Multi-omics based intelligent domestication system and method for zhang-flavor liquor yeast population

Abstract

The present application belongs to the technical field of liquor intelligent brewing, and discloses a system and method for intelligent domestication of a yeast population for producing Maotai-flavor liquor based on multi-omics, which comprises the following steps: collecting multi-source data, constructing a multi-dimensional data framework, correcting the time-space delay of sensors, generating environment-metabolism correlation data sets, defining the response threshold of each partition to perform partition environment risk assessment, and constructing a three-dimensional responsibility library; determining the responsibility partition risk state by combining the symbiotic performance index, generating a bacterial population binding rule table; dividing conflict interception levels, generating a three-dimensional topology; performing four-level structured gate interception, and generating a dynamic domestication instruction set; analyzing the response deviation of the bacterial population by simulating a hierarchical stress scene, and locating the root cause of the deviation by combining transcriptomics and metabolomics; and dynamically updating the performance baseline and hierarchical strategy container parameters, which solves the problems of low yeast fermentation efficiency, poor stress resistance, and extensive flavor regulation in traditional processes, and realizes dynamic optimization of microbial communities and system adaptive capacity.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent liquor brewing, more particularly, the present application relates to a system and method for intelligent domestication of a yeast population for producing Maotai-flavor liquor based on multi-omics. BACKGROUND

[0002] Maotai-flavor liquor production relies on complex microbial communities, especially the metabolic activities of yeast strains for producing liquor. The performance of the yeast strains directly determines the flavor of the liquor and the stability of the quality. In traditional processes, the selection and optimization of yeast strains rely on empirical operations and single-omics (such as metagenomics or metabolomics) analysis, which has problems such as low fermentation efficiency of yeast strains, poor stress resistance (such as sensitivity to high temperature and high acid environment), and unstable synthesis of flavor substances. In addition, in an open fermentation system, the dynamic changes of the microbial community make it difficult for traditional methods to optimize the core microbial community, resulting in frequent abnormal fermentation (such as rancidity and moldy smell) and significant fluctuations in the quality of the liquor.

[0003] To address the above problems, the existing technology still has obvious shortcomings in yeast strain selection and process control: first, the selection method (such as WL agar medium) relies on phenotypic observation, which is low in efficiency and poor in specificity, making it difficult to quickly identify high-performance strains; second, the regulation of fermentation parameters (such as temperature and pH) lacks theoretical basis, which can easily cause metabolic pathway blockage or accumulation of off-flavor substances (such as acetic acid and isovaleric acid); third, the flavor regulation means is rough, and there is a lack of systematic modeling research on the synthesis path of key flavor substances (such as esters and alcohols) and the microorganism-metabolism-environment system, making it difficult to achieve precise intervention. These limitations seriously restrict the improvement of the production efficiency and quality stability of Maotai-flavor liquor. SUMMARY

[0004] In order to overcome the above-mentioned defects of the prior art, in order to achieve the above-mentioned purposes, the present application provides the following technical scheme: a system for intelligent domestication of a yeast population for producing Maotai-flavor liquor based on multi-omics, comprising:

[0005] An environment-metabolism fusion unit: a multi-source biosensor network is deployed to collect multi-source data of the environment and metabolism of the pit pool, access a historical multi-omics database to construct a multi-dimensional data framework, then correct the time-space delay of the sensor to generate time-space aligned environment-metabolism correlation data sets, and define response thresholds for each partition to perform partition environment risk assessment and construct a three-dimensional responsibility library;

[0006] A microbial community topology binding unit: based on the three-dimensional responsibility library, the responsibility partition risk state is determined in combination with the yeast-lactic acid bacteria symbiotic performance index, a hierarchical strategy container is dynamically loaded and real-time parameters are injected to generate a microbial community binding rule table; according to the microbial community binding rule table, a conflict interception level is divided through priority arbitration to generate a three-dimensional topology of time-space-function-performance, and the three-dimensional topology is tabled into a strategy container scheduling instruction set;

[0007] The structured gate interception unit: according to the policy container scheduling instruction set, four-level structured gate interception of gene compatibility, metabolic pathway saturation rate, attenuation risk value and dynamic arbitration is performed, and then dynamic domestication instruction set is generated, and responsibility transfer path is recorded synchronously;

[0008] The multi-omics closed-loop evolution unit: based on the dynamic domestication instruction set and the responsibility transfer path, the response deviation of the bacterial community is analyzed by simulating the hierarchical stress scenario, the deviation source is located by combining the transcriptome-metabolome tracing, the efficiency baseline and the hierarchical policy container parameters are dynamically updated, the responsibility system evolution package is generated, and the three-dimensional responsibility library and the hierarchical policy container are reversely updated.

[0009] Further, the generation mode of the environment-metabolism correlation data set comprises:

[0010] The multi-source data comprises physical environment data, biological metabolism data and chemical parameter data of the upper layer, the middle layer and the lower layer of the pit pool;

[0011] The collected multi-source data is subjected to data cleaning and standardization processing;

[0012] The historical multi-omics database is accessed, and the metagenome functional annotation library and the metabolome flavor correlation spectrum are extracted;

[0013] The multi-dimensional data fusion framework is formed by correlating and mapping the standardized multi-source data with the functional gene annotation of the metagenome functional annotation library and the metabolic pathway information of the metabolome flavor correlation spectrum;

[0014] The transmission delay time of the lower layer sensor data is calculated, and then the compensation coefficient is calculated according to the delay time, and the compensation coefficient is taken as the weight of the corresponding sensor;

[0015] The multi-source data collected by different sensors is unified into a standard unit, the time stamp is corrected, and the spatiotemporally aligned multi-source data is generated;

[0016] The pit pool is divided into different partitions, the correlation model of dissolved oxygen concentration and acetic acid accumulation accident in each partition of the pit pool is established based on the historical multi-omics database, and the response threshold of each partition is defined;

[0017] Based on the spatiotemporally aligned multi-source data, the environmental risk assessment of each partition is obtained;

[0018] The time stamp, the coordinates of each partition, the spatiotemporally aligned multi-source data and the environmental risk assessment are integrated to generate the spatiotemporally aligned environment-metabolism correlation data set; meanwhile, the response threshold of each partition is integrated to generate the environmental risk threshold table.

[0019] Further, the construction mode of the three-dimensional responsibility library comprises:

[0020] Based on the spatio-temporal alignment of multi-source data in the environmental-metabolic correlation dataset and the environmental risk threshold table, the responsibility type of each partition is defined and labeled;

[0021] Using the historical metabolome flavor correlation spectrum to set the performance baseline of each partition;

[0022] At the same time, combined with the environmental risk assessment in the environmental-metabolic correlation dataset, the maximum metabolic load of each partition is labeled;

[0023] Integrate the spatial coordinates, responsibility type, response threshold, performance baseline and maximum metabolic load of the partition to form a three-dimensional responsibility library.

[0024] Further, the generation method of the bacterial community binding rule table comprises:

[0025] According to the three-dimensional responsibility library, based on the yeast-lactic acid bacteria symbiotic model, the symbiotic performance index is obtained;

[0026] Based on the three-dimensional responsibility library, each partition is regarded as a responsibility partition, and when the symbiotic performance index does not reach the expectation, a high-risk warning is triggered, and the corresponding responsibility partition is marked as a high-risk area;

[0027] Record the responsibility partition triggering the alarm and the symbiotic performance index to generate a high-risk responsibility partition log;

[0028] When the symbiotic performance index reaches the expectation, the responsibility partition is determined to be normal;

[0029] For the responsibility partition determined to be normal, according to the responsibility type in the three-dimensional responsibility library, the functional strain of each responsibility partition is matched from the functional gene annotation in the multi-dimensional data fusion framework, and the corresponding binding strain is obtained;

[0030] At the same time, based on the metabolic pathway information and functional gene annotation in the multi-dimensional data fusion framework, the metabolic target of each responsibility type is defined;

[0031] According to the responsibility type and the 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-metabolic correlation dataset and injected into the hierarchical strategy container;

[0032] Integrate the partition coordinates, responsibility type, binding strain, metabolic target and hierarchical strategy container to establish a bacterial community binding rule table.

