline; A responsibility matrix map is generated simultaneously to visualize the distribution of responsibility types in each cellar zone (e.g., the lower layer - XY32 is marked as the "acid suppression and spoilage prevention" area); According to the three-dimensional responsibility library and the yeast-lactic acid bacteria symbiotic model, the symbiotic effectiveness index is calculated; Specifically, an exemplary method for calculating the symbiotic effectiveness index is: Symbiotic performance index = (ethanol output / lactic acid intake) × temperature adaptation coefficient; The calculation method of temperature adaptation coefficient is: ; For example, when the optimal temperature is 28℃ and the actual temperature is 25℃, the temperature adaptation coefficient is 0.8; Based on the three-dimensional responsibility database, each zone is treated as a responsibility zone, and an efficiency threshold for the symbiotic effectiveness index is set. When the symbiotic effectiveness index is lower than the efficiency threshold, a high-risk threshold is triggered, and it is judged as not meeting expectations. The corresponding responsibility zone is marked as a high-risk area. Record the responsible zones and symbiotic performance indexes that trigger alarms, generate high-risk zone responsibility logs (e.g., "Upper Layer - XY01, Symbiotic Performance Index = 0.55"), and send these logs to the processing end to alert relevant personnel for manual processing. When the symbiotic effectiveness index is greater than or equal to the effectiveness threshold, it is determined to have met expectations and the responsibility partition is determined to be normal; For the responsibility partitions judged to be normal, according to the responsibility type in the three-dimensional responsibility library (such as "acid suppression and anti-failure"), functional strains (such as CRISPR acid-resistant Clostridium acs gene) are matched for each responsibility partition from the functional gene annotations in the multi-dimensional data fusion framework to obtain the corresponding binding strains; At the same time, based on the metabolic pathway information and functional gene annotations in the multi-dimensional data fusion framework, the metabolic targets for each responsibility type are defined (e.g., “acid suppression and anti-pollution” requires an acetic acid conversion rate ≥ 55%). The process is "acid suppression to prevent spoilage" → activation of the acs gene's acetate → ethanol conversion pathway; Exemplary: Acid suppression and spoilage prevention: reducing acidity through acetic acid → ethanol conversion mediated by acs gene; Oxygenation and fermentation maintenance: Optimize yeast metabolic efficiency by regulating NADH activity; According to the responsibility type and environmental risk assessment, the hierarchical strategy container (such as lower_acid conversion.SPM) is dynamically loaded through the preset loading strategy, and real-time parameters (such as acetic acid concentration and temperature) are extracted from the environmental-metabolism correlation dataset and injected into the hierarchical strategy container; Specifically, the layered strategy containers are classified according to the physical level of the cellar (upper, middle, lower) and the type of responsibility (such as oxygenation to maintain fermentation, acid suppression to prevent spoilage); It is divided into three types: upper layer, middle layer, and lower layer, each with a hierarchical policy container; Exemplary divisions and nomenclature: upper_oxygenation.SPM, middle_signaling.SPM, lower_acid conversion.SPM; Each hierarchical strategy container contains the metabolic regulation logic (e.g., acid conversion pathway optimization, dissolved oxygen regulation algorithm) and parameter binding rules (e.g., acetate concentration threshold, temperature adaptation coefficient) at that level; For different partitions in the same layer, the tiering strategy containers will be slightly different. For example, for the lower layer - XY32, its tiering strategy container strategy is: Use CRISPR acid-tolerant Clostridium to achieve an acetic acid conversion rate ≥ 55%. The tiering strategy container belongs to Acid Conversion_SPM, with the goal of inhibiting rancidity and improving acetic acid conversion efficiency. The loading conditions for the loading strategy are to select the corresponding container (such as Lower Layer_Acid Conversion.SPM) based on the responsibility type (such as "acid suppression and spoilage prevention") in the three-dimensional responsibility library; when the environmental risk assessment is greater than the preset environmental risk warning threshold (such as >80), load the enhanced layered strategy container (such as Lower Layer_Acid Conversion_SPM_Enhanced Version); The loading granularity of the loading policy is to load a layered policy container for each layer, rather than loading each partition independently; the container is only loaded / unloaded when the responsibility type or environmental risk assessment changes (such as switching from upper_oxygen_SPM_emergency_version to upper_oxygen_SPM_emergency_version); The process of dynamically loading the hierarchical policy container is to obtain the responsibility type of the current partition (such as "acid suppression and anti-failure") from the three-dimensional responsibility library; Match policy containers based on responsibility type (e.g., layer_acidconversion.SPM); If the environmental risk assessment is greater than the preset environmental risk warning threshold, the enhanced hierarchical policy container is loaded; Use asynchronous loading mechanisms (such as dynamic class loading technology) to asynchronously load layered policy container files to avoid blocking the main thread; The control logic inside the layered strategy container is dynamically adjusted according to the injected parameters (for example, the acid conversion acceleration mode is activated when the acetic acid concentration is >6 mg / L); Define the unloading conditions for the layered strategy container as a change in responsibility type (e.g., from "acid suppression to prevent spoilage" to "oxygenation to maintain fermentation"), or a decrease in the environmental risk assessment to a safety threshold (e.g., <50); Uninstall unused containers and free up memory; When the layered policy container is updated (for example, a new acid conversion algorithm is added), the new container (for example, the policy container named: Lower Layer_Acid Conversion_SPM_v2) is loaded by matching the version number. If a new version of the layered policy container causes an exception, it will automatically roll back to the old version (for example, the policy container named: Lower Layer_Acid Conversion_SPM_v1); Integrate partition coordinates, responsibility