Prediction method and system for utilization rate of amino acid in multi-stage feed of laying hens

Abstract

The invention provides a method and system for predicting the utilization rate of feed amino acid in multiple stages of laying hens, and relates to the field of bioinformatics, and the method comprises the following steps: obtaining chemical component measured values of feed raw material samples in different growth stages of the laying hens and in-vivo measured values of standard ileum amino acid digestibility; performing predictive factor screening according to the chemical component measured value and the standard ileum amino acid digestibility in-vivo measured value to obtain a predictive factor combination; constructing a prediction model according to the prediction factor combination to obtain a candidate prediction equation; optimizing according to the candidate prediction equation, and screening to obtain an optimized equation; performing model verification according to the optimization equation to obtain a target prediction model; and performing digestibility prediction according to the target prediction model to obtain a predicted standard ileum amino acid digestibility value. According to the method, the standard ileum amino acid digestibility is accurately predicted based on in-vitro detection data, and the limitation of a traditional in-vivo determination method is effectively overcome.

Description

Technical Field

[0001] This invention belongs to the field of bioinformatics, and specifically relates to a method and system for predicting the utilization rate of amino acids in multi-stage feed for laying hens. Background Technology

[0002] In the field of precision layer hen feeding, accurately assessing feed amino acid utilization is crucial for achieving nutritional optimization and reducing feeding costs. Standard ileal amino acid digestibility is a core indicator for evaluating the biological value of amino acids. Currently, this indicator is mainly obtained through traditional in vivo assays. While these methods are accurate, they suffer from problems such as long experimental cycles, high costs, and significant impacts on animal welfare, making them unsuitable for the rapid assessment of large quantities of feed ingredients in actual production. Although in vitro prediction methods developed in recent years have improved detection efficiency to some extent, their models are mostly based on static nutrient data and fail to fully consider the dynamic differences in metabolic characteristics at different physiological stages of layer hens, as well as the complex influence of ileal microbial communities and endogenous amino acid loss on the digestion process. This limits the universality and accuracy of the prediction models, especially when facing different laying cycles and different combinations of feed ingredients, resulting in significant prediction deviations and making it difficult to support the practical application of precision feed formulation.

[0003] Based on the shortcomings of the existing technology, there is an urgent need for a method and system for predicting the amino acid utilization rate of multi-stage feed for laying hens. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for predicting the amino acid utilization rate of multi-stage feed for laying hens, in order to improve the aforementioned problems. To achieve the above objective, the technical solution adopted by this invention is as follows: In a first aspect, this application provides a method and system for predicting the amino acid utilization rate of multi-stage feed for laying hens, including: Chemical composition and in vivo digestibility of standard ileal amino acids were determined from feed ingredients samples at different growth stages of laying hens. Predictive factors were screened based on the measured values ​​of the chemical composition and the measured values ​​of the standard ileal amino acid digestibility in vivo. The correlation between the nutritional components of feed ingredients and the metabolic characteristics of laying hens during the egg production cycle was analyzed using the Akaike Information Criterion to obtain a combination of predictive factors. A prediction model was constructed based on the combination of prediction factors. By establishing a multiple regression relationship between the amino acid flow characteristics of ileal digesta in laying hens and the combination of prediction factors, candidate prediction equations were obtained. The candidate prediction equations were optimized by comparing the coefficients of determination and biological rationality of each equation at different growth stages of laying hens, and the optimized equations were selected. The model was validated based on the optimization equation. The relative deviation between the predicted value and the in vivo measured value was calculated, and a validation mechanism was constructed by combining the correction of endogenous amino acid loss in laying hens to obtain the target prediction model. Digestibility is predicted based on the target prediction model. By inputting the predictive factor content of the feed ingredients for laying hens to be tested into the target prediction model, the predicted standard ileal amino acid digestibility value of the current feed formula is obtained.