[0033] Further, the generation method of the three-dimensional topology comprises:

[0034] Extract the flavor correlation spectrum of the partition from the historical metabolome flavor correlation spectrum, and extract the flavor improvement potential of the current flavor correlation spectrum as the flavor weight of the partition;

[0035] According to the delay time and response threshold of real-time data transmission of the sensors in the partition, timeliness is obtained as the timeliness evaluation;

[0036] The colony binding rule corresponding to each partition in the colony binding rule table and the corresponding hierarchical strategy container are taken as the current strategy, and the conflict risk is obtained based on the resource occupation of the current strategy and the historical conflict record;

[0037] The flavor weight, timeliness evaluation and conflict risk are combined to obtain the strategy priority through priority arbitration;

[0038] Based on the strategy priority, the hierarchical strategy container is divided into different conflict interception levels;

[0039] Then, the partition coordinates, hierarchical strategy container ID and metabolic target are extracted from the colony binding rule table, and the conflict interception level is allocated according to the strategy priority, and a three-dimensional topology of space-time-functionality-performance is generated;

[0040] The three-dimensional topology is constructed into a table form to obtain a strategy container scheduling instruction set.

[0041] Further, the four-level structured gate interception is defined as: the matching degree of the functional gene of the foreign bacteria and the receptor of the indigenous bacteria is compared through the gene compatibility gate; the target metabolite synthesis bottleneck is scanned through the metabolic pathway saturation rate gate;

[0042] The gene compatibility gate information and the metabolic pathway saturation rate gate information are integrated to generate a gene-metabolism conflict detection report;

[0043] According to the gene-metabolism conflict detection report, the strain environmental adaptability is calculated through the attenuation risk value gate, and the conflict level is determined through the dynamic arbitration gate;

[0044] The way of comparing the matching degree of the functional gene of the foreign bacteria and the receptor of the indigenous bacteria through the gene compatibility gate includes:

[0045] The functional gene of the foreign bacteria and the receptor protein of the indigenous bacteria are extracted from the strategy container scheduling instruction set;

[0046] The sequence similarity of the functional gene of the foreign bacteria and the receptor protein of the indigenous bacteria is extracted;

[0047] The protein three-dimensional structure of the functional gene of the foreign bacteria and the receptor protein of the indigenous bacteria is predicted, and the structure adaptation degree of the binding site of the two is analyzed;

[0048] Combined with the metabolome flavor correlation spectrum, the functional adaptation degree of the metabolic demand of the functional gene of the foreign bacteria and the receptor protein of the indigenous bacteria is verified;

[0049] Combined with the sequence similarity, the structure adaptation degree and the functional adaptation degree, the matching degree evaluation of the functional gene is obtained;

[0050] When the matching degree evaluation of the partition does not reach the expectation, activate the gene compatibility gate alarm;

[0051] The way of scanning the target metabolite synthesis bottleneck through the metabolic pathway saturation rate gate includes:

[0052] Extract the target metabolite from the environment-metabolism correlation data set;

[0053] Further detect the target metabolite, and obtain the pathway saturation rate combined with the metabolic flux model;

[0054] If the pathway saturation rate does not reach the expectation, it is determined that there is a metabolic bottleneck in the partition, and then the bottleneck type is identified; according to the bottleneck type, the preset buffering strategy is started.

[0055] Further, the generation way of the dynamic domestication instruction set includes:

[0056] According to the gene-metabolism conflict detection report, the way of obtaining the strain environmental adaptability through the attenuation risk value gate includes:

[0057] Obtain the temperature fluctuation and pH value change of the partition, and combine the matching degree evaluation to evaluate the strain attenuation risk;

[0058] If the strain attenuation risk does not reach the expectation, trigger the high-risk early warning of the partition, and add the cooling sub-strategy;

[0059] The way of determining the conflict level through the dynamic arbitration gate includes:

[0060] According to the matching degree evaluation and the metabolic bottleneck information, the conflict level is determined, and different conflict levels are divided;

[0061] Take the conflict interception level as the core, take the conflict level as the strategy detail adjustment factor, then perform rule linkage between the conflict interception level and the conflict level, activate the corresponding responsibility transfer strategy of different partitions, and generate the responsibility transfer path synchronously;

[0062] Based on the result of the four-level gate conflict interception, generate the dynamic domestication instruction set.

[0063] Further, the way of analyzing the group response deviation by simulating hierarchical stress scenarios, and locating the deviation source combined with the transcriptomics-metabolomics tracing includes:

[0064] According to the dynamic domestication instruction set, set the environmental parameters of the simulated hierarchical stress scenario, then verify the intervention effect of the dynamic domestication instruction set on the group metabolic pathway through the simulated hierarchical stress scenario, simulate the group response deviation under environmental stress, and obtain the deviation performance;

[0065] The transcriptional dimension and the metabolic dimension of the transcriptome-metabolome are combined to trace the source of the response deviation of the flora and attribute it to a specific responsible partition, and the attribution partition is obtained;

[0066] The transcriptome and the metabolome are taken as analysis dimensions, the deviation performance of the correspondence analysis dimension, the attribution partition and the corresponding correction suggestion are integrated, a deviation source analysis report is generated, the responsibility transfer path is updated synchronously, and the hierarchical strategy container is corrected.

[0067] Further, the generation mode of the responsibility system evolution package comprises:

[0068] According to the correction effect of the hierarchical strategy container, the performance baseline of the corresponding partition in the three-dimensional responsibility library is dynamically adjusted to form a new performance baseline;

[0069] At the same time, according to the deviation source analysis report, the hierarchical strategy container parameters are optimized, and a high-conflict responsibility partition is split out;

[0070] The split-out high-conflict responsibility partition coordinates, the new performance baseline and the hierarchical strategy container parameters are integrated to generate the responsibility system evolution package, and then the three-dimensional responsibility library and the hierarchical strategy container are updated reversely.

[0071] Further, the multi-omics based intelligent domestication method of the Maotai-flavor liquor producing yeast flora comprises:

[0072] S1: Collecting multi-source data, correcting the space-time delay of sensors to generate environment-metabolism correlation data sets, and defining partition response thresholds to calculate partition environment risk assessment, and constructing a three-dimensional responsibility library;

[0073] S2: Based on the three-dimensional responsibility library, the risk state of the responsibility partition is determined by combining the symbiotic performance index, the functional strain is matched, the hierarchical strategy container is dynamically loaded, the flora binding rule table is generated, and the three-dimensional topology and the strategy container scheduling instruction set are generated;

[0074] S3: According to the strategy container scheduling instruction set, four-level structured gate interception is performed, and then a dynamic domestication instruction set is generated, and a responsibility transfer path is recorded synchronously;

[0075] S4: Based on the dynamic domestication instruction set and the responsibility transfer path, the response deviation of the flora is analyzed by simulating hierarchical stress scenes, the deviation source is located by combining transcriptome-metabolome, the performance baseline and the hierarchical strategy container parameters are dynamically updated, and a closed-loop iterative responsibility system evolution package is generated.

[0076] The technical effects and advantages of the multi-omics based intelligent domestication system and method of the Maotai-flavor liquor producing yeast flora of the present application are as follows:

[0077] The present application significantly improves the data integration efficiency and risk assessment accuracy by deploying a multi-source biosensor network to collect multi-source data in real time, constructing a three-dimensional responsibility library, and providing a high-precision decision basis for subsequent optimization of the microbial population.

[0078] Secondly, based on the symbiotic performance index to determine the risk state of the partition, the priority arbitration formula is used to calculate the priority based on the flavor weight, time limit evaluation and conflict risk calculation strategy, and the strategy priority is divided to ensure that the key task (such as acetate synthesis) is executed preferentially, and the precise matching of the microbial population and the responsibility partition is realized, the conflict management efficiency is optimized, and the cooperation and stability of the fermentation process are greatly improved.

[0079] Then, through the four-level gate (gene compatibility, metabolic bottleneck, attenuation risk, dynamic arbitration), targeted domestication instructions (such as buffer protocol, cooling strategy) are generated to accurately solve the microbial population conflict problem, repair the ethyl acetate synthesis gap, reduce the odor risk and ensure the timeliness of the task.

[0080] Finally, through the simulation of stress scenarios and transcriptomic-metabolomic tracing, the performance baseline is dynamically updated, the strategy parameters are optimized, and the high-conflict responsibility unit is split to form a closed-loop iteration, realizing the continuous evolution of the system.

[0081] The present application combines genome, transcriptome and metabolome with three-dimensional responsibility library to construct an "environment-metabolism-microbial population" dynamic model, realizes the whole-chain analysis of yeast function, solves the problems of low fermentation efficiency, poor stress resistance and rough flavor regulation in traditional processes through the whole-process design of "environment fusion → microbial population binding → gate interception → closed-loop evolution", significantly improves the survival rate, flavor stability and production efficiency, and realizes the dynamic optimization of microbial community and the system self-adaptation ability. BRIEF DESCRIPTION OF DRAWINGS

[0082] Figure 1 The present application is a schematic diagram of a multi-omics based yeast population intelligent domestication system for Maotai-flavor liquor production.