types, bound strains, metabolic targets, and hierarchical strategy containers to establish a bacterial community binding rule table; Example flora binding rule table: [partition coordinates, responsibility type, binding strain, metabolic target, hierarchical strategy container]; [Upper layer - XY01, oxygenation and fermentation, yeast S.cerevisiae, ethanol yield ≥ 70%, upper layer_oxygenation.SPM]; [Lower layer - XY32, acid suppression and anti-failure, CRISPR acid-tolerant Clostridium, acetic acid conversion rate ≥ 55%, lower layer_acid conversion.SPM]; [middle-XY20, signaling, Lactobacillus brevis, signaling efficiency ≥ 60%, middle_signal.SPM]; Extract the flavor association profile of the partition (e.g., ethyl acetate / ethyl lactate ratio) from the historical metabolome flavor association profile, and calculate the flavor enhancement potential of the current flavor association profile (e.g., ratio improvement magnitude) as the flavor weight of the partition; An exemplary calculation method is: Set a flavor enhancement threshold range, such as [5%, 20%]. If the ethyl acetate / ethyl lactate ratio increases by >20%, the flavor weight is 0.9; if the ethyl acetate / ethyl lactate ratio increases by 5%-20%, the flavor weight is 0.6; if the ethyl acetate / ethyl lactate ratio increases by <5%, the flavor weight is 0.3. Timeliness is calculated based on the delay time and response threshold of real-time data transmission of sensors in the partition as a timeliness evaluation; Specifically, the required response time is defined for the response threshold, such as 10 minutes for dissolved oxygen < 5%. The ratio of the actual delay time to the required response time is then calculated, and the difference between 1 and the ratio is taken as the timeliness assessment, that is, timeliness assessment = 1-(delay time / required response time); The microbial binding rules and hierarchical strategy containers corresponding to each partition in the microbial binding rule table are used as the current strategy (these strategies aim to achieve specific metabolic goals (such as increasing the acetic acid conversion rate to 70%)). Based on the resource usage of the current strategy and historical conflict records, the conflict risk is calculated. Specifically, it detects whether the current strategy in the partition shares key resources with other strategies in the partition. If so, the resource conflict coefficient is increased; if not, the resource conflict coefficient is decreased. The specific calculation method is: First, identify all key parameters that may affect the same resource (e.g., temperature, pH, dissolved oxygen concentration); Then, check whether the current strategy and other strategies have opposite requirements on these key parameters or need to be adjusted significantly in the same time period; According to the degree of resource conflict, a resource conflict coefficient between 0 and 1 is assigned, for example, no conflict at all is 0.1, high conflict is 0.9, or no conflict at all is 0.1, high conflict is 0.5; Extract the historical conflict count of the partition (for example, the number of conflicts in the past 30 days for "lower layer-XY32" is 2), and then calculate the conflict risk based on the historical conflict count and resource conflict coefficient; Specifically, conflict risk = resource conflict coefficient × (number of historical conflicts / 100); It should be noted that multiple different operations or control measures may exist simultaneously in the same partition, which is considered as other strategies; For example: different types of metabolic activity control (such as temperature regulation strategies and pH regulation strategies); different stage strategies implemented at different time points within the same partition (such as early oxygenation to maintain fermentation strategy, late acid suppression to prevent spoilage strategy); independent but mutually influential growth regulation strategies in a multi-microbial symbiotic system; Each strategy uses certain environmental resources (e.g., temperature regulation, dissolved oxygen level adjustment, etc.). When two or more strategies need to adjust the same resources, resource competition or conflict may occur; For example, if one strategy requires increasing dissolved oxygen to promote the growth of a certain microorganism, while another strategy hopes to reduce dissolved oxygen to inhibit the growth of another microorganism, this constitutes a resource conflict; Then, a priority arbitration formula is constructed based on flavor weight, timeliness evaluation, and conflict risk to calculate the policy priority; The priority arbitration formula is constructed as: Strategy priority = proportional factor 1 × flavor weight + proportional factor 2 × timeliness assessment - proportional factor 3 × conflict risk; Among them, scale factor 1, scale factor 2 and scale factor 3 are preset values, such as 0.4, 0.3 and 0.3, used to emphasize the importance of different values, and the sum of the three scale factors is 1; Based on policy priorities, hierarchical policy containers are divided into different conflict interception levels; Specifically, an exemplary method for dividing conflict interception levels is as follows: Set a policy priority threshold range, assuming the policy priority threshold range is 0.6-0.8; Strategies with a strategy priority ≥ 0.8 are defined as high priority; The policy with a policy priority of 0.6 ≤ policy priority < 0.8 is defined as medium priority; Strategies with a strategy priority less than 0.6 are defined as low priority; Exemplary: partition = lower-XY32; Flavor weight = 0.87 (ethyl acetate / ethyl lactate ratio increased significantly); Timeliness assessment = 0.9 (response time < 5 minutes); Conflict risk = 0.01 (resource conflict coefficient = 0.5, number of historical conflicts = 2); Policy priority = 0.83 → Conflict interception level = high priority; Then, partition coordinates, hierarchical strategy container IDs, and metabolic targets are extracted from the bacterial community binding rule table, and conflict interception levels (high, medium, and low) are assigned according to the strategy priority to generate a three-dimensional topology of time, space, function, and effectiveness. Exemplary conflict interception rules are: High priority: Execute the policy container (such as acid conversion_SPM) immediately and freeze other low-priority policies in the same partition; Medium priority: The policy container is executed after a certain delay (for example, 10 minutes), and a conflict log is recorded (for example, "Lower-XY32 policy delayed 10 minutes"). Low priority: Pauses policy container execution and triggers manual review (e.g., "Mid-level-XY20 policy requires manual intervention"). The three-dimensional topology is constructed