A single-objective optimization method for plant tissue culture based on artificial intelligence

Abstract

The present invention discloses a method for single-target optimization of plant tissue culture based on artificial intelligence, which relates to the field of plant tissue culture, including: building a plant tissue culture environment, collecting environmental parameters and preprocessing; using mass spectrometry and liquid chromatography to analyze the metabolomics data of plant tissues, and setting a single-target optimization culture target in combination with environmental parameters; based on the environmental parameters and metabolomics data, constructing a metabolomics analysis model through a random forest algorithm, and extracting a preliminary optimization path; according to the preliminary optimization path, using a sparse optimization algorithm to minimize plant culture resources and obtain a minimum resource solution; using the minimum resource solution to culture plant tissues, continuously collecting environmental parameters and metabolomics data, and updating the metabolomics analysis model parameters. The present invention improves the intelligence, precision and efficient resource utilization of plant tissue culture by comprehensively utilizing artificial intelligence, metabolomics and optimization algorithms.

Description

Technical Field

[0001] The present invention relates to the field of plant tissue culture, and in particular to a method for single-objective optimization of plant tissue culture based on artificial intelligence. Background Art

[0002] As an important branch of modern plant biotechnology, plant tissue culture technology has been widely used in agriculture, horticulture, drug production and other fields. Early plant tissue culture mainly relied on manual experience and experiments. Researchers manually adjusted environmental conditions such as temperature, humidity, and light to try to improve the growth rate and quality of plants. However, this empirical method has significant limitations. It cannot fully utilize the large amount of environmental parameters and metabolic data during plant growth, and the optimization of culture conditions lacks systematicity and automation. With the advancement of modern biotechnology and information technology, plant tissue culture optimization methods based on big data analysis and machine learning have gradually become a research hotspot. For example, in recent years, the application of high-throughput data such as metabolomics and genomics has enabled the growth process of plants to be accurately monitored and analyzed, providing an important theoretical basis and technical means for optimizing plant tissue culture conditions.

[0003] Despite this, existing technologies still face many challenges. On the one hand, traditional plant tissue culture techniques often rely on manual experience or simple experimental designs, and are unable to achieve intelligent and automated optimization under complex environmental conditions. On the other hand, although there are some plant growth prediction models based on metabolomics and environmental parameters, most of these models lack real-time updating capabilities and are unable to cope with complex dynamic changes. Existing optimization methods usually only focus on a single goal (such as improving plant growth rate or quality), and fail to comprehensively consider the synergistic effects of multiple factors such as environmental parameters, culture resources, and metabolites. In addition, many optimization algorithms do not take into account the needs of sparse optimization when allocating resources, resulting in inefficient resource utilization. Therefore, existing technologies have problems such as poor accuracy, low efficiency, and insufficient resource utilization in practical applications. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a method for single-objective optimization of plant tissue culture based on artificial intelligence to solve the problem of inefficient resource utilization.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In the first aspect, the present invention provides a method for single-objective optimization of plant tissue culture based on artificial intelligence, which includes establishing a plant tissue culture environment, collecting environmental parameters and preprocessing them; using mass spectrometry and liquid chromatography to analyze the metabolomics data of plant tissues, and setting a single-objective optimization culture target in combination with environmental parameters; based on the environmental parameters and metabolomics data, constructing a metabolomics analysis model through a random forest algorithm, and extracting a preliminary optimization path; based on the preliminary optimization path, using a sparse optimization algorithm to minimize plant culture resources and obtain a minimum resource solution; using the minimum resource solution to culture plant tissues, continuously collecting environmental parameters and metabolomics data, and updating the metabolomics analysis model parameters.

[0008] As a preferred solution of the method for single-objective optimization of plant tissue culture based on artificial intelligence described in the present invention, wherein: the steps of building a plant tissue culture environment, collecting environmental parameters and preprocessing them are as follows:

[0009] Select a plant incubator as the culture equipment, set up the initial environmental conditions and inoculate the plant tissue;

[0010] Install the sensor in the plant incubator and perform calibration and accuracy verification on the sensor;

[0011] The environmental parameters in the plant incubator are collected through sensors and transmitted to the data processing terminal;

[0012] In the data processing terminal, the environmental parameters are denoised and outliers are detected, and missing values ​​are processed using interpolation methods;

[0013] The processed data are normalized and standardized, and data integration is performed.

[0014] As a preferred solution of the method for single-objective optimization of plant tissue culture based on artificial intelligence described in the present invention, the environmental parameters include temperature, humidity, light intensity and carbon dioxide concentration.

[0015] As a preferred embodiment of the method for single-target optimization of plant tissue culture based on artificial intelligence described in the present invention, wherein: the metabolomics data of plant tissues are determined by mass spectrometry and liquid chromatography, and the single-target optimization cultivation target is set in combination with environmental parameters, and the specific steps are:

[0016] Samples were collected from plant tissues, sample interferences were removed by filtration and centrifugation, and metabolites in the samples were extracted into methanol solution using a cooling centrifuge;

[0017] Configure a liquid chromatography column and mobile phase according to the metabolites, inject the metabolites using an automatic injector, perform liquid chromatography analysis, and obtain separated metabolites;

[0018] The separated metabolites are sent to a mass spectrometer for mass spectrometry analysis to extract metabolomics data;

[0019] Combining environmental parameters and metabolomics data, and searching the literature for recommended temperature and humidity for plant growth, we set a single-objective optimization function for plant tissue culture, which is expressed as:

[0020] ;

[0021] in, For plant tissue culture single objective optimization function, is the length of cultivation time, is the weight coefficient of temperature for target optimization, is the natural base, is the sensitivity coefficient of temperature deviation from the recommended value, is the temperature inside the incubator, Recommended temperature for plant growth, is the weight coefficient of humidity for target optimization, is the sensitivity coefficient of humidity deviation from the recommended value, is the humidity in the incubator, Recommended humidity for plant growth, is the influence coefficient of light intensity on target optimization, is the light intensity in the incubator, is the suppression coefficient when the light intensity is too high, is the influence coefficient of carbon dioxide on target optimization, is the carbon dioxide concentration, is the inhibition coefficient when the carbon dioxide concentration is too high, is the contribution coefficient of metabolites to target optimization, is the absolute concentration of the target metabolite in the metabolomics data, is the score identifier.

[0022] As a preferred embodiment of the method for single-objective optimization of plant tissue culture based on artificial intelligence of the present invention, wherein: based on environmental parameters and metabolomics data, a metabolomics analysis model is constructed by a random forest algorithm, and the specific steps are:

[0023] Time series segmentation of environmental parameters and metabolomics data was performed to obtain sliding window datasets;

[0024] For the sliding window data set, wavelet transform is used to extract features, and feature dimension reduction is performed through principal component analysis algorithm to obtain the sliding window feature set;

[0025] Split the sliding window feature set into multiple subsets and use different subsets to train different decision trees;

[0026] For each decision tree, bootstrap sampling was used to randomly extract samples and split them using a randomly selected feature subset to obtain a metabolomics analysis model;

[0027] Through feature importance analysis, the influence of each input variable on the single objective optimization function is quantified. The expression is:

[0028] ;

[0029] in, For the The importance weights of the input variables, For the The index in the split The reduction in the Gini index of the single-objective optimization function brought about by is the total number of splits of the decision tree, is the total number of trees in the random forest, is the input variable index, is the index of the number of splits;

[0030] According to the importance weight, the input variables are sorted in descending order, and the input variables with the highest ranking are extracted as key variables.

[0031] As a preferred solution of the method for single-objective optimization of plant tissue culture based on artificial intelligence described in the present invention, wherein: the extracting of the preliminary optimization path comprises the following specific steps:

[0032] Perform sensitivity analysis on the single-objective optimization function and divide the key variables into positive variables and negative variables. The expression is:

[0033] ;

[0034] in, Key variables The sensitivity coefficient of is the key variable, Index for the number of key variables;

[0035] The key variables are combined synergistically to perform interactive sensitivity analysis, and the expression is:

[0036] ;

[0037] in, Key variables and The synergistic sensitivity coefficient, is the key variable for collaboration, is the index of the number of collaborative key variables;

[0038] According to the sensitivity coefficient and the co-sensitivity coefficient, the values ​​of key variables are adjusted to generate the optimization principle of key variables;

[0039] According to the optimization principle of key variables, combined with time series data, the preliminary optimization value of key variables is extracted, and the expression is:

[0040] ;

[0041] in, For the Preliminary optimization values ​​of key variables at the time step, In order to maximize the objective function, is a set of key variables;

[0042] The preliminary optimized values ​​of the key variables and the key variable set are combined into a preliminary optimization path.

[0043] As a preferred solution of the method for single-objective optimization of plant tissue culture based on artificial intelligence described in the present invention, wherein: according to the preliminary optimization path, a sparse optimization algorithm is used to minimize plant culture resources to obtain a minimum resource solution, the specific steps are:

[0044] Based on the preliminary optimization path, the resource consumption coefficient of each key variable is defined, and the total consumption function in the plant cultivation process is set. The expression is:

[0045] ;

[0046] in, is the total consumption function, is the number of key variables, For the The resource consumption coefficient of the key variables, is the resource consumption coefficient index;

[0047] Based on the preliminary optimization path, obtain a set of key variables and preliminary optimization values ​​of the key variables;

[0048] Based on the key variable set and the preliminary optimized values ​​of the key variables, the sparsity objective function is defined, and the LASSO sparse optimization algorithm is used to solve it iteratively to obtain the final optimized value of the key variables. The expression is:

[0049] ;

[0050] in, To get the minimum value, is the regularization strength, is the L1 norm;

[0051] The final optimized values ​​of key variables and the corresponding total consumption are sorted into the minimum resource solution.

[0052] As a preferred embodiment of the method for single-objective optimization of plant tissue culture based on artificial intelligence described in the present invention, the method comprises the following steps: using a minimum resource solution to carry out plant tissue culture, continuously collecting environmental parameters and metabolomics data, and updating metabolomics analysis model parameters.

[0053] Configure the culture environment according to the minimum resource plan and adjust the environmental parameters and metabolomics data;

[0054] During the plant tissue culture process, environmental parameters were re-collected and metabolomics data were measured regularly;

[0055] Update metabolomics analysis model parameters based on newly collected environmental parameters and metabolomics data.

[0056] In a second aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the method for single-objective optimization of plant tissue culture based on artificial intelligence as described in the first aspect of the present invention is implemented.

[0057] In a third aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, any step of the method for single-objective optimization of plant tissue culture based on artificial intelligence as described in the first aspect of the present invention is implemented.