[0083] Figure 2 The present application is a flowchart of the structured gate interception unit in the multi-omics based yeast population intelligent domestication system for Maotai-flavor liquor production.

[0084] Figure 3 The present application is a schematic diagram of the multi-omics based yeast population intelligent domestication method for Maotai-flavor liquor production. DETAILED DESCRIPTION

[0085] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0086] Embodiment one

[0087] Please refer to Figure 1 and 2 The present embodiment is a multi-omics-based intelligent domestication system for wine yeast flora in Maotai-flavor liquor production, comprising:

[0088] Environment-metabolism fusion unit: Deploy a multi-source biosensor network to collect multi-source data of pit pool environment and metabolism, access historical multi-omics database to construct a multi-dimensional data framework; then correct the sensor space-time delay to generate space-time aligned environment-metabolism correlation data set; and define the response threshold of each partition to assess the partition environment risk and construct a three-dimensional responsibility library;

[0089] Flora topology binding unit: Based on the three-dimensional responsibility library, combine the yeast-lactic acid bacteria symbiotic performance index to determine the responsibility partition risk state, dynamically load the hierarchical strategy container and inject real-time parameters to generate the flora binding rule table; according to the flora binding rule table, divide the conflict interception level through priority arbitration, generate a three-dimensional topology of space-time-function-performance, and tabulate it into a strategy container scheduling instruction set;

[0090] Structured gate interception unit: According to the strategy container scheduling instruction set, execute four-level structured gate interception of gene compatibility, metabolic pathway saturation rate, attenuation risk value and dynamic arbitration, and then generate a dynamic domestication instruction set and record the responsibility transfer path synchronously;

[0091] Multi-omics closed-loop evolution unit: Based on the dynamic domestication instruction set and the responsibility transfer path, analyze the flora response deviation by simulating hierarchical stress scenarios, and combine transcription-metabolomics to trace the source of deviation; dynamically update the performance baseline and hierarchical strategy container parameters to generate a responsibility system evolution package, and update the three-dimensional responsibility library and hierarchical strategy container reversely.

[0092] The multi-source data includes physical environment data, biological metabolism data and chemical parameter data of the upper layer, middle layer and lower layer of the pit pool;

[0093] It should be noted that high-precision sensors can be buried at key coordinate points of the pit pool to collect data streams (including temperature, dissolved oxygen, pH value, metabolite concentration, etc.) in real time;

[0094] The dynamic monitoring range needs to cover different depth areas of the upper, middle and lower layers of the cellar to ensure the comprehensiveness of multi-modal data (such as physical environment parameters, biological metabolic signals, chemical reaction rates, etc.);

[0095] Specifically, physical environment data is collected by physical environment sensors, including but not limited to temperature and humidity sensors, dissolved oxygen probes, covering key coordinates of the upper, middle and lower layers of the cellar, and installed at intervals (such as one temperature sensor every 30 cm), and collecting physical environment data at regular intervals, including but not limited to cellar temperature and humidity, dissolved oxygen level;

[0096] Biological metabolic sensors (such as NADH fluorescence probes) are deployed on the upper, middle and lower layers of the cellar to monitor biological metabolic data such as biological metabolic activity, focusing on monitoring active areas of ethanol synthesis (such as the middle layer of the cellar);

[0097] Chemical parameter sensors (such as pH probes, acetic acid concentration detection modules, etc.) are used to collect chemical parameter data (such as pH value, metabolite (such as acetic acid) concentration) of each layer in the cellar;

[0098] A multi-modal data acquisition network is formed by physical environment sensors, biological metabolic sensors and chemical parameter sensors;

[0099] The collected multi-source data is subjected to data cleaning and standardization processing;

[0100] Specifically, data cleaning includes missing value filling (using linear interpolation to supplement missing data), outlier removal (identifying and removing outliers by Z-score method or IQR method) and noise filtering (using sliding average filtering for biological metabolic data (such as NADH activity)), and data standardization includes time and space reference unification (aligning time stamps to UTC time and converting coordinate systems to WGS84 coordinate system) and unit normalization (converting different sensor outputs to uniform units (such as temperature -> ℃, dissolved oxygen -> mg / L, pH -> pH value));

[0101] Access to historical multi-omics databases to extract metagenomic functional annotation library (such as CRISPR acid-resistant Clostridium acs gene) and metabolomic flavor correlation spectrum (such as ethyl acetate / ethyl lactate ratio);

[0102] The standardized multi-source data is associated and mapped with the functional gene annotation of the metagenomic functional annotation library and the metabolic pathway information of the metabolomic flavor correlation spectrum to form a multi-dimensional data fusion framework, providing data support for subsequent bacterial population binding, risk modeling;

[0103] It should be noted that this multi-dimensional data fusion framework is used to provide data support for subsequent bacterial population binding and metabolic correlation, while ensuring the consistency of global data;

[0104] The time stamp method is used to calculate the transmission delay time of the lower sensor data (the difference between the receiving time and the sending time is taken as the delay time), and the fixed delay (sampling period) and random delay (network jitter) are distinguished to provide the basis for the compensation mechanism; then the compensation coefficient is calculated according to the delay time (the delay time (such as seconds) is multiplied by the preset proportion factor (such as 0.15, 0.2) to obtain the compensation coefficient, for example, when the lower sensor delay is 4 seconds, the compensation coefficient = 0.6), and the compensation coefficient is used as the weight of the corresponding sensor to balance the space-time deviation; at the same time, the sensor sampling frequency is adjusted according to the real-time delay, so as to shorten the delay time as much as possible (such as delay > 2 seconds, the sampling time is shortened to 5 seconds);

[0105] The multi-source data collected by different sensors is unified into a standard unit, the time stamp is corrected, and the multi-source data aligned in space and time is generated;

[0106] The pit is divided into different partitions (i.e. each layer of the pit is divided into an equilateral grid (such as 30cm×30cm), each grid is taken as a partition, and an independent ID is assigned, for example, upper layer-XY01, upper layer-XY02, lower layer-XY-34), based on the historical multi-omics database, an association model of dissolved oxygen concentration and acetic acid accumulation accident in each partition of the pit is established, and the response threshold of each partition is defined (such as the dissolved oxygen in the upper layer-XY01 area <18% to trigger the response);

[0107] The exemplary association model is that if the dissolved oxygen <5%, the acetic acid exceeds the standard probability = 82%; if 5%≤dissolved oxygen<10%, the acetic acid exceeds the standard probability = 35%; if the dissolved oxygen ≥10%, the acetic acid exceeds the standard probability = 5%;

[0108] Based on the multi-source data aligned in space and time, the weight of each partition of the same type sensor and the corresponding multi-source data are calculated respectively, and then the average of all types of data and the average weight of the corresponding type sensor are weighted and fused to obtain the environmental risk score of each partition, which is used as the environmental risk assessment of each partition;

[0109] The weighted fusion method is to multiply the average of each type of data with the average weight, and then add all types of multiplication results;

[0110] A dynamic early warning mechanism can be defined, when the environmental risk assessment is greater than the preset environmental risk warning threshold, the pit environment control strategy is automatically triggered;

[0111] The time stamp, the coordinates of each partition, the multi-source data aligned in space and time, and the environmental risk assessment are integrated to generate the environmental-metabolic correlation data set aligned in space and time; at the same time, the response threshold of each partition is integrated to generate the environmental risk threshold table;

[0112] Based on the spatio-temporal alignment of multi-source data in the environmental-metabolic correlation dataset and the environmental risk threshold table, define and label the responsibility type of each partition metabolism, such as oxygenation and fermentation maintenance, acid inhibition and spoilage prevention;

[0113] Exemplary:

[0114] Oxygenation and fermentation maintenance: for the area where the dissolved oxygen is lower than the threshold value (such as lower layer-XY32), maintain yeast activity by increasing oxygen;

[0115] Acid inhibition and spoilage prevention: for the area with high risk of acetic acid accumulation (such as lower layer-XY32), inhibit spoilage by metabolic pathway regulation;

[0116] Set the performance baseline of each partition using historical metabolomics flavor correlation spectrum (such as ethyl acetate / ethyl lactate ratio, acetic acid conversion rate), for example, the acetic acid conversion rate baseline of lower layer-XY32 = 55%, temperature adaptation coefficient = 0.9 (when the temperature gradient is ≤2℃);

[0117] Specifically, the performance baseline is set by taking the data of the same responsibility type in the historical data, for example, for the "acid inhibition and spoilage prevention" partition, extract historical records (such as ethyl acetate / ethyl lactate ratio, or acetic acid conversion rate), and take the average or median of the historical data as the performance limit value;

[0118] At the same time, combined with the environmental risk assessment in the environmental-metabolic correlation dataset, label the maximum metabolic load of each partition (such as the upper limit of acetic acid conversion rate = 65%);