into a table format to obtain a policy container scheduling instruction set, including partition coordinates, container ID, performance target, conflict interception level, and policy priority; The four-level gate conflict interception is defined as: comparing the matching degree between the functional genes of the foreign bacteria and the receptors of the native bacteria through the gene compatibility gate; scanning the bottleneck of target metabolite synthesis through the metabolic pathway saturation rate gate; Integrate gene compatibility gate information and metabolic pathway saturation gate information to generate a gene-metabolism conflict detection report; According to the gene-metabolism conflict detection report, the strain environmental adaptability is calculated through the attenuation risk value gate; the conflict level is determined through the dynamic arbitration gate; Methods for comparing the degree of compatibility between exogenous bacterial functional genes and native bacterial receptors through gene compatibility gates include: Extract the functional genes of the bound strain (such as the acs gene of CRISPR acid-resistant Clostridium) from the strategic container scheduling instruction set as the functional genes of the exogenous bacteria, and call the receptor proteins of the indigenous bacteria in the cellar (such as the SdiA protein) in the historical multi-omics database; Use the BLAST tool to calculate the sequence similarity between the functional gene of the foreign bacteria and the receptor protein of the native bacteria (e.g., the similarity between the acs gene and the SdiA protein = 58%); The AlphaFold tool was used to predict the three-dimensional protein structures of the exogenous bacterial functional gene and the native bacterial receptor protein, and the structural compatibility of the binding sites (such as the number of hydrogen bonds and hydrophobic interactions) between the two was analyzed (e.g., structural compatibility = 60%). Combined with the metabolome flavor association profile, the functional fitness of the exogenous bacterial functional genes and the metabolic requirements of the native bacterial receptor proteins was verified (e.g., functional fitness = 55%). Specifically, target metabolites (e.g., ethyl acetate, ethyl lactate) were identified based on metabolomic flavor association profiles in historical multi-omics databases (e.g., the correlation between the ethyl acetate / ethyl lactate ratio and flavor quality); For example, the ethyl acetate / ethyl lactate ratio directly affects the flavor of the fermentation product (such as the body or acidity of the wine); Predict the metabolic potential of exogenous bacterial functional genes (e.g., whether they can synthesize / decompose ethyl acetate) through metagenomic functional annotation libraries (e.g., the acs gene of CRISPR acid-tolerant Clostridium); After introducing exogenous bacteria into the pit, samples are collected and the actual concentrations of target metabolites are detected using targeted metabolomics techniques (e.g., LC-MS); Experimental group settings: ethyl acetate / ethyl lactate ratio = actual ratio after introduction of exogenous bacteria; Set an ideal value: a ratio based on the metabolic needs of the indigenous flora determined by historical data (e.g., 1.2:1); The ratio of the experimental group ratio to the ideal value is taken as the functional fitness; Combining sequence similarity, structural fitness and functional fitness, the matching score of functional genes was calculated by weighted fusion method, which was used as the matching evaluation of functional genes; Specifically, a weight ratio is assigned to sequence similarity, structural fitness, and functional fitness respectively, and then multiplied respectively. The three multiplication results are then added together to obtain the matching evaluation; If the partition's matching evaluation is less than the preset matching threshold, it is determined to have not met expectations and the gene compatibility gate alarm is activated; Methods for scanning the target metabolite synthesis bottleneck through metabolic pathway saturation rate gates include: Extract target metabolites (e.g., ethyl acetate, glycerol, serine) from environmental-metabolism association datasets; Then, the target metabolites were detected by LC-MS / GC-MS (liquid chromatography-mass spectrometry / gas chromatography-mass spectrometry), and the pathway saturation rate was calculated in combination with the metabolic flux model (FBA); If the pathway saturation rate is greater than a preset saturation rate threshold (e.g., 85%), it is determined that the expected rate has not been achieved, that is, a metabolic bottleneck exists in the partition (e.g., GPD1 enzyme overexpression leads to glycerol diversion >12%), and the bottleneck type (e.g., substrate competition, product inhibition) is identified. Based on the bottleneck type, the preset caching strategy (e.g., serine buffering protocol, NADH regeneration enhancement) is initiated. It should be noted that the gene-metabolism conflict detection report includes gene compatibility gate information (partition coordinates, exogenous bacterial functional genes, native bacterial receptor proteins, and matching evaluation (e.g., the matching degree between the XY32 CRISPR acid-resistant Clostridium acs gene and SdiA is 58.2 points)) and metabolic pathway saturation gate information (target metabolites with metabolic bottlenecks, pathway saturation rate, and triggered caching strategies (e.g., the overexpression of GPD1 results in glycerol diversion >12%, triggering the serine buffering protocol)). According to the gene-metabolism conflict detection report, the methods for calculating the strain environmental adaptability through the attenuation risk value gate include: Obtain the temperature fluctuation and pH value change of the partition, and combine it with the matching evaluation to calculate the strain attenuation risk value as the strain attenuation risk; A formula for calculating the strain attenuation risk value is: ; in, represents the strain attenuation risk value, e represents the natural constant, Indicates the difference between the actual temperature of the partition and the optimal temperature. It represents the difference between the actual pH value and the metabolically adapted pH value; and is the proportional constant (usually set to 0.3 and 0.2), and 100 is the maximum value of the matching evaluation; The logic of the formula is that if the genetic compatibility score is lower than a certain score (such as 60 points), the strain attenuation risk value will increase, triggering a higher risk warning; It should be noted that the actual pH value and metabolically adapted pH value refer to the pH value of the target strain. This target