[0058] The beneficial effects of the present invention are as follows: by precisely constructing a plant culture environment and collecting environmental parameters, the high quality and accuracy of the data are ensured, providing a reliable basis for subsequent optimization. By analyzing metabolomics data through mass spectrometry and liquid chromatography, and setting optimization targets in combination with environmental parameters, a scientific and reasonable optimization target setting is achieved. The random forest algorithm is used to extract key features from environmental parameters and metabolomics data, and automatically identify variables related to the optimization targets, thereby improving the accuracy and efficiency of data analysis. The use of a sparse optimization algorithm to minimize resource consumption not only reduces costs but also improves resource utilization, meeting the needs of sustainable development. By dynamically adjusting the minimum resource plan and continuously monitoring the environment and metabolomics data, real-time optimization and continuous improvement of the culture process are achieved, ensuring higher plant growth quality and yield. By comprehensively utilizing artificial intelligence, metabolomics and optimization algorithms, the intelligence, precision and efficient resource utilization of plant tissue culture are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0060] Figure 1 This is a flow chart of the method for single-objective optimization of plant tissue culture based on artificial intelligence in Example 1.

[0061] Figure 2 This is a flowchart for extracting the preliminary optimization path in Example 1. DETAILED DESCRIPTION

[0062] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0063] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0064] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0065] Example 1, with reference to Figure 1 and Figure 2 , which is the first embodiment of the present invention, provides a method for single-objective optimization of plant tissue culture based on artificial intelligence, comprising the following steps:

[0066] S1: Set up a plant tissue culture environment, collect environmental parameters and pre-process them.

[0067] Specifically, the following steps are included:

[0068] S1.1: Select a plant incubator as the culture equipment, set up the initial environmental conditions, and inoculate the plant tissue.

[0069] Specifically, the selected plant incubator needs to meet the following conditions:

[0070] Temperature controllable range: 10-40℃ or above, with an accuracy requirement of 0.1℃;

[0071] Humidity controllable range: 30%-90%RH or above, accuracy ±1%RH;

[0072] Adjustable lighting conditions: light intensity range 0-1000 or above, supports photoperiod control (e.g. 16h light / 8h dark);

[0073] Carbon dioxide concentration control: 0-5000ppm, accuracy ±50ppm;

[0074] The equipment must have good sealing and uniform airflow distribution.

[0075] According to the cultivation requirements of the target plant tissue, set the initial environmental parameters (such as temperature, humidity, light intensity, and carbon dioxide concentration):

[0076] Temperature: Set the appropriate cultivation temperature based on the plant type (such as tropical plants or temperate plants). For example, tropical plants should be 25-28°C, and temperate plants should be 20-25°C.

[0077] Humidity: Set to 60%-80% RH to keep the environment moist and reduce evaporation;

[0078] Light conditions: Set the light cycle to 16h light / 8h dark and adjust the light intensity to 200-400 ;

[0079] Carbon dioxide concentration: set to 400-600ppm.

[0080] Furthermore, under a sterile operating hood, plant tissue (such as buds, callus, or embryos) is inoculated into the prepared culture medium. After inoculation, the culture container (such as a culture flask or dish) is placed in an incubator to ensure that the sample is evenly distributed and to avoid uneven light or airflow.

[0081] S1.2: Install the sensor in the plant incubator and perform calibration and accuracy verification on the sensor.

[0082] Specifically, the following sensors are installed in the incubator to monitor environmental parameters in real time:

[0083] Temperature sensors: installed at different heights inside the incubator to ensure uniform temperature distribution;

[0084] Humidity sensor: Place in the center of the incubator to avoid being affected by airflow or condensation;

[0085] Light intensity sensor: installed above the sample placement area to directly measure the light intensity received by the sample;

[0086] Carbon dioxide sensor: Placed in the incubator's airflow circulation path to measure uniform carbon dioxide concentration.

[0087] The temperature sensor was calibrated at multiple temperature points using a platinum resistance thermometer; the humidity sensor was calibrated using a 75%RH saturated sodium chloride solution; and the humidity sensor was calibrated using a standard light intensity meter at 200, 400, and 600°C. The light intensity sensor was calibrated under light intensity; the carbon dioxide sensor was calibrated using 400ppm carbon dioxide gas.

[0088] During the incubator operation, compare the sensor readings with the standard instrument measurement data to ensure that the error is within the allowable range.

[0089] S1.3: Collect environmental parameters in the plant incubator through sensors and transmit them to the data processing terminal.

[0090] Specifically, the acquisition frequency is set to once per minute and stored as time series data. A wireless connection is used to transmit the sensor data to a data processing terminal (such as a computer or cloud server).

[0091] Furthermore, environmental parameters refer to:

[0092] S1.3.1: Environmental parameters include temperature, humidity, light intensity, and carbon dioxide concentration.

[0093] S1.4: In the data processing terminal, perform denoising and outlier detection on environmental parameters, and use interpolation to handle missing values.

[0094] It should be understood that denoising refers to removing noise from sensor data using a sliding average filtering algorithm (taking the average of five consecutive time points as the smoothed value at the current time point).

[0095] Outlier detection refers to the use of statistical methods to detect outliers, such as converting data into a standard distribution and marking data points that deviate from the mean by more than 3 standard deviations as outliers.

[0096] Missing value processing refers to the use of linear interpolation to fill in missing data caused by equipment failure or transmission problems during the collection process.

[0097] S1.5: Normalize and standardize the processed data and perform data integration.

[0098] Specifically, normalization refers to converting data of different dimensions (such as temperature and humidity) to the same range (such as [0, 1]) using the Min-Max normalization algorithm.