[0119] Specifically, the maximum metabolic load = historical average value × (environmental risk assessment × attenuation coefficient); wherein the historical average value is the average of all historical performance baseline values set in the historical data within a certain time, and the attenuation coefficient is a pre-set attenuation ratio, for example, 0.1%;

[0120] It should be noted that in the dynamic baseline adjustment logic of the maximum metabolic load, the environmental risk assessment is used as a dynamic variable to real-time correct the upper limit of the metabolic load, which can ensure the stability of the system when running in different areas;

[0121] Integrate the spatial coordinates, responsibility type, response threshold, performance baseline and maximum metabolic load of the partition to form a three-dimensional responsibility library with performance baseline;

[0122] Synchronously generate a responsibility matrix mapping diagram for visualizing the responsibility type distribution of each partition of the cellar (such as the "acid inhibition and spoilage prevention" area marked in lower layer-XY32);

[0123] According to the three-dimensional responsibility library, calculate the symbiotic performance index based on the yeast-lactic acid bacteria symbiotic model;

[0124] Specifically, an exemplary method for calculating the symbiotic performance index is:

[0125] Symbiotic performance index = (ethanol output / lactic acid intake) x temperature adaptation coefficient;

[0126] Wherein, the calculation method of temperature adaptation coefficient is: ; If the optimal temperature = 28℃, the actual temperature = 25℃, the temperature adaptation coefficient = 0.8;

[0127] Based on the three-dimensional responsibility library, each partition is set as a responsibility partition, and the performance threshold of the symbiotic performance index is set. When the symbiotic performance index is less than the performance threshold, a high risk threshold is triggered, and it is determined that the expectation is not reached, and the corresponding responsibility partition is marked as a high-risk area;

[0128] The responsibility partition and the symbiotic performance index triggering the alarm are recorded, and a high-risk area responsibility partition log (such as "upper layer-XY01, symbiotic performance index = 0.55") is generated. The high-risk responsibility partition log is sent to the processing end to remind the relevant personnel to manually process;

[0129] When the symbiotic performance index is greater than or equal to the performance threshold, it is determined that the expectation is reached, and the responsibility partition is determined to be normal;

[0130] For the responsibility partition determined to be normal, according to the responsibility type (such as "acid suppression and failure prevention") in the three-dimensional responsibility library, the functional strain (such as CRISPR acid-resistant Clostridium acs gene) is matched for each responsibility partition from the functional gene annotation in the multi-dimensional data fusion framework, and the corresponding binding strain is obtained;

[0131] At the same time, based on the metabolic pathway information and functional gene annotation in the multi-dimensional data fusion framework, the metabolic target of each responsibility type (such as "acid suppression and failure prevention" needs to achieve acetic acid conversion rate ≥ 55%) is defined;

[0132] The process is "acid suppression and failure prevention"→ activation of acs gene acetic acid→ ethanol conversion path;

[0133] Exemplary:

[0134] Acid suppression and failure prevention: reduce acidity through acs gene-mediated acetic acid→ ethanol conversion;

[0135] Oxygen increase and fermentation preservation: optimize yeast metabolic efficiency through NADH activity regulation;

[0136] According to the responsibility type and environmental risk assessment, through the preloaded loading strategy, the hierarchical strategy container (such as lower layer_acid conversion.SPM) is dynamically loaded, and the real-time parameters (such as acetic acid concentration, temperature) are extracted from the environment-metabolic correlation data set and injected into the hierarchical strategy container;

[0137] Specifically, the hierarchical strategy container is classified by the physical level of the cellar (upper, middle, lower) and the responsibility type (such as oxygenation and fermentation preservation, acid inhibition and spoilage prevention);

[0138] There are three types, upper, middle, and lower, each of which divides a hierarchical strategy container;

[0139] Exemplary division and naming: upper_oxygenation.SPM, middle_signal transmission.SPM, and lower_acid conversion.SPM;

[0140] Each hierarchical strategy container contains the metabolic regulation logic of the level (such as acid conversion path optimization, dissolved oxygen regulation algorithm) and parameter binding rules (such as acetic acid concentration threshold, temperature adaptation coefficient);

[0141] For different partitions of the same layer, the hierarchical strategy container will have slight differences, for example, for lower-XY32, the strategy of its hierarchical strategy container is: using CRISPR acid-resistant Clostridium to achieve acetic acid conversion rate ≥55%, the hierarchical strategy container is acid conversion_SPM, the goal is to inhibit spoilage and improve acetic acid conversion efficiency;

[0142] The loading condition of the loaded strategy is to select the corresponding container (such as lower_acid conversion.SPM) according to the responsibility type (such as "acid inhibition 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 version of the hierarchical strategy container (such as lower_acid conversion.SPM_enhanced);

[0143] The loading granularity of the loaded strategy is to load one hierarchical strategy container for each level, rather than to load each partition independently; only when the responsibility type or environmental risk assessment changes, load / unload the container (such as switching from upper_oxygenation.SPM to upper_oxygenation_SPM_emergency);

[0144] The process of dynamically loading hierarchical strategy containers is to obtain the responsibility type (such as "acid inhibition and spoilage prevention") of the current partition from the three-dimensional responsibility library;

[0145] According to the responsibility type, match the strategy container (such as lower_acid conversion.SPM);

[0146] If the environmental risk assessment is greater than the preset environmental risk warning threshold, load the enhanced version of the hierarchical strategy container;

[0147] Use an asynchronous loading mechanism (such as dynamic class loading technology) to asynchronously load the hierarchical strategy container file, avoiding blocking the main thread;

[0148] Inside the hierarchical strategy container, dynamically adjust the regulation logic according to the injected parameters (such as starting the acid conversion acceleration mode when the acetic acid concentration is >6mg / L);

[0149] Define the unloading condition of the hierarchical strategy container as a change in responsibility type (such as from "acid suppression and failure prevention" to "oxygen increase and fermentation preservation") or a decrease in environmental risk assessment to a safety threshold (such as <50);

[0150] Unload the useless container and release the memory;

[0151] When the hierarchical strategy container is updated (such as adding an acid conversion algorithm), load the new container by matching the version number (such as the strategy container named: lower layer_acid conversion_SPM_v2);

[0152] If the new version of the hierarchical strategy container causes an exception, automatically roll back to the old version (such as the strategy container named: lower layer_acid conversion_SPM_v1);

[0153] Integrate the partition coordinates, responsibility type, bound strain, metabolic target, and hierarchical strategy container to establish a bacterial community binding rule table;

[0154] An exemplary bacterial community binding rule table:

[0155] [partition coordinates, responsibility type, bound strain, metabolic target, hierarchical strategy container];

[0156] [upper layer-XY01, oxygen increase and fermentation preservation, Saccharomyces cerevisiae, ethanol production rate ≥70%, upper layer_oxygen increase_SPM];

[0157] [lower layer-XY32, acid suppression and failure prevention, CRISPR acid-resistant Clostridium, acetic acid conversion rate ≥55%, lower layer_acid conversion_SPM];

[0158] [middle layer-XY20, signal transmission, Lactobacillus brevis, signal transmission efficiency ≥60%, middle layer_signal_SPM];

[0159] Extract the partition flavor correlation spectrum (such as the ethyl acetate / ethyl lactate ratio) from the historical metabolome flavor correlation spectrum, calculate the flavor improvement potential (such as the ratio improvement range) of the current flavor correlation spectrum, and use it as the flavor weight of the partition;

[0160] An exemplary calculation method is:

[0161] Set a flavor improvement threshold range, such as [5%, 20%], if the ethyl acetate / ethyl lactate ratio improves by >20%, the flavor weight = 0.9; if the ethyl acetate / ethyl lactate ratio improves by 5%-20%, the flavor weight = 0.6, and if the ethyl acetate / ethyl lactate ratio improves by <5%, the flavor weight = 0.3;

[0162] Calculate the timeliness based on the delay time and response threshold of real-time data transmission of the sensors in the partition, and use it as the timeliness evaluation;

[0163] Specifically, the required response time is defined in response to the threshold value, such as 10 minutes of response when the dissolved oxygen <5%, then the ratio of the actual delay time to the required response time is calculated, and the difference between 1 and the ratio is taken as the timeliness evaluation, i.e., timeliness evaluation = 1-(delay time / required response time);

[0164] The microbial population binding rule corresponding to each partition in the microbial population binding rule table and the corresponding hierarchical strategy container are taken as the current strategy (these strategies aim to achieve specific metabolic goals (such as increasing the acetic acid conversion rate to 70%)), and based on the resource occupation of the current strategy and the historical conflict record, the conflict risk is calculated;