strain can be an exogenous strain or an indigenous strain, depending on actual needs. In genetic compatibility testing through a genetic compatibility gate, the functional genes of the exogenous bacteria (such as the acs gene of CRISPR-resistant Clostridium) are compared with the metabolic requirements of the indigenous bacteria. At this time, the metabolically adapted pH value of the exogenous bacteria is usually deduced from the characteristics of its gene sequence and metabolic pathway. For example, the metabolic pathway of the strain is queried through gene annotation databases (such as KEGG and UniProt) to determine the optimal pH value of its key enzyme. For example, if the key enzyme of acetate metabolism in the exogenous bacteria is acetyl-CoA synthetase (ACS), its optimal pH may be 6.5-7.0. The metabolically adapted pH value of the indigenous bacteria can be derived from historical environmental data (e.g., the pH tolerance range of the pit microbiome) or a table of bacterial community binding rules. For example, if the indigenous bacteria are acid-tolerant Clostridium, their metabolically adapted pH value may be 4.0-5.0. When calculating the strain attenuation risk value, the metabolic adaptation pH value is used to measure the degree to which the actual pH value deviates from the strain's optimal metabolic environment; If the strain decay risk value is greater than the preset strain decay threshold (e.g., 1.0), it is determined that the strain decay risk has not met expectations, triggering a high-risk warning for the partition (e.g., lower layer - XY32, sj=1.10), and adding a cooling sub-strategy (e.g., cooling rate = 0.5°C / h, target temperature = 25°C); The methods for determining conflict levels through dynamic arbitration gates include: Based on the matching evaluation and metabolic bottleneck information, the conflict level is determined and classified into different conflict levels; Specifically, one division method is to set a matching evaluation threshold interval (e.g., 60 points to 80 points); If the matching evaluation is greater than or equal to the maximum value of the matching evaluation threshold interval, and there is no metabolic bottleneck (i.e., the pathway saturation rate is less than or equal to the preset saturation rate threshold), it is classified as conflict level 1; If the matching evaluation is greater than or equal to the minimum value of the matching evaluation threshold interval and less than or equal to the maximum value of the matching evaluation threshold interval, or the path saturation rate is greater than or equal to the preset saturation rate threshold and less than or equal to the maximum path saturation rate (i.e., 100%), the conflict level is classified as level 2. If the matching evaluation is less than the minimum value of the matching evaluation threshold interval, or the path saturation rate is greater than the maximum path saturation rate (i.e., 100%), the conflict level is classified as level 3; Taking the conflict interception level as the core and the conflict level as the adjustment factor for policy details, the conflict interception level and the conflict level are then linked in rules to activate the corresponding responsibility transfer strategies for different partitions and simultaneously generate responsibility transfer paths (according to the bacterial community binding rule table, the responsibility type of the partition that needs to split the responsibility unit is transferred to the adjacent partition, and the three-dimensional responsibility library is updated); An exemplary liability transfer strategy is: For situations where the conflict interception level is high priority and the policy priority is immediate execution, if the conflict level is level 1, the responsibility transfer policy is to directly execute the policy container and no adjustment is required (e.g., acid conversion_SPM starts immediately); If the conflict level is level 2, the responsibility transfer strategy is to execute the strategy container, but with additional local intervention (e.g., the cooling sub-strategy rate is increased to 0.8°C / h); If the conflict level is level 3, the responsibility transfer strategy is to execute the strategy container urgently, triggering large-scale responsibility transfer (such as splitting the responsibility unit and injecting buffer); For situations where the conflict interception level is medium priority and the policy priority is to delay for a certain period of time, if the conflict level is level 1, the responsibility transfer policy is to delay the execution of the policy container and record the log (for example, "lower layer - XY32 policy delayed for 10 minutes"); If the conflict level is level 2, the responsibility transfer strategy is to delay the execution of the strategy container and dynamically adjust the metabolic target (e.g., lower the lactate conversion efficiency target to 85%); If the conflict level is level 3, the responsibility transfer strategy is to delay the execution of the policy container, but it is upgraded to a high-priority interception (such as triggering manual review); For situations where the conflict interception level is low priority and the policy priority is suspended execution, if the conflict level is level 1, the responsibility transfer policy is to suspend the policy container and wait for manual review (e.g., "Middle-level - XY20 policy requires manual intervention"); If the conflict level is level 2, the responsibility transfer strategy is to suspend the strategy container, but automatic compensation is allowed (such as temporarily enabling the buffer protocol); If the conflict level is level 3, the responsibility transfer strategy is to suspend the policy container and force manual review (e.g., "lower-level-XY32 policy requires expert intervention"); It should be noted that the design logic of the responsibility transfer strategy is to unconditionally prioritize the case where the conflict interception level is high, and the conflict level only affects the intervention intensity of the specific execution (such as whether to split the responsibility unit); For conflicts with a medium priority level, execution is delayed, but the conflict level determines whether it needs to be upgraded to "high priority" (e.g., a level 3 conflict triggers manual review); For cases where the conflict interception level is low priority, execution is suspended, and the conflict level determines whether automatic compensation is allowed (e.g., a buffer protocol can be temporarily enabled for a level 2 conflict); There is no need to define separate rules for each combination of conflict level and interception level. The core logic is driven only by the conflict interception level, and the conflict level serves as a detail correction factor.