[0099] Standardization refers to converting data into a distribution with a mean of 0 and a standard deviation of 1, eliminating the dimensionality of the data.

[0100] All processed environmental parameters (temperature, humidity, light intensity, and carbon dioxide concentration) are integrated into a unified time series data set and stored in a CSV data table.

[0101] Optimally, the consistency and stability of environmental conditions are ensured by precisely constructing and controlling the plant tissue culture environment and installing high-precision sensors to monitor environmental parameters such as temperature, humidity, light intensity, and carbon dioxide concentration in real time. Data quality and integrity are guaranteed through data acquisition, denoising, outlier detection, and missing value interpolation, while normalization, standardization, and data integration provide a unified, standardized data set for subsequent analysis and optimization. This series of refined steps provides a reliable data foundation for environmental control and optimization decisions during plant tissue culture, improving the controllability and repeatability of the process, thereby increasing culture efficiency, reducing resource waste, and ensuring the predictability and accuracy of results.

[0102] S2: Use mass spectrometry and liquid chromatography to analyze the metabolomics data of plant tissues and set single-target optimization cultivation targets in combination with environmental parameters.

[0103] Specifically, the following steps are included:

[0104] S2.1: Collect samples from plant tissues, remove sample interferences by filtration and centrifugation, and extract metabolites from the samples into methanol solution using a cooled centrifuge.

[0105] Specifically, fresh samples are obtained from plant tissues (such as leaves, roots, stems, etc.). Appropriate plant parts should be selected for sample collection to avoid damage and contamination. Aseptic operation should be used to prevent interference from exogenous microorganisms. After collection, the samples are filtered to remove larger particulate impurities to ensure that the metabolites in the samples are effectively extracted. Then, centrifugation is performed in a cooled centrifuge to remove most of the cell residues and obtain a relatively pure metabolite sample. Metabolites are extracted from the cells by adding methanol solution to the sample and using a cooled centrifuge. The methanol solution should be kept at a low temperature to prevent metabolite degradation. The operating temperature is -20°C, the centrifugation rate is 3000-5000rpm, and the centrifugation time is 5-10 minutes.

[0106] S2.2: Configure the liquid chromatography column and mobile phase according to the metabolites, inject the metabolites using an automatic injector, perform liquid chromatography analysis, and obtain the separated metabolites.

[0107] Specifically, select an appropriate chromatographic column based on the properties of the metabolites. These include C18 reversed-phase columns (suitable for separating most polar and non-polar compounds) and HILIC columns (suitable for polar compounds). The mobile phase should be optimized based on the polarity of the metabolites being analyzed, including water / methanol / acetonitrile mixtures.

[0108] Prepare an injection volume of 10-50 μL and inject it into the chromatographic column via an autosampler. The injection method should be linear isocratic elution or gradient elution to separate different metabolites.

[0109] Set the flow rate to 0.2-1.0 mL / min and the temperature to around 30°C. Gradient elution conditions (e.g., starting with 0% acetonitrile and increasing to 80% acetonitrile after 5 minutes) can be optimized based on the characteristics of the metabolites to be analyzed.

[0110] S2.3: Send the separated metabolites to the mass spectrometer for mass spectrometry analysis and extract metabolomics data.

[0111] Specifically, a high-resolution mass spectrometer (such as a Q-TOF, Orbitrap, or ion trap mass spectrometer) is selected. Mass spectrometry analysis typically uses electrospray ionization (ESI) or atmospheric pressure chemical ionization (APCI) mode, selecting an appropriate ion source based on the ionization characteristics of the metabolites.

[0112] ESI source settings should be optimized based on the characteristics of the target metabolites, including spray voltage (set to 3-5 kV), inlet gas flow (30-40 L / min), and auxiliary gas flow (10-15 L / min). Mass spectrometry analysis should scan the full mass spectral range (usually from m / z 50 to 1000).

[0113] Analyze in both positive and negative ion modes to cover as many metabolites as possible. The mass spectrometer should be used in scan mode, recording the mass-to-charge ratio (m / z) and relative intensity (intensity) data of each ion to obtain the mass spectrum of the metabolites.

[0114] In a mass spectrum, each metabolite's ion peak has a different peak area or height, which is proportional to the metabolite's concentration. By comparing the peak areas or heights of different samples or under different conditions, the relative concentrations of each metabolite can be estimated. Furthermore, by using the internal standard method, the peak area ratio of the target metabolite to the internal standard can be compared to quantitatively analyze the metabolites and determine their absolute concentrations.

[0115] S2.4: Combine environmental parameters and metabolomics data, and search the literature for recommended temperature and humidity for plant growth to establish a single-objective optimization function for plant tissue culture. The expression is:

[0116] ;

[0117] in, For plant tissue culture single objective optimization function, is the length of cultivation time, is the weight coefficient of temperature for target optimization, is the natural base, is the sensitivity coefficient of temperature deviation from the recommended value, is the temperature inside the incubator, Recommended temperature for plant growth, is the weight coefficient of humidity for target optimization, is the sensitivity coefficient of humidity deviation from the recommended value, is the humidity in the incubator, Recommended humidity for plant growth, is the influence coefficient of light intensity on target optimization, is the light intensity in the incubator, is the suppression coefficient when the light intensity is too high, is the influence coefficient of carbon dioxide on target optimization, is the carbon dioxide concentration, is the inhibition coefficient when the carbon dioxide concentration is too high, is the contribution coefficient of metabolites to target optimization, is the absolute concentration of the target metabolite in the metabolomics data, is the score identifier.