[0165] Specifically, it is detected whether the current strategy in the partition and other strategies in the partition share key resources, and if so, the resource conflict coefficient is increased, and if not, the resource conflict coefficient is decreased;

[0166] The specific calculation method is:

[0167] First, identify all key parameters (such as temperature, pH value, and dissolved oxygen concentration) that can affect the same resource;

[0168] Then, check whether the current strategy and other strategies make opposite requirements or need to make large adjustments within the same time period on these key parameters;

[0169] According to the degree of resource conflict, a resource conflict coefficient between 0 and 1 is assigned, for example, no conflict is 0.1 and high conflict is 0.9, or no conflict is 0.1 and high conflict is 0.5;

[0170] Extract the historical conflict frequency of the partition (such as "Lower Layer-XY32" past 30 days conflict frequency = 2 times), and then calculate the conflict risk through the historical conflict frequency and the resource conflict coefficient;

[0171] Specifically, the conflict risk = resource conflict coefficient x (historical conflict frequency / 100);

[0172] It should be noted that multiple different operations or control measures may exist simultaneously in the same partition, which is considered as other strategies;

[0173] For example: different types of metabolic activity control (such as temperature adjustment strategy and pH value adjustment strategy); different stages of strategies implemented at different time points in the same partition (such as pre-fermentation strategy and post-acid inhibition strategy); independent but mutually influenced growth control strategies in a variety of microbial symbiotic systems;

[0174] Each strategy uses certain environmental resources (such as temperature regulation, dissolved oxygen level adjustment, etc.), and when two or more strategies need to adjust the same resource, resource competition or conflict may occur;

[0175] Example: If one strategy requires increasing the dissolved oxygen level to promote the growth of a certain microorganism, and another strategy wants to reduce the dissolved oxygen level to inhibit the reproduction of another microorganism, this constitutes a resource conflict;

[0176] Further, a priority arbitration formula is constructed according to the flavor weight, time evaluation and conflict risk, and the strategy priority is calculated;

[0177] The priority arbitration formula is constructed as:

[0178] Strategy priority = proportion factor 1 × flavor weight + proportion factor 2 × time evaluation - proportion factor 3 × conflict risk;

[0179] Wherein, the proportion factor 1, the proportion factor 2 and the proportion factor 3 are preset values, for example, 0.4, 0.3 and 0.3, which are used to emphasize the importance of different values, and the sum of the three proportion factors is 1;

[0180] Based on the strategy priority, the hierarchical strategy container is divided into different conflict interception levels;

[0181] Specifically, an exemplary conflict interception level division method is:

[0182] Set a strategy priority threshold interval, assuming that the strategy priority threshold interval is 0.6-0.8;

[0183] Strategies with a strategy priority ≥ 0.8 are defined as high priority;

[0184] Strategies with a strategy priority of 0.6≤strategy priority<0.8 are defined as medium priority;

[0185] Strategies with a strategy priority <0.6 are defined as low priority;

[0186] Example:

[0187] Partition = lower layer-XY32;

[0188] Flavor weight = 0.87 (significant increase in ethyl acetate / ethyl lactate ratio);

[0189] Time evaluation = 0.9 (response time <5 minutes);

[0190] Conflict risk = 0.01 (resource conflict coefficient = 0.5, historical conflict times = 2);

[0191] Strategy priority = 0.83→ conflict interception level = high priority;

[0192] Further, extract the partition coordinates, hierarchical strategy container ID and metabolic target from the flora binding rule table, and assign conflict interception levels (high, medium, low) according to the strategy priority, generate a three-dimensional topology of space-time-functionality-performance;

[0193] The exemplary conflict interception rule is:

[0194] High priority: immediately execute the strategy container (such as acid conversion_SPM), and freeze other low-priority strategies in the same partition;

[0195] Medium priority: execute the strategy container after a certain time (such as 10 minutes), and record the conflict log (such as "lower layer-XY32 strategy delayed for 10 minutes");

[0196] Low priority: suspend the execution of the strategy container, and trigger manual review (such as "middle layer-XY20 strategy needs manual intervention");

[0197] Build a three-dimensional topology into a table form to get a strategy container scheduling instruction set, including partition coordinates, container ID, performance target, conflict interception level and strategy priority;

[0198] The four-level gate conflict interception is defined as: through the gene compatibility gate, compare the matching degree of the functional gene of the foreign bacteria with the receptor of the indigenous bacteria; through the metabolic pathway saturation rate gate, scan the bottleneck of target metabolite synthesis;

[0199] Integrate the gene compatibility gate information and the metabolic pathway saturation rate gate information to generate a gene-metabolism conflict detection report;

[0200] According to the gene-metabolism conflict detection report, calculate the strain environmental adaptability through the attenuation risk value gate, and determine the conflict level through the dynamic arbitration gate;

[0201] The way of comparing the matching degree of the functional gene of the foreign bacteria with the receptor of the indigenous bacteria through the gene compatibility gate includes:

[0202] Extract the functional gene of the binding strain (such as the acs gene of CRISPR acid-resistant Clostridium) from the strategy container scheduling instruction set as the functional gene of the foreign bacteria, and call the pit indigenous bacteria receptor protein (such as SdiA protein) in the historical multi-omics database;

[0203] Use BLAST tool to calculate the sequence similarity of the functional gene of the foreign bacteria and the receptor protein of the indigenous bacteria (such as the similarity of acs gene and SdiA protein = 58%);

[0204] Predict the protein three-dimensional structure of the exogenous bacterial functional gene and the receptor protein of the indigenous bacteria by the AlphaFold tool, and analyze the structural fitness of the binding site (such as the number of hydrogen bonds, hydrophobic interactions) between the two (such as structural fitness = 60%);

[0205] Combined with the metabolic flavor correlation spectrum, the functional fitness of the exogenous bacterial functional gene and the metabolic demand of the indigenous bacterial receptor protein is verified (such as functional fitness = 55%);

[0206] Specifically, according to the metabolic flavor correlation spectrum (such as the correlation between the ratio of ethyl acetate / ethyl lactate and flavor quality) in the historical multi-omics database, the target metabolite (such as ethyl acetate, ethyl lactate) is determined;

[0207] For example, the ratio of ethyl acetate / ethyl lactate directly affects the flavor of the fermentation product (such as the body of the wine or the acidity of the wine);

[0208] Predict the metabolic potential of the exogenous bacterial functional gene (such as whether it can synthesize / decompose ethyl acetate) through the macro-genome functional annotation library (such as the acs gene of CRISPR acid-resistant Clostridium);

[0209] After introducing the exogenous bacteria into the pit, samples are collected, and the actual concentration of the target metabolite is detected using targeted metabolomics technology (such as LC-MS);

[0210] Set the experimental group: the ratio of ethyl acetate / ethyl lactate = the actual ratio after the introduction of the exogenous bacteria;

[0211] Set the ideal value: the metabolic demand ratio of the indigenous bacterial group based on historical data (such as 1.2:1);

[0212] Take the ratio of the experimental group ratio and the ideal value as the functional fitness;

[0213] Combined with sequence similarity, structural fitness and functional fitness, the matching degree score of the functional gene is calculated by weighted fusion method, which is used as the matching degree evaluation of the functional gene;

[0214] Specifically, sequence similarity, structural fitness and functional fitness are respectively given a weight ratio, then multiplied, and the three multiplication results are added to obtain the matching degree evaluation;

[0215] If the matching degree evaluation of the partition is less than the preset matching degree threshold, it is determined that the expected value has not been reached, and the gene compatibility gate alarm is activated;

[0216] The way to scan the target metabolite synthesis bottleneck through metabolic pathway saturation rate gate includes:

[0217] Extract the target metabolite (such as ethyl acetate, glycerol, serine) from the environment-metabolism correlation data set;

[0218] Further, the target metabolite is detected by LC-MS / GC-MS (liquid chromatography-mass spectrometer / gas chromatography-mass spectrometer), and the pathway saturation rate is calculated by combining the metabolic flux model (FBA);

[0219] If the pathway saturation rate is greater than the preset saturation rate threshold (such as 85%), it is determined that the expected value has not been reached, that is, there is a metabolic bottleneck in the partition (such as GPD1 enzyme overexpression leading to glycerol diversion > 12%), and further the bottleneck type (such as substrate competition, product inhibition) is identified; according to the bottleneck type, a preset buffer strategy (such as serine buffer protocol, NADH regeneration enhancement) is started;