[0020] The specific responsibility transfer strategy can be flexibly changed according to actual conditions. For example, when the conflict interception level = medium and the conflict level = 3, it can be automatically upgraded to high-priority interception to avoid missing critical scenarios; Generate dynamic domestication instruction sets based on the results of the four-level gate conflict interception; The dynamic domestication instruction set includes target coordinates (partition locations where anomalies, conflicts, and adjustments occur (e.g., lower layer - XY32)), conflict interception level, buffer protocol (dynamically selected buffer type and concentration (e.g., 0.1M serine buffer), high priority triggers emergency buffer protocol; medium priority triggers standard buffer protocol; low priority no buffer protocol), cooling sub-strategy (temperature control scheme (e.g., rate 0.5°C / h → target temperature 25°C), high priority triggers rapid cooling; medium priority triggers gradual cooling; low priority no cooling strategy), responsibility transfer path (transfer range of conflict responsibility (e.g., lower layer - XY32 → lower layer - XY33), high priority triggers large-scale responsibility transfer; medium priority triggers local responsibility transfer; low priority no transfer), metabolic bottleneck target metabolites (metabolic bottleneck substances caused by conflicts (e.g., glycerol)), and manual review flag (whether manual intervention is required (e.g., "expert intervention"), low priority triggers manual review; medium priority upgrades to manual review if the conflict level = level 3). According to the dynamic domestication instruction set, the environmental parameters of the simulated layered stress scenario are set, and then the intervention effect of the dynamic domestication instruction set on the bacterial metabolic pathway is verified by simulating the layered stress scenario. The deviation of the bacterial response under environmental stress is simulated to obtain the deviation performance; Specifically, an exemplary environmental parameter setting is as follows: pH value: 3.5 (adjust according to the buffer protocol in the dynamic acclimation instruction set, such as serine buffer to maintain the pH at around 4.0); Dissolved oxygen content: 0.5% (adjusted based on the target temperature of 25°C in the cooling sub-strategy; low temperatures inhibit oxygen dissolution); Metabolic bottleneck target metabolite concentration: Set the initial concentration (e.g., 120% saturation rate) based on the metabolic bottleneck target metabolite (e.g., glycerol) in the dynamic acclimation instruction set; Then, based on the COBRA framework, the microbial gene-metabolic network model and the parameters of the dynamic domestication instruction set were input. Combined with environmental parameters such as pH and dissolved oxygen, the enzyme activity and reaction rate of the metabolic pathway were dynamically adjusted through the environmental stress simulator to simulate the intervention effect of the dynamic domestication instruction set on the microbial metabolic pathway. In the obtained simulation results, focus on whether HSP104 overexpression and metabolic pathway inhibition occur. If so, it indicates that there is a bacterial community response deviation, and the corresponding deviation manifestation is recorded simultaneously; Among them, HSP104 overexpression refers to the overexpression of HSP104 triggered by high acid stress, and the concentration of the metabolic bottleneck target metabolite is greater than the initial concentration (e.g., >120%). Due to high acid stress (e.g., pH = 3.5), HSP104 (heat shock protein) of the acid-tolerant strain is activated, inhibiting the expression of ATF1 (anti-stress transcription factor); Metabolic pathway inhibition refers to the partial blockage of the glycerol metabolic pathway due to the introduction of buffer (serine), resulting in an increase in the lactate synthesis gap (>35%); Combining the transcriptome and metabolome dimensions of transcriptomics-metabolomics, we can trace the root causes of microbial response deviations and attribute them to specific responsibility partitions to obtain attribution partitions. Transcriptome-wide tracing and localization analyzed the overexpression of HSP104. The analysis logic is that high acid stress (such as pH = 3.5) triggers the overexpression of HSP104 (such as > 120%). At this time, the inhibitory correlation between HSP104 and ATF1 shows that the expression of ATF1 is significantly inhibited, resulting in a decrease in the anti-stress ability of the bacterial community. The attribution is achieved by dynamically taming the target coordinates in the instruction set to locate the specific responsible partition, marking the corresponding responsible partition as the source of HSP104 overexpression, and then adjusting the strategies within the hierarchical strategy container of the responsible partition according to the preset strategy correction suggestions (adding the expression regulation strategy of ATF1 (such as introducing the ATF1 promoter enhancer)). The metabolomics dimension traceability focuses on metabolic pathway inhibition, specifically analyzing the ethyl acetate synthesis gap. The analysis logic is based on the metabolic bottleneck target metabolites in the dynamic domestication instruction set (e.g., the glycerol saturation rate of -XY32 = 120%). Due to the excessively high glycerol saturation rate, the metabolic pathway is blocked, and the acetate inhibition synthesis gap widens (>35%). The buffering protocol (serine buffer) further inhibits the activity of glycerol dehydrogenase, exacerbating the metabolic bottleneck. The attribution is achieved by dynamically locating the target coordinates in the instruction set to a specific responsibility partition, marking the corresponding responsibility partition as the source of the ethyl acetate synthesis gap. Then, the strategy within the hierarchical strategy container of the responsibility partition is adjusted synchronously based on the threshold strategy correction suggestions (such as optimizing the buffer protocol (such as reducing the serine concentration to 0.05M) and introducing a glycerol dehydrogenase activator); Using the transcriptome and metabolome as analysis dimensions, the deviation manifestations, attribution partitions, and corresponding correction suggestions of the corresponding analysis dimensions are integrated to generate a deviation root cause analysis report (including analysis dimensions, deviation manifestations, attribution partitions, and correction suggestions). The responsibility transfer path is also updated and the hierarchical strategy container is corrected simultaneously. Based on the modified effect of the layered policy container, the performance baseline of the corresponding partition in the three-dimensional responsibility library is dynamically adjusted to form a new performance baseline; The dynamic adjustment method for the new performance baseline is: New performance baseline = original performance baseline × (1 + scaling factor × traceability correction factor); The scaling factor is a preset constant value used to adjust the correction amplitude of the performance baseline, usually set to 0.2. The traceability correction factor is defined as the traceability correction factor = 1-the degree of repair of the deviation root cause; The degree of repair of the deviation root is determined by the correction effect of the sub-volume strategy container. For example, if the ethyl acetate synthesis gap is repaired from 35% to 10%, the degree of repair of the deviation root is 0.1, and the traceability correction factor is 1-0.1 = 0.9. The updated content of the responsibility matrix includes the partition coordinates (such as the lower layer - XY32), the original performance baseline (such as 55%), the correction factor (such as 0.9), and the new performance baseline (such as 66%); At the same time, based on the correction suggestions in the deviation root cause analysis report, the layered policy container parameters are optimized and high-conflict responsibility partitions are split; Specifically, to optimize the tiered strategy container parameters, assume that the correction suggestion in the deviation root cause analysis report is to adjust the concentration of serine buffer_SPM from 0.1M to 0.05M. Then, according to this correction suggestion, adjust the tiered strategy container accordingly; Partitions with a conflict level of 3 or a traceability correction factor less than a preset correction threshold (such as 0.5) are regarded as high-conflict responsibility partitions. As for the splitting operation, the high-conflict responsibility partitions are split into different types of partitions. For example, the lower layer - XY32 is split into XY32A (main metabolic pathway) and XY32B (auxiliary buffer regulation); Integrate the split high-conflict responsibility partition coordinates, new performance baselines and hierarchical strategy container parameters to generate a responsibility system evolution package, and then reversely update the three-dimensional responsibility library and hierarchical strategy container.