[0118] Preferably, a single-objective optimization function is established by combining environmental parameters (such as temperature, humidity, light intensity, and carbon dioxide concentration) with the concentration of the target metabolite from metabolomics data. This function can describe the relationship between environmental parameters and metabolite production through a mathematical model. Using a polynomial or exponential decay function in conjunction with the environmental parameters, optimal environmental control conditions are determined. Metabolomics data can then be monitored to optimize the production of the target metabolite. This optimization function can directly guide the regulation of the plant tissue culture environment, and the optimization path can be verified experimentally to ultimately maximize the accumulation of the target metabolite during the culture process.

[0119] Ideally, refined plant tissue culture and metabolite analysis processes, combined with liquid chromatography and mass spectrometry, can efficiently extract, separate, and quantify metabolites in plant tissues, ensuring the accuracy of metabolite data. By optimizing multiple parameters of the culture environment (such as temperature, humidity, light, and carbon dioxide concentration) and combining them with metabolomics data, mathematical modeling can be used to set optimization functions, enabling precise control of environmental conditions to maximize the yield of target metabolites. This approach not only improves the accuracy of metabolite extraction and analysis but also continuously optimizes the culture environment through experimental verification and data-driven approaches, promoting the efficient production and application of plant metabolites.

[0120] S3: Based on environmental parameters and metabolomics data, a metabolomics analysis model was constructed using the random forest algorithm to extract preliminary optimization pathways.

[0121] Specifically, the following steps are included:

[0122] S3.1: Perform time series segmentation on environmental parameters and metabolomics data to obtain sliding window datasets.

[0123] It should be understood that the time series data is segmented using a sliding window, where each window contains a fixed-length time step. The data in each window is considered a sample, containing environmental parameters and corresponding metabolomics data.

[0124] S3.2: For the sliding window data set, wavelet transform is used to extract features, and feature dimension reduction is performed through the principal component analysis algorithm to obtain the sliding window feature set.

[0125] Specifically, a wavelet transform was performed on the environmental parameters and metabolomics data within each window to extract time-domain and frequency-domain features. The wavelet basis function was the Haar wavelet, which was decomposed into different frequency components (low-frequency and high-frequency) to extract features such as energy distribution and signal trends.

[0126] The time-frequency features extracted from each window are summarized to form a time-frequency feature vector. The time-frequency feature vectors of all windows are subjected to principal component analysis, and the principal components with a cumulative contribution rate of 95% are retained to generate a sliding window feature set after dimensionality reduction.

[0127] S3.3: Based on the sliding window feature set, a metabolomics analysis model is constructed using the random forest algorithm. The influence of each input variable on the single-objective optimization function is quantified through feature importance analysis. The expression is:

[0128] ;

[0129] in, For the The importance weights of the input variables, For the The index in the split The reduction in the Gini index of the single-objective optimization function brought about by is the total number of splits of the decision tree, is the total number of trees in the random forest, is the input variable index, The index of the number of splits.

[0130] Preferably, the training dataset is split into multiple subsets, and different decision trees are trained using different subsets. For each decision tree, bootstrap sampling is used to randomly extract samples from the dataset and split using a randomly selected subset of features. Each decision tree is trained based on the input features and output targets, forming the splitting nodes of the decision tree. To prevent overfitting, a maximum depth is set for the decision tree to control tree growth and avoid excessive complexity.

[0131] Furthermore, after random forest training is complete, feature importance analysis is used to quantify the contribution of each input variable to the optimization function. This is achieved by calculating the change in the Gini index caused by each input variable when the decision tree splits. The Gini index is a metric used to measure the purity of the categories in a dataset; smaller values ​​indicate purer data.

[0132] S3.4: Sort the input variables in descending order according to their importance weights, and extract the input variables with the highest ranking as key variables.

[0133] Specifically, by calculating the All input variables are sorted in descending order, and the top 30% of the input variables are selected to obtain the key environmental parameters or metabolite indicators that have the greatest impact on the single-objective optimization function.

[0134] Furthermore, extracting the preliminary optimization path includes the following steps:

[0135] S3.4.1: Perform sensitivity analysis on the single-objective optimization function, dividing the key variables into positive variables and negative variables. The expression is:

[0136] ;

[0137] in, Key variables The sensitivity coefficient of is the key variable, Index of the number of key variables.

[0138] Specifically, when When , the key variables are divided into positive variables, and increasing the variable value will improve the single-objective optimization function; when When , the key variables are classified as negative variables, and increasing the variable value will reduce the single objective optimization function.

[0139] S3.4.2: Combine key variables together to perform interactive sensitivity analysis, the expression is,

[0140] ;

[0141] in, Key variables and The synergistic sensitivity coefficient, is the key variable for collaboration, is the index of the number of collaborative key variables.

[0142] Specifically, when When , the synergy of variables has a positive impact on the objective function; when When , the synergy of variables has a negative impact on the objective function.

[0143] S3.4.3: Adjust the values ​​of key variables based on the sensitivity coefficients and co-sensitivity coefficients and generate optimization principles for key variables.

[0144] Specifically, the forward variable optimization principle is: if , then increase the variable value appropriately until it reaches the upper limit of the value range.

[0145] The principle of negative variable optimization is: if , then appropriately reduce the variable value until it reaches the lower limit of the value range.