[0220] It should be noted that the content of the gene-metabolic conflict detection report includes gene compatibility gate information (partition coordinates, exogenous bacterial functional genes, indigenous bacterial receptor proteins, and matching degree evaluation (such as the following layer-XY32 CRISPR acid-resistant Clostridium acs gene and SdiA matching degree = 58.2 points)) and metabolic pathway saturation rate gate information (target metabolite with metabolic bottleneck, pathway saturation rate, triggered buffer strategy (such as GPD1 overexpression leading to glycerol diversion > 12%, triggering serine buffer protocol));

[0221] According to the gene-metabolic conflict detection report, the way to calculate the strain environmental fitness through the attenuation risk value gate includes:

[0222] The temperature fluctuation and pH value change of the partition are obtained, and the strain attenuation risk value is calculated by combining the matching degree evaluation, which is used as the strain attenuation risk;

[0223] A formula for calculating the strain attenuation risk value is:

[0224] ;

[0225] Wherein, represents the strain attenuation risk value, e represents the natural constant, represents the difference between the actual temperature of the partition and the optimal temperature, represents the difference between the actual pH value and the metabolic adaptation pH value; and are proportional constants (usually set to 0.3 and 0.2), and 100 is the maximum value of the matching degree evaluation;

[0226] The formula logic is that if the gene compatibility score is lower than a certain score (such as 60 points), the strain attenuation risk value will increase, triggering a higher risk warning;

[0227] It should be noted that the actual pH value and the metabolic adaptation pH value refer to the pH value of the target strain, which can be the strain of the exogenous bacteria or the strain of the indigenous bacteria according to the actual demand; in the gene compatibility detection through the gene compatibility gate, the functional gene of the exogenous bacteria (such as the acs gene of the CRISPR acid-resistant Clostridium) is compared with the metabolic demand of the indigenous bacteria, at this time, the metabolic adaptation pH value of the exogenous bacteria is usually derived from the characteristics of its gene sequence and metabolic pathway, for example: by querying the metabolic pathway of the strain through a gene annotation database (such as KEGG, UniProt) to determine the optimum pH value of the key enzyme; for example, if the key enzyme for acetic acid metabolism of the exogenous bacteria is acetyl-CoA synthetase (ACS), its optimum pH value can be 6.5-7.0;

[0228] The metabolic adaptation pH value of the indigenous bacteria can be derived from historical environmental data (such as the pH tolerance range of the pit microbial community) or the bacterial community binding rule table; for example, if the indigenous bacteria is acid-resistant Clostridium, its metabolic adaptation pH value can be 4.0-5.0;

[0229] In the calculation of the strain attenuation risk value, the metabolic adaptation pH value is used to measure the degree of deviation of the actual pH value from the optimum metabolic environment of the strain;

[0230] If the strain attenuation risk value is greater than the preset strain attenuation threshold (such as 1.0), it is determined that the strain attenuation risk does not meet the expectation, triggering the high-risk early warning of the partition (such as the lower layer-XY32, sj=1.10), and adding the cooling sub-strategy (such as the cooling rate=0.5℃ / h, the target temperature=25℃);

[0231] The way of determining the conflict level through the dynamic arbitration gate includes:

[0232] According to the matching degree evaluation and the metabolic bottleneck information, the conflict level is determined, and different conflict levels are divided;

[0233] Specifically, one division method is to set a matching degree evaluation threshold interval (such as 60-80);

[0234] If the matching degree evaluation is greater than or equal to the maximum value of the matching degree 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;

[0235] If the matching degree evaluation is greater than or equal to the minimum value of the matching degree evaluation threshold interval and less than the maximum value of the matching degree evaluation threshold interval, or the pathway saturation rate is greater than the preset saturation rate threshold and less than or equal to the maximum value of the pathway saturation rate (i.e. 100%), it is classified as conflict level 2;

[0236] If the matching degree evaluation is less than the minimum value of the matching degree evaluation threshold interval, or the pathway saturation rate is greater than the maximum value of the pathway saturation rate (i.e. 100%), it is classified as conflict level 3;

[0237] The conflict interception level is taken as the core, the conflict level is taken as the strategy detail adjustment factor, and then the conflict interception level and the conflict level are linked by rules to activate the corresponding responsibility transfer strategy of different partitions and synchronously generate the responsibility transfer path (according to the bacteria colony binding rule table, the responsibility type of the partition needing to split the responsibility unit is transferred to the adjacent partition, and the three-dimensional responsibility library is updated);

[0238] An exemplary responsibility transfer strategy is:

[0239] For the case that the conflict interception level is high priority and the strategy priority is immediate execution, if the conflict level is level 1, the responsibility transfer strategy is to directly execute the strategy container without adjustment (such as acid conversion_SPM immediate start);

[0240] If the conflict level is level 2, the responsibility transfer strategy is to execute the strategy container, but with local intervention (such as reducing the cooling sub-strategy rate to 0.8℃ / h);

[0241] If the conflict level is level 3, the responsibility transfer strategy is to urgently execute the strategy container, triggering large-scale responsibility transfer (such as splitting the responsibility unit and injecting buffer solution);

[0242] For the case that the conflict interception level is medium priority and the strategy priority is delayed for a certain time, if the conflict level is level 1, the responsibility transfer strategy is to delay the execution of the strategy container and record the log (such as "lower layer-XY32 strategy delayed for 10 minutes");

[0243] 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 (such as reducing the lactic acid conversion efficiency target to 85%);

[0244] If the conflict level is level 3, the responsibility transfer strategy is to delay the execution of the strategy container, but upgrade to high priority interception (such as triggering manual review);

[0245] For the case that the conflict interception level is low priority and the strategy priority is suspended execution, if the conflict level is level 1, the responsibility transfer strategy is to suspend the strategy container and wait for manual review (such as "middle layer-XY20 strategy needs manual intervention");

[0246] If the conflict level is level 2, the responsibility transfer strategy is to suspend the strategy container, but allow automatic compensation (such as temporarily enabling the buffer protocol);

[0247] If the conflict level is level 3, the responsibility transfer strategy is to suspend the strategy container and force manual review (such as "lower layer-XY32 strategy needs expert intervention");

[0248] It should be noted that the design logic of the responsibility transfer strategy is that for high priority conflict interception level, unconditional priority execution is performed, and the conflict level only affects the specific execution of the intervention intensity (such as whether to split the responsibility unit);

[0249] For the case of medium priority conflict interception level, delayed execution, but the conflict level determines whether it needs to be upgraded to "high priority" (such as level 3 conflict triggering manual review);

[0250] For the case of low priority conflict interception level, suspend execution, and the conflict level determines whether automatic compensation is allowed (such as level 2 conflict can temporarily enable buffer protocol);

[0251] There is no need to define rules for each combination of conflict level and interception level, only the core logic is driven by the conflict interception level, and the conflict level is used as a detail correction factor.

[0252] 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;

[0253] Based on the results of the four-level gate conflict interception, a set of dynamic domestication instructions is generated;

[0254] The dynamic domestication instruction set includes target coordinates (abnormal, conflict, and adjusted partition positions (such as lower layer-XY32)), conflict interception level, buffer protocol (dynamically selected buffer type and concentration (such as serine buffer 0.1M), high priority triggers emergency buffer protocol; medium priority triggers standard buffer protocol; low priority has no buffer protocol), cooling sub-strategy (temperature control scheme (such as rate 0.5℃ / h→target temperature 25℃), high priority triggers rapid cooling; medium priority triggers gradual cooling; low priority has no cooling strategy), responsibility transfer path (conflict responsibility transfer range (such as lower layer-XY32→lower layer-XY33), high priority triggers large-scale responsibility transfer; medium priority triggers local responsibility transfer; low priority has no transfer), metabolic bottleneck target metabolite (metabolic bottleneck substance caused by conflict (such as glycerol)), and manual review identifier (whether human intervention is needed (such as "expert intervention"), low priority triggers manual review; medium priority upgrades to manual review if the conflict level = level 3);

[0255] According to the dynamic domestication instruction set, the environmental parameters of the simulation hierarchical stress scenario are set, and then the intervention effect of the dynamic domestication instruction set on the metabolic pathway of the microbial community is verified through the simulation of the hierarchical stress scenario, and the response deviation of the microbial community under environmental stress is simulated to obtain the deviation performance;

[0256] Specifically, an example of environmental parameter setting is as follows:

[0257] pH: 3.5 (adjusted according to the buffer protocol in the dynamic acclimation instruction set, such as serine buffer can maintain the pH at about 4.0);

[0258] dissolved oxygen: 0.5% (adjusted according to the target temperature 25℃ in the cooling sub-strategy, low temperature inhibits oxygen dissolution);

[0259] metabolic bottleneck target metabolite concentration: set the initial concentration (such as 120% saturation) according to the metabolic bottleneck target metabolite (such as glycerol) in the dynamic acclimation instruction set;