[0021] Example 2 See also Figure 3 As shown, for the parts not described in detail in this embodiment, please refer to the description of Example 1. A multi-omics-based intelligent domestication method for a Maotai-flavor liquor-producing yeast population is provided, comprising: S1: Deploy a multi-source biosensor network to collect multi-source data on cellar environment and metabolism, access a historical multi-omics database to build a multi-dimensional data framework; then correct for sensor spatiotemporal delays to generate a spatiotemporally aligned environmental-metabolism association dataset; define response thresholds for each partition to conduct regional environmental risk assessments, and construct a three-dimensional responsibility library; S2: Based on the three-dimensional responsibility library and the yeast-lactic acid bacteria symbiotic performance index, the risk status of the responsibility partition is determined. The hierarchical policy container is dynamically loaded and injected with real-time parameters to generate a bacterial community binding rule table. Based on the bacterial community binding rule table, the conflict interception level is divided using a priority arbitration formula, and a three-dimensional topology of time, space, function, and performance is generated and tabulated as a policy container scheduling instruction set. S3: Based on the policy container scheduling instruction set, it executes the four-level structured gate interception of gene compatibility, metabolic pathway saturation rate, attenuation risk value and dynamic arbitration, thereby generating a dynamic domestication instruction set and simultaneously recording the responsibility transfer path; S4: Based on the dynamic domestication instruction set and responsibility transfer path, the microbial response deviation is analyzed by simulating the hierarchical stress scenario, and the root cause of the deviation is located by combining transcriptomics-metabolomics tracing; the performance baseline and hierarchical strategy container parameters are dynamically updated, the responsibility system evolution package is generated, and the three-dimensional responsibility library and hierarchical strategy container are reversely updated.