[0146] The principle of collaborative variable optimization is: if , then increase and If the value of , then reduce and value.

[0147] S3.4.4: Based on the optimization principle of key variables and combined with time series data, extract the preliminary optimized values ​​of key variables, which are expressed as:

[0148] ;

[0149] in, For the Preliminary optimization values ​​of key variables at the time step, In order to maximize the objective function, A set of key variables.

[0150] Preferably, at each time step, the values ​​of the key variables are adjusted according to the optimization principle of the key variables to maximize the single-objective optimization function, and the preliminary optimized values ​​of the key variables at each time step are output.

[0151] S3.4.5: Combine the preliminary optimized values ​​of key variables and the set of key variables into a preliminary optimization path.

[0152] Specifically, the key variables at each time step are initially optimized The combination forms a continuous preliminary optimization path. The preliminary optimization path includes the time series optimization value of each key variable and the change trend of the single objective optimization function over time.

[0153] Optimally, by building and optimizing a random forest algorithm model based on environmental parameters and metabolomics data, using sliding window technology to segment time series data and extract wavelet transform features, combined with principal component analysis for dimensionality reduction, key trends and features in the data can be effectively captured. Through random forest modeling and feature importance analysis, key variables influencing the objective optimization function can be identified, and optimization strategies for these key variables can be developed through sensitivity analysis and co-sensitivity analysis. By dynamically adjusting the values ​​of key variables to generate a continuous optimization path, the single-objective optimization function is maximized at each time step, providing a precise and actionable optimization solution.

[0154] S4: Based on the preliminary optimization path, a sparse optimization algorithm is used to minimize plant cultivation resources and obtain the minimum resource solution.

[0155] Specifically, the following steps are included:

[0156] S4.1: Based on the preliminary optimization path, define the resource consumption coefficient of each key variable and set the total consumption function during the plant cultivation process. The expression is:

[0157] ;

[0158] in, is the total consumption function, is the number of key variables, For the The resource consumption coefficient of the key variables, It is the resource consumption coefficient index.

[0159] Specifically, different key variables consume different types of resources. Set resource consumption coefficient , as follows:

[0160] Temperature regulation: Power consumption coefficient , the unit is power (kWh / ℃).

[0161] Humidity control: water resources or power consumption of humidification equipment , the unit is water volume or electricity volume (L / %RH or kWh / %RH).

[0162] Light intensity adjustment: light source power consumption , the unit is power (kWh / μmol / m² / s).

[0163] CO2 concentration regulation: CO2 supply costs , the unit is gas cost (￥ / ppm).

[0164] Use the device technical specifications or policies to obtain resource consumption coefficients, for example:

[0165] Temperature regulation: 0.5kWh is consumed for every 1°C increase in temperature;

[0166] Light source adjustment: Every increase of 100μmol / m² / s consumes 0.8kWh;

[0167] Humidity adjustment: Every 10% increase in RH consumes 1L of water.

[0168] S4.2: Based on the preliminary optimization path, obtain a set of key variables and preliminary optimized values ​​of the key variables.

[0169] S4.3: Based on the set of key variables and the preliminary optimized values ​​of the key variables, define the sparsity objective function and use the LASSO sparse optimization algorithm to solve it iteratively to obtain the final optimized value of the key variables. The expression is:

[0170] ;

[0171] in, To get the minimum value, is the regularization strength, is the L1 norm, which represents the sum of the absolute values ​​of all key variable values ​​in the key variable set.

[0172] Preferably, on the basis of the preliminary optimization path, further adjust the key variable values , in order to minimize the total resource consumption while ensuring that the concentration of metabolites reaches the target value. Use the key variables to preliminarily optimize the value As a starting point, set the regularization strength to ,According to the iterative rules of the LASSO algorithm, the variable values ​​are gradually adjusted to converge the sparsity objective function.,The variable range restriction is applied to ensure that the metabolomics target value (such as,metabolite concentration) meets the optimization requirements.

[0173] S4.4: Arrange the final optimized values ​​of key variables and the corresponding total consumption into the minimum resource solution.

[0174] Specifically, the final optimized value of the key variable is output through the sparsity objective function, and the final optimized value is substituted into the total consumption function to calculate the total resource consumption.

[0175] Substitute the final optimized value into the single-objective optimization function for verification to ensure that the target meets the requirements of the preliminary optimization path. If the target is not achieved, adjust the regularization strength and re-optimize.

[0176] The final optimized values ​​of key variables and the corresponding total consumption are integrated and stored in JSON format.

[0177] Optimally, by applying a sparse optimization algorithm in combination with a preliminary optimization path, resource consumption during plant cultivation was effectively reduced while ensuring that the metabolite concentration reached the target. Based on the resource consumption coefficients of each key variable, a total consumption function was constructed, which comprehensively considered the resource consumption of factors such as temperature, humidity, light, and carbon dioxide. The LASSO sparse optimization algorithm was used to iteratively optimize the values ​​of key variables to minimize resource consumption while ensuring that the target metabolite concentration was not affected. By combining the optimization results with resource consumption, a solution that meets the optimization goals and minimizes resource consumption was output, providing an efficient and sustainable resource management strategy for plant tissue culture.

[0178] S5: Use minimal resource protocols for plant tissue culture, continuously collect environmental parameters and metabolomics data, and update metabolomics analysis model parameters.

[0179] Specifically, the following steps are included:

[0180] S5.1: Configure the culture environment according to the minimum resource plan and adjust the environmental parameters and metabolomics data.