[0260] Then based on the COBRA framework, input the microbial community gene-metabolic network model and dynamic acclimation instruction set parameters, combined with pH, dissolved oxygen and other environmental parameters, through the environmental stress simulator to dynamically adjust the enzyme activity and reaction rate of the metabolic pathway, simulate the intervention effect of the dynamic acclimation instruction set on the metabolic pathway of the microbial community;

[0261] In the obtained simulation results, focus on whether HSP104 overexpression and metabolic pathway inhibition occur, if they occur, it indicates that there is a microbial community response deviation, and the corresponding deviation performance is recorded simultaneously;

[0262] Among them, HSP104 overexpression refers to the overexpression of HSP104 triggered by high acid stress, and the concentration of metabolic bottleneck target metabolite is greater than the initial concentration (such as >120%), due to high acid stress (such as pH=3.5), HSP104 (heat shock protein) of acid-tolerant strains is activated, and the expression of ATF1 (anti-stress transcription factor) is inhibited;

[0263] Metabolic pathway inhibition refers to the partial blockage of the glycerol metabolic pathway due to the introduction of buffer (serine), resulting in an enlarged gap (>35%) in lactic acid synthesis;

[0264] Combined with the transcriptomic dimension and metabolomic dimension of transcript-metabolomics, trace the source of the microbial community response deviation and attribute it to specific responsibility partitions to obtain the attribution partitions;

[0265] Transcriptomic dimension traceability and positioning is to analyze HSP104 overexpression, the analysis logic is that high acid stress (such as pH=3.5) triggers HSP104 overexpression (such as >120%), at this time, through the correlation between HSP104 and ATF1 inhibition, it is shown that the expression of ATF1 is significantly inhibited, resulting in a decrease in the stress resistance of the microbial community;

[0266] The attribution is to locate the specific responsibility partition through the target point coordinates in the dynamic domestication instruction set, mark the corresponding responsibility partition as the source of HSP104 overexpression, and then adjust the strategy in the hierarchical strategy container of the responsibility partition according to the preset strategy correction suggestion (increase the expression regulation strategy of ATF1 (such as introducing ATF1 promoter enhancer));

[0267] The metabolic dimension traceability positioning is to analyze the metabolic pathway inhibition, that is, the ethyl acetate synthesis gap, and the analysis logic is to analyze the metabolic bottleneck target metabolite (glycerol saturation rate = 120% of the lower layer-XY32) in the dynamic domestication instruction set. Due to the high glycerol saturation rate, the metabolic pathway is blocked, the acetic acid synthesis gap is expanded (> 35%), the buffer protocol (serine buffer) further reduces the activity of glycerol dehydrogenase, and the metabolic bottleneck is aggravated;

[0268] The attribution is to locate the specific responsibility partition through the target point coordinates in the dynamic domestication instruction set, mark the corresponding responsibility partition as the source of ethyl acetate synthesis gap, and then adjust the strategy in the hierarchical strategy container of the responsibility partition according to the threshold value of the strategy correction suggestion (such as optimizing the buffer protocol (such as reducing the serine concentration to 0.05M), and introducing glycerol dehydrogenase activator);

[0269] Taking the transcriptome and the metabolome as the analysis dimensions, integrating the deviation performance, attribution partition and corresponding correction suggestion of the corresponding analysis dimension, generating a deviation root cause analysis report (including analysis dimension, deviation performance, attribution partition and correction suggestion), synchronously updating the responsibility transfer path and correcting the hierarchical strategy container;

[0270] According to the correction effect of the hierarchical strategy container, dynamically adjust the efficiency baseline of the corresponding partition in the three-dimensional responsibility library to form a new efficiency baseline;

[0271] The dynamic adjustment method of the new efficiency baseline is:

[0272] New efficiency baseline = original efficiency baseline × (1 + proportional factor × traceability correction factor);

[0273] Wherein, the proportional factor is a preset constant value, used to adjust the correction amplitude of the efficiency baseline, usually set to 0.2, and the traceability correction factor is defined as traceability correction factor = 1 - repair degree of deviation root cause;

[0274] The repair degree of the deviation root cause is obtained according to the correction effect of the hierarchical strategy container, for example, the ethyl acetate synthesis gap is repaired from 35% to 10%, the repair degree of the deviation root cause = 0.1, and the traceability correction factor = 1-0.1 = 0.9;

[0275] The content of the responsibility matrix update includes the partition coordinates (such as 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%);

[0276] At the same time, according to the correction suggestion in the deviation root cause analysis report, the hierarchical strategy container parameters are optimized, and high-conflict responsibility partitions are split out;

[0277] Specifically, the optimization of the hierarchical strategy container parameters is that, assuming that the correction suggestion in the deviation root cause analysis report is that the concentration of serine buffer_SPM is adjusted from 0.1M to 0.05M, the hierarchical strategy container is adjusted accordingly according to this correction suggestion;

[0278] The 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, lower layer-XY32 is split into XY32A (main control metabolic pathway) and XY32B (auxiliary buffer regulation).

[0279] The coordinates of the split-out high-conflict responsibility partitions, the new performance baseline, and the hierarchical strategy container parameters are integrated to generate an evolution package of the responsibility system, and then the three-dimensional responsibility library and the hierarchical strategy container are updated reversely.

[0280] Embodiment Two

[0281] Please refer to Figure 3 The part not described in detail in the present embodiment can be seen from the description of embodiment 1. The present embodiment provides a multi-omics-based intelligent domestication method of wine yeast flora in Maotai-flavor liquor production, which comprises:

[0282] S1: Deploy a multi-source biosensor network to collect multi-source data of pit environment and metabolism, access a historical multi-omics database to construct a multi-dimensional data framework; then correct the time-space delay of the sensor to generate a time-space aligned environment-metabolism correlation data set; and define the response threshold of each partition to perform partition environment risk assessment, thereby constructing a three-dimensional responsibility library;

[0283] S2: Based on the three-dimensional responsibility library, in combination with the yeast-lactic acid bacteria symbiotic performance index, determine the risk state of the responsibility partition, dynamically load the hierarchical strategy container and inject real-time parameters, and generate a flora binding rule table; according to the flora binding rule table, divide the conflict interception level through a priority arbitration formula, generate a three-dimensional topology of time-space-function-performance, and tabulate it into a strategy container scheduling instruction set;

[0284] S3: According to the strategy container scheduling instruction set, perform 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 synchronously recording the responsibility transfer path;

[0285] S4: Based on the dynamic domestication instruction set and the responsibility transfer path, analyze the response deviation of the bacterial community by simulating the hierarchical stress scenario, combine the transcription-metabolomics tracing to locate the root cause of the deviation, dynamically update the efficiency baseline and the hierarchical strategy container parameters, generate the responsibility system evolution package, and reversely update the three-dimensional responsibility library and the hierarchical strategy container.

[0286] Embodiment three

[0287] The embodiment discloses an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the running mode of the above-mentioned provided multi-omics based on the intelligent domestication method of the yeast bacterial community of Maotai-flavor liquor producing yeast.

[0288] Since the electronic device introduced in the embodiment is the electronic device used to implement the multi-omics based on the intelligent domestication method of the yeast bacterial community of Maotai-flavor liquor producing yeast in the embodiment, based on the multi-omics based on the intelligent domestication method of the yeast bacterial community of Maotai-flavor liquor producing yeast introduced in the embodiment, the person skilled in the art can understand the specific implementation mode of the electronic device of the embodiment and its various forms, so the implementation of the electronic device in the method of the embodiment is not introduced in detail. As long as the electronic device used to implement the multi-omics based on the intelligent domestication method of the yeast bacterial community of Maotai-flavor liquor producing yeast in the embodiment is implemented by the person skilled in the art, it belongs to the scope of protection of the present application.

[0289] The above formulas are dimensionless values calculated, and the formulas are obtained by collecting a large amount of data to simulate a formula of the nearest real situation, and the preset parameters and threshold values in the formula are set by the person skilled in the art according to the actual situation.

[0290] The above is only the preferred embodiment of the present application, and the protection scope of the present application is not limited to the above-mentioned embodiments, and any technical solution belonging to the idea of the present application belongs to the protection scope of the present application. It should be noted that for ordinary technical users in the technical field, some improvements and decorations without departing from the principle of the present application are also considered as the protection scope of the present application.