[0022] Example 3 This embodiment discloses an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the operation mode of the multi-omics-based intelligent domestication method for a sauce-flavor liquor-producing yeast population is implemented.

[0023] Since the electronic device introduced in this embodiment is an electronic device used to implement the multi-omics-based intelligent domestication method for the production of yeast flora of sauce-flavor liquor in the embodiment of this application, based on the multi-omics-based intelligent domestication method for the production of yeast flora of sauce-flavor liquor in the embodiment of this application, those skilled in the art can understand the specific implementation of the electronic device of this embodiment and its various variations, so how the electronic device implements the method in the embodiment of this application will not be introduced in detail here. As long as those skilled in the art implement the electronic device used in the multi-omics-based intelligent domestication method for the production of yeast flora of sauce-flavor liquor in the embodiment of this application, they all fall within the scope of protection to be protected by this application.

[0024] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters and thresholds in the formulas are set by technicians in this field according to actual conditions.

[0025] The above are merely preferred embodiments of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions based on the principles of the present invention are within the scope of protection of the present invention. It should be noted that for users of ordinary skill in the art, various improvements and modifications that do not depart from the principles of the present invention are also within the scope of protection of the present invention.

Claims

1. A multi-omics-based intelligent domestication system for yeast flora in the production of Maotai-flavor liquor, characterized by: include: Environmental-Metabolism Fusion Unit: This unit collects multi-source data on pit environment and metabolism, connects to a historical multi-omics database, and constructs a multi-dimensional data framework. This framework then corrects for sensor spatiotemporal delays to generate an environmental-metabolism correlation dataset. Furthermore, it defines response thresholds for each partition, conducts regional environmental risk assessments, and constructs a three-dimensional responsibility database. The microbial topology binding unit determines the risk status of the responsibility partition based on the three-dimensional responsibility library and the symbiotic performance index, matches functional strains, and dynamically loads the hierarchical policy container to generate a microbial binding rule table. It also divides the conflict interception level through priority arbitration and generates a three-dimensional topology and policy container scheduling instruction set. Structured gate interception unit: This unit performs four-level structured gate interception based on the policy container scheduling instruction set, generates a dynamic domestication instruction set, and simultaneously records the responsibility transfer path. Multi-omics closed-loop evolution unit: Based on a dynamic domestication instruction set and responsibility transfer path, the unit analyzes microbial response deviations by simulating hierarchical stress scenarios and locates the root causes of the deviations by combining transcriptomics and metabolomics. It dynamically updates the performance baseline and hierarchical strategy container parameters to generate a closed-loop iterative responsibility system evolution package.

2. The multi-omics-based intelligent domestication system for Maotai-flavor liquor-producing yeast flora according to claim 1 is characterized in that: The generation method of the environment-metabolism association dataset includes: Deploy a multi-source biosensor network to collect multi-source data in real time, including physical environment data, biological metabolism data, and chemical parameter data from the upper, middle, and lower layers of the cellar; Access historical multi-omics databases, extract metagenomic functional annotation libraries and metabolomics flavor association profiles, and then associate and map with multi-source data to form a multi-dimensional data fusion framework; Standardize the multi-source data collected by different sensors, correct the timestamps, and generate multi-source data aligned in time and space; The cellars were divided into different zones, and environmental risk assessments for each zone were obtained based on spatiotemporally aligned multi-source data. Response thresholds for each zone were defined based on a historical multi-omics database. The timestamps, coordinates of each partition, spatiotemporally aligned multi-source data and environmental risk assessment are integrated to generate a spatiotemporally aligned environmental-metabolic association dataset; at the same time, the response thresholds of each partition are integrated to generate an environmental risk threshold table.

3. The multi-omics-based intelligent domestication system for Maotai-flavor liquor-producing yeast flora according to claim 2 is characterized in that: The construction method of the three-dimensional responsibility library includes: Based on the environmental-metabolism association dataset and environmental risk threshold table, the responsibility type of each partition is defined and labeled; the performance baseline of each partition is set using the historical metabolomics flavor association profile; At the same time, the maximum metabolic load of each partition is annotated by combining the environmental risk assessment in the environment-metabolism association dataset; The spatial coordinates, responsibility type, response threshold, performance baseline and maximum metabolic load of the partition are integrated to construct a three-dimensional responsibility library.

4. The multi-omics-based intelligent domestication system for Maotai-flavor liquor-producing yeast flora according to claim 3 is characterized in that: The generation method of the bacterial community binding rule table includes: Based on the three-dimensional responsibility database and the yeast-lactic acid bacteria symbiosis model, a symbiotic performance index is derived. When the symbiotic performance index fails to meet expectations, a high-risk warning is triggered and a high-risk area responsibility zoning log is generated. When the symbiotic effectiveness index reaches the expected level, the responsibility partition is judged to be normal; Based on the normal responsibility partitions, functional strains were matched from the functional gene annotations of the multi-dimensional data fusion framework and metabolic targets were defined; Dynamically load hierarchical policy containers and inject real-time parameters based on liability type and environmental risk assessment; Integrate partition coordinates, responsibility types, binding strains, metabolic targets and hierarchical strategy containers to establish a bacterial community binding rule table.