[0181] Specifically, according to the minimum resource plan, the environmental parameters in the plant incubator are adjusted, the temperature control of the incubator is adjusted to the optimal temperature value, the humidity is set to the optimal humidity value through a humidity adjustment device (such as a humidifier or dehumidification module), the light source (such as an LED lamp) is set to the optimal light, and the photoperiod is configured, and the carbon dioxide concentration is adjusted to the optimal level through the carbon dioxide supply.

[0182] S5.2: During the plant tissue culture process, resample environmental parameters and measure metabolomics data regularly.

[0183] Specifically, the configured sensors were used to continuously collect environmental parameters at a frequency of once per minute, and the data was stored in CSV format. Metabolomics data were re-measured every 10 days.

[0184] During this period, check the sample status in the incubator regularly to avoid contamination or excessive consumption of culture medium.

[0185] S5.3: Update metabolomics analysis model hyperparameters based on newly acquired environmental parameters and metabolomics data.

[0186] Specifically, the metabolomics analysis model is continuously iterated based on the newly collected environmental parameters and metabolomics data, and the metabolomics analysis model hyperparameters are continuously updated during the iteration process to improve the metabolomics analysis model's ability to describe the relationship between metabolomics characteristics and environmental parameters.

[0187] Optimally, the minimum resource solution provides a scientific basis for adjusting environmental parameters, ensuring that the culture environment is maintained at an optimal state, thereby improving culture efficiency and reducing resource waste. Regularly collecting environmental parameters and metabolomics data, and promptly updating metabolomics analysis models, can continuously optimize parameters and maintain an accurate description of environmental and metabolic characteristics. This continuous optimization helps to dynamically adjust culture conditions, maintain stable concentrations of target metabolites, and further improve yield and quality. Model adjustments based on real-time data feedback can effectively respond to changes that may occur during the culture process, ensuring the efficiency and controllability of the entire process.

[0188] This embodiment also provides a computer device suitable for the method of performing single-objective optimization of plant tissue culture based on artificial intelligence, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the method of performing single-objective optimization of plant tissue culture based on artificial intelligence proposed in the above embodiment.

[0189] The computer device may be a terminal, comprising a processor, memory, a communication interface, a display, and an input device connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores an operating system and computer programs. The internal memory provides an environment for the operating system and computer programs stored in the non-volatile storage media. The communication interface of the computer device is used to communicate with external terminals via wired or wireless communication. Wireless communication may be achieved via Wi-Fi, a carrier network, NFC (near-field communication), or other technologies. The display of the computer device may be a liquid crystal display or an electronic ink display. The input device may be a touchscreen overlay on the display, buttons, a trackball, or a touchpad on the computer device housing, or an external keyboard, touchpad, or mouse.

[0190] This embodiment also provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for implementing single-objective optimization of plant tissue culture based on artificial intelligence as proposed in the above embodiment; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0191] In summary, the present invention ensures the high quality and accuracy of the data by: accurately building a plant culture environment and collecting environmental parameters, providing a reliable basis for subsequent optimization. By analyzing metabolomics data through mass spectrometry and liquid chromatography, and setting optimization targets in combination with environmental parameters, a scientific and reasonable optimization target setting is achieved. The random forest algorithm is used to extract key features from environmental parameters and metabolomics data, and automatically identify variables related to the optimization target, thereby improving the accuracy and efficiency of data analysis. The use of a sparse optimization algorithm to minimize resource consumption not only reduces costs, but also improves resource utilization, meeting the needs of sustainable development. By dynamically adjusting the minimum resource plan and continuously monitoring the environment and metabolomics data, real-time optimization and continuous improvement of the cultivation process are achieved, ensuring higher plant growth quality and yield. By comprehensively utilizing artificial intelligence, metabolomics and optimization algorithms, the intelligence, precision and efficient use of resources of plant tissue culture are improved.

[0192] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for single-objective optimization of plant tissue culture based on artificial intelligence, characterized in that: include, Set up plant tissue culture environment, collect environmental parameters and pre-process; Use mass spectrometry and liquid chromatography to analyze the metabolomics data of plant tissues and combine them with environmental parameters to set single-target optimization cultivation goals; Based on environmental parameters and metabolomics data, a metabolomics analysis model was constructed using the random forest algorithm to extract the preliminary optimization path. The specific steps are as follows: Perform sensitivity analysis on the single-objective optimization function and divide the key variables into positive variables and negative variables. The expression is: ; in, Key variables The sensitivity coefficient, is the key variable, Index for the number of key variables; The key variables are combined synergistically to perform interactive sensitivity analysis, and the expression is: ; in, Key variables and The synergistic sensitivity coefficient, is the key variable for collaboration, is the index of the number of collaborative key variables; According to the sensitivity coefficient and the co-sensitivity coefficient, the values ​​of key variables are adjusted to generate the optimization principle of key variables; According to the optimization principle of key variables, combined with time series data, the preliminary optimization value of key variables is extracted, and the expression is: ； in, For the Preliminary optimization values ​​of key variables at the time step, In order to maximize the objective function, is a set of key variables; Combining the preliminary optimization values ​​of key variables and the key variable set into a preliminary optimization path; Based on the preliminary optimization path, a sparse optimization algorithm is used to minimize plant cultivation resources and obtain the minimum resource solution; Use minimal resource protocols for plant tissue culture, continuously collect environmental parameters and metabolomics data, and update metabolomics analysis model parameters.