Claims

1. A multi-omics-based intelligent domestication system for a Maotai-flavor liquor-producing Saccharomyces cerevisiae bacterial population, characterized in that, The method comprises the following steps: An environment-metabolism fusion unit is used to collect multi-source data of pit environment and metabolism, access historical multi-omics databases to construct a multi-dimensional data framework, and then correct sensor space-time delay to generate an environment-metabolism correlation data set; And define the response threshold of each partition to carry out partition environment risk assessment and construct a three-dimensional responsibility library; The generation mode of the environment-metabolism correlation data set comprises: 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 of the upper, middle and lower layers of the pit; Access historical multi-omics databases to extract a macro-genome function annotation library and a metabolomics flavor correlation spectrum, and then correlate and map the multi-source data to form a multi-dimensional data fusion framework; Standardize the multi-source data collected by different sensors, correct the time stamp, and generate space-time aligned multi-source data; Divide the pit into different partitions, obtain the environment risk assessment of each partition based on the space-time aligned multi-source data, and define the response threshold of each partition based on the historical multi-omics database; Integrate the time stamp, coordinates of each partition, space-time aligned multi-source data and environment risk assessment to generate a space-time aligned environment-metabolism correlation data set; and integrate the response threshold of each partition to generate an environment risk threshold table; The construction mode of the three-dimensional responsibility library comprises: Based on the environment-metabolism correlation data set and the environment risk threshold table, define and label the responsibility type of each partition; and set the performance baseline of each partition by using the historical metabolomics flavor correlation spectrum; At the same time, combined with the environment risk assessment in the environment-metabolism correlation data set, the maximum metabolic load of each partition is labeled; Integrate the spatial coordinates, responsibility type, response threshold, performance baseline and maximum metabolic load of the partition to construct a three-dimensional responsibility library; A microbial community topology binding unit is used to determine the responsibility partition risk state based on the three-dimensional responsibility library and the symbiotic performance index, match functional strains and dynamically load a hierarchical strategy container to generate a microbial community binding rule table; and divide conflict interception levels by priority arbitration to generate a three-dimensional topology and strategy container scheduling instruction set; The generation mode of the three-dimensional topology comprises: Extract the flavor weight of each partition from the historical metabolomics flavor correlation spectrum to obtain time efficiency evaluation and conflict risk; Combined with the flavor weight, time efficiency evaluation and conflict risk, the strategy priority is obtained by priority arbitration; and based on the strategy priority, the hierarchical strategy container is divided into different conflict interception levels; From the microbial community binding rule table, the partition coordinates, hierarchical strategy container ID and metabolic target are extracted, and the conflict interception level is allocated according to the strategy priority to generate a three-dimensional topology of space-function-performance; The three-dimensional topology is constructed into a table form to obtain a strategy container scheduling instruction set; A structured gate interception unit is used to execute four-level structured gate interception according to the strategy container scheduling instruction set, generate a dynamic domestication instruction set, and record the responsibility transfer path synchronously; A multi-omics closed-loop evolution unit is used to analyze the microbial community response deviation by simulating hierarchical stress scenarios based on the dynamic domestication instruction set and the responsibility transfer path, locate the deviation source by combining transcriptomics and metabolomics, dynamically update the performance baseline and hierarchical strategy container parameters, and generate a closed-loop iterative responsibility system evolution package.

2. The multi-omics based intelligent domestication system for Jiang-flavor liquor Saccharomyces cerevisiae bacterial population of claim 1, wherein, The generation mode of the bacterial community binding rule table comprises: According to the three-dimensional responsibility library, based on the yeast-lactic acid bacteria symbiotic model, the symbiotic performance index is obtained, when the symbiotic performance index does not reach the expectation, a high-risk early warning is triggered, and a high-risk area responsibility partition log is generated; When the symbiotic performance index reaches the expectation, it is determined that the responsibility partition is normal; Based on the responsibility partition determined to be normal, the functional strain is matched from the functional gene annotation of the multi-dimensional data fusion framework, and the metabolic target is defined; According to the responsibility type and the environmental risk assessment, the hierarchical strategy container is dynamically loaded, and the real-time parameters are injected; Integrate partition coordinates, responsibility type, binding strain, metabolic target and hierarchical strategy container to establish a bacterial community binding rule table. 3.The multi-omics based intelligent domestication system of zhang-flavor liquor Saccharomyces cerevisiae population according to claim 1, characterized in that, The four-level structured gate intercept definition is: gene compatibility gate, metabolic pathway saturation rate gate, attenuation risk value gate and dynamic arbitration gate; By comparing the functional gene of the exogenous bacteria with the receptor matching degree of the indigenous bacteria through the gene compatibility gate, including: Extract the indigenous bacteria receptor protein from the strategy container scheduling instruction set, and synchronously extract the functional gene of the exogenous bacteria; Obtain the sequence similarity, structure adaptation and function adaptation of the functional gene of the exogenous bacteria and the receptor protein of the indigenous bacteria, and then evaluate the matching degree of the functional gene; when the matching degree evaluation of the partition does not reach the expectation, activate the gene compatibility gate alarm; Through the metabolic pathway saturation rate gate, the target metabolite synthesis bottleneck is scanned, including: Extract the target metabolite from the environment-metabolism correlation data set; obtain the pathway saturation rate of the target metabolite by combining the metabolic flux model; if the pathway saturation rate does not reach the expectation, it is determined that there is a metabolic bottleneck in the partition, and a preset buffer strategy is started.

4. The multi-omics based intelligent domestication system for Jiang-flavor liquor Saccharomyces cerevisiae bacterial population of claim 3, characterized in that, The generation mode of the dynamic domestication instruction set comprises: According to the gene-metabolic conflict detection report, the strain environmental adaptability is obtained through the attenuation risk value gate, including: Obtain the temperature fluctuation and pH change of the partition, combine the matching degree evaluation, and evaluate the strain attenuation risk; if the strain attenuation risk does not reach the expectation, trigger the high-risk early warning of the partition, and add a cooling sub-strategy; Through the dynamic arbitration gate, the conflict level is determined, including: According to the matching degree evaluation and the metabolic bottleneck information, different conflict levels are divided; Link the conflict interception level and the conflict level to activate the responsibility transfer strategy, and synchronously generate the responsibility transfer path; based on the results of the four-level gate conflict interception, generate the dynamic domestication instruction set.

5. The multi-omics based intelligent domestication system of zhang-flavor liquor Saccharomyces cerevisiae population according to claim 4, characterized in that, The way of analyzing the bacterial community response deviation by simulating hierarchical stress scenarios, combined with the transcription-metabolomics locating the root cause of the deviation comprises: Based on the dynamic domestication instruction set, set the simulation hierarchical stress scenario, then verify the intervention effect of the dynamic domestication instruction set on the bacterial community metabolic pathway, detect the bacterial community response deviation, and obtain the deviation performance; Combine the transcriptome dimension and the metabolome dimension of the transcription-metabolomics to locate the root cause of the bacterial community response deviation, and attribute it to a specific partition to obtain the attribution partition; Take the transcriptome and the metabolome as the analysis dimension, integrate the deviation performance of the corresponding analysis dimension, the attribution partition and the corresponding correction suggestion to generate a deviation root cause analysis report, synchronously update the responsibility transfer path and correct the hierarchical strategy container. 6.The multi-omics based intelligent domestication system of zhang-flavor liquor Saccharomyces cerevisiae population according to claim 5, characterized in that, The generation mode of the responsibility system evolution package comprises: According to the modified effect of the hierarchical strategy container, the partition performance baseline is dynamically adjusted; According to the bias source analysis report, the hierarchical strategy container parameters are optimized, and the high conflict responsibility partition is split out; Integrate the split out high conflict responsibility partition coordinates, new performance baseline and hierarchical strategy container parameters to generate responsibility system evolution package, and then update the three-dimensional responsibility library and hierarchical strategy container in reverse.

7. The method for intelligent domestication of the yeast population for liquor production of Maotai-flavor liquor based on multi-omics, which is realized based on the system for intelligent domestication of the yeast population for liquor production of Maotai-flavor liquor based on multi-omics according to any one of claims 1 to 6, characterized in that, Comprise: S1: Collect multi-source data, correct sensor space-time delay and generate environment-metabolism correlation data set; And define each partition response threshold to calculate partition environmental risk assessment and build three-dimensional responsibility library; S2: Based on the three-dimensional responsibility library, combined with the symbiotic performance index to determine the risk state of the responsibility partition, match the functional strain and dynamically load the hierarchical strategy container, generate the bacteria colony binding rule table, and generate the three-dimensional topology and strategy container scheduling instruction set; S3: According to the strategy container scheduling instruction set, execute four-level structured gate interception, and then generate dynamic domestication instruction set, and record the responsibility transfer path synchronously; S4: Based on the dynamic domestication instruction set and the responsibility transfer path, analyze the bacteria colony response deviation by simulating hierarchical stress scene, locate the bias source combined with transcriptomics and metabolomics; Dynamically update the performance baseline and hierarchical strategy container parameters to generate closed-loop iterative responsibility system evolution package.

Images