5. The multi-omics-based intelligent domestication system for Maotai-flavor liquor-producing yeast flora according to claim 4 is characterized in that: The three-dimensional topology is generated by: Extract the flavor weight of each partition from the historical metabolome flavor association spectrum to obtain the timeliness evaluation and conflict risk; Combining flavor weight, timeliness evaluation, and conflict risk, priority arbitration is performed to obtain policy priorities. Based on policy priorities, layered policy containers are divided into different conflict interception levels. Extract partition coordinates, hierarchical strategy container IDs, and metabolic targets from the bacterial community binding rule table, assign conflict interception levels based on strategy priorities, and generate a three-dimensional topology of spatiotemporal-functional-efficacy. The three-dimensional topology is constructed into a table form to obtain a policy container scheduling instruction set.

6. The multi-omics-based intelligent domestication system for Maotai-flavor liquor-producing yeast flora according to claim 5 is characterized in that: The four-level structured gate interception is defined as: gene compatibility gate, metabolic pathway saturation rate gate, attenuation risk value gate and dynamic arbitration gate; Compare the matching degree between the functional genes of foreign bacteria and the receptors of indigenous bacteria through the gene compatibility gate, including: Extract the native bacteria receptor proteins from the strategic container scheduling instruction set and simultaneously extract the exogenous bacteria functional genes; Obtain the sequence similarity, structural compatibility, and functional compatibility of the functional genes of the exogenous bacteria and the receptor proteins of the native bacteria, and then evaluate the matching degree of the functional genes; when the matching degree evaluation of the partition does not meet the expectations, activate the gene compatibility gate alarm; Scan target metabolite synthesis bottlenecks through metabolic pathway saturation rate gates, including: Target metabolites are extracted from the environmental-metabolism association dataset; the pathway saturation rate of the target metabolites is obtained by combining the metabolic flux model; if the pathway saturation rate does not meet expectations, it is determined that there is a metabolic bottleneck in the partition, and the preset caching strategy is activated.

7. The multi-omics-based intelligent domestication system for Maotai-flavor liquor-producing yeast flora according to claim 6 is characterized in that: The generation method of the dynamic domestication instruction set includes: According to the gene-metabolism conflict detection report, the strain environmental adaptability is obtained through the attenuation risk value gate, including: Obtain the temperature fluctuations and pH changes of each partition, and combine them with the matching evaluation to assess the risk of strain attenuation. If the risk of strain attenuation does not meet expectations, a high-risk warning for the partition is triggered, and a cooling sub-strategy is added. Determine the conflict level through dynamic arbitration gate, including: Different conflict levels are divided according to the matching evaluation and metabolic bottleneck information; The conflict interception level is linked to the conflict level, the responsibility transfer strategy is activated, and the responsibility transfer path is generated synchronously; based on the results of the four-level gate conflict interception, a dynamic domestication instruction set is generated.

8. The multi-omics-based intelligent domestication system for Maotai-flavor liquor-producing yeast flora according to claim 7 is characterized in that: The method of analyzing microbial response deviations by simulating hierarchical stress scenarios and combining transcriptomics and metabolomics to locate the root causes of the deviations includes: Based on the dynamic domestication instruction set, a simulated hierarchical stress scenario was set to verify the intervention effect of the dynamic domestication instruction set on the bacterial metabolic pathway, and the bacterial response deviation was detected to obtain the deviation performance; Combining the transcriptome and metabolome dimensions of transcriptomics-metabolomics, we can locate the root cause of the microbial response deviation and attribute it to a specific partition to obtain the attribution partition. Taking the transcriptome and metabolome as analysis dimensions, the deviation performance, attribution partitioning and corresponding correction suggestions of the corresponding analysis dimensions are integrated to generate a deviation root cause analysis report, synchronously update the responsibility transfer path and correct the hierarchical strategy container.

9. The multi-omics-based intelligent domestication system for Maotai-flavor liquor-producing yeast flora according to claim 8 is characterized in that: The generation method of the responsibility system evolution package includes: Dynamically adjust the partition performance baseline based on the modified effect of the layered policy container; Based on the deviation root cause analysis report, optimize the layered policy container parameters and split out high-conflict responsibility partitions; Integrate the split high-conflict responsibility partition coordinates, new performance baselines and hierarchical strategy container parameters to generate a responsibility system evolution package, and then reversely update the three-dimensional responsibility library and hierarchical strategy container.

10. A multi-omics-based intelligent domestication method for a Maotai-flavor liquor-producing yeast community, which is implemented based on the multi-omics-based intelligent domestication system for a Maotai-flavor liquor-producing yeast community according to any one of claims 1 to 9, characterized in that: include: S1: Collect multi-source data, correct sensor spatiotemporal delays, and generate environment-metabolism correlation datasets; And define the response threshold of each zone to calculate the zone environmental risk assessment and build a three-dimensional responsibility library; S2: Based on the three-dimensional responsibility library and combined with the symbiotic performance index, the risk status of the responsibility partition is determined, the functional strains are matched and the hierarchical strategy container is dynamically loaded to generate a bacterial community binding rule table, and a three-dimensional topology and strategy container scheduling instruction set; S3: Based on the policy container scheduling instruction set, it executes four-level structured gate interception, generates a dynamic domestication instruction set, and simultaneously records the responsibility transfer path; S4: Based on a dynamic domestication instruction set and responsibility transfer path, the microbial response deviation is analyzed by simulating hierarchical stress scenarios, and the root cause of the deviation is located in combination with transcriptomics-metabolomics; the performance baseline and hierarchical strategy container parameters are dynamically updated to generate a closed-loop iterative responsibility system evolution package.

Images