2. The method for single-objective optimization of plant tissue culture based on artificial intelligence according to claim 1, wherein: The plant tissue culture environment is constructed, and environmental parameters are collected and pre-processed. The specific steps are: Select a plant incubator as the culture equipment, set up the initial environmental conditions and inoculate the plant tissue; Install the sensor in the plant incubator and perform calibration and accuracy verification on the sensor; The environmental parameters in the plant incubator are collected through sensors and transmitted to the data processing terminal; In the data processing terminal, the environmental parameters are denoised and outliers are detected, and missing values ​​are processed using interpolation methods; The processed data are normalized and standardized, and data integration is performed.

3. The method for single-objective optimization of plant tissue culture based on artificial intelligence according to claim 2, wherein: The environmental parameters include temperature, humidity, light intensity and carbon dioxide concentration.

4. The method for single-objective optimization of plant tissue culture based on artificial intelligence according to claim 3, wherein: The method uses mass spectrometry and liquid chromatography to analyze the metabolomics data of plant tissues, and combines environmental parameters to set a single-target optimization cultivation target. The specific steps are: Samples were collected from plant tissues, sample interferences were removed by filtration and centrifugation, and metabolites in the samples were extracted into methanol solution using a cooling centrifuge; Configure a liquid chromatography column and mobile phase according to the metabolites, inject the metabolites using an automatic injector, perform liquid chromatography analysis, and obtain separated metabolites; The separated metabolites are sent to a mass spectrometer for mass spectrometry analysis to extract metabolomics data; Combining environmental parameters and metabolomics data, and searching the literature for recommended temperature and humidity for plant growth, we set a single-objective optimization function for plant tissue culture, which is expressed as: ； in, For plant tissue culture single objective optimization function, is the length of cultivation time, is the weight coefficient of temperature for target optimization, is the natural base, is the sensitivity coefficient of temperature deviation from the recommended value, is the temperature inside the incubator, Recommended temperature for plant growth, is the weight coefficient of humidity for target optimization, is the sensitivity coefficient of humidity deviation from the recommended value, is the humidity in the incubator, Recommended humidity for plant growth, is the influence coefficient of light intensity on target optimization, is the light intensity in the incubator, is the suppression coefficient when the light intensity is too high, is the influence coefficient of carbon dioxide on target optimization, is the carbon dioxide concentration, is the inhibition coefficient when the carbon dioxide concentration is too high, is the contribution coefficient of metabolites to target optimization, is the absolute concentration of the target metabolite in the metabolomics data, is the score identifier.

5. The method for single-objective optimization of plant tissue culture based on artificial intelligence according to claim 4, characterized in that: The metabolomics analysis model is constructed based on environmental parameters and metabolomics data by using a random forest algorithm. The specific steps are: Time series segmentation of environmental parameters and metabolomics data was performed to obtain sliding window datasets; For the sliding window data set, wavelet transform is used to extract features, and feature dimension reduction is performed through principal component analysis algorithm to obtain the sliding window feature set; Split the sliding window feature set into multiple subsets and use different subsets to train different decision trees; For each decision tree, bootstrap sampling was used to randomly extract samples and split them using a randomly selected feature subset to obtain a metabolomics analysis model; Through feature importance analysis, the influence of each input variable on the single-objective optimization function is quantified. The expression is: ； in, For the The importance weights of the input variables, For the The index in the split The reduction in the Gini index of the single-objective optimization function brought about by is the total number of splits of the decision tree, is the total number of trees in the random forest, is the input variable index, is the index of the number of splits; According to the importance weight, the input variables are sorted in descending order, and the input variables with the highest ranking are extracted as key variables.

6. The method for single-objective optimization of plant tissue culture based on artificial intelligence according to claim 1, wherein: According to the preliminary optimization path, the sparse optimization algorithm is used to minimize plant cultivation resources and obtain the minimum resource solution. The specific steps are: Based on the preliminary optimization path, the resource consumption coefficient of each key variable is defined, and the total consumption function in the plant cultivation process is set. The expression is: ； in, is the total consumption function, is the number of key variables, For the The resource consumption coefficient of the key variables, is the resource consumption coefficient index; Based on the preliminary optimization path, obtain a set of key variables and preliminary optimization values ​​of the key variables; Based on the key variable set and the preliminary optimized values ​​of the key variables, the sparsity objective function is defined, and the LASSO sparse optimization algorithm is used to solve it iteratively to obtain the final optimized value of the key variables. The expression is: ； in, To get the minimum value, is the regularization strength, is the L1 norm; The final optimized values ​​of key variables and the corresponding total consumption are sorted into the minimum resource solution.

7. The method for single-objective optimization of plant tissue culture based on artificial intelligence according to claim 6, characterized in that: Use the minimum resource solution to culture plant tissues, continuously collect environmental parameters and metabolomics data, and update metabolomics analysis model parameters. The specific steps are: Configure the culture environment according to the minimum resource plan and adjust the environmental parameters and metabolomics data; During the plant tissue culture process, environmental parameters were re-collected and metabolomics data were measured regularly; Update metabolomics analysis model parameters based on newly collected environmental parameters and metabolomics data.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method for single-objective optimization of plant tissue culture based on artificial intelligence according to any one of claims 1 to 7 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for single-objective optimization of plant tissue culture based on artificial intelligence according to any one of claims 1 to 7 are implemented.

Images