[0005] Secondly, this application also provides a method and system for predicting the amino acid utilization rate of multi-stage feed for laying hens, including: The acquisition module is used to acquire the chemical composition values ​​and in vivo intestinal amino acid digestibility values ​​of feed ingredient samples from different growth stages of laying hens. The screening module is used to screen predictive factors based on the measured values ​​of the chemical components and the measured values ​​of the standard ileal amino acid digestibility in vivo. The module also uses the Akaike Information Criterion to analyze the correlation between the nutritional components of feed ingredients and the metabolic characteristics of laying hens during the egg production cycle, and then performs dimensionality reduction mining to obtain a combination of predictive factors. The construction module is used to construct a prediction model based on the combination of prediction factors. By establishing a multiple regression relationship between the amino acid flow characteristics of ileal digesta in laying hens and the combination of prediction factors, candidate prediction equations are obtained. The optimization module is used to optimize the candidate prediction equations by comparing the coefficients of determination and biological rationality of each equation at different growth stages of laying hens, and then selecting the optimal equation. The verification module is used to verify the model according to the optimization equation. It calculates the relative deviation between the predicted value and the in vivo measured value, and constructs a verification mechanism by combining the correction of endogenous amino acid loss in laying hens to obtain the target prediction model. The prediction module is used to predict digestibility based on the target prediction model. By inputting the predictor content of the feed ingredients for laying hens to be tested into the target prediction model, the predicted standard ileal amino acid digestibility value of the current feed formula is obtained.

[0006] The beneficial effects of this invention are as follows: This invention establishes a dynamic correlation model between the metabolic characteristics of laying hens during the egg-laying cycle and the nutritional components of feed. Combined with a multi-stage screening and verification mechanism and an endogenous amino acid loss compensation algorithm, it achieves accurate prediction of the digestibility of standard ileal amino acids based on in vitro detection data, effectively overcoming the limitations of traditional in vivo assay methods. Attached Figure Description

[0007] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0008] Figure 1 This is a flowchart illustrating a method for predicting the multi-stage feed amino acid utilization rate of laying hens as described in an embodiment of the present invention. Figure 2 This is a schematic diagram of the structure of a multi-stage feed amino acid utilization prediction system for laying hens as described in an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of a device for predicting the amino acid utilization rate of multi-stage feed for laying hens, as described in an embodiment of the present invention.

[0009] The diagram is labeled as follows: 800, a predictive device for amino acid utilization rate in multi-stage feed for laying hens; 801, processor; 802, memory; 803, multimedia component; 804, I / O interface; 805, communication component; 901, acquisition module; 902, screening module; 903, construction module; 904, optimization module; 905, verification module; 906, prediction module. Detailed Implementation

[0010] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0011] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this invention, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0012] Example 1

[0013] This embodiment provides a method and system for predicting the utilization rate of amino acids in multi-stage feed for laying hens.

[0014] See Figure 1 The figure shows that the method includes steps S100 to S600.

[0015] Step S100: Obtain the chemical composition values ​​and in vivo intestinal amino acid digestibility values ​​of feed ingredient samples from different growth stages of laying hens. Understandably, in precision layer hen feeding practices, while traditional in vivo assays can obtain accurate data on the standard ileal amino acid digestibility, they suffer from limitations such as long experimental cycles and high costs. Step S100 addresses these pain points by designing a systematic data collection scheme: First, representative layer hen feed ingredient samples are selected, covering ingredients from different origins, batches, and processing techniques to ensure sample diversity and representativeness. Second, ileal digesta samples are systematically collected at different growth stages of layer hens (such as brooding, rearing, peak laying, and molting), and standardized assay procedures are used to obtain chemical composition values, including conventional nutrients (such as crude protein, crude fat, and crude fiber) and specific indicators (such as neutral detergent fiber, acid detergent fiber, and mineral content). Simultaneously, the standard ileal amino acid digestibility values ​​are determined through strictly controlled in vivo experiments to ensure data accuracy and reliability. This step establishes a complete dataset covering the entire growth cycle of layer hens, including both chemical characteristic data of raw materials and corresponding biological potency data, laying the foundation for subsequent development of precise prediction models.

[0016] Step S200: Based on the chemical composition determination value and the in vivo determination value of standard ileal amino acid digestibility, predictive factors are screened. The correlation between the nutritional components of feed ingredients and the metabolic characteristics of laying hens during the egg production cycle is analyzed by the Akaike Information Criterion to obtain a combination of predictive factors. It should be noted that this step uses the Akaike Information Criterion to perform dimensionality reduction mining on multidimensional data. This method can effectively identify key predictive factors that are closely related to the metabolic characteristics of laying hens during the egg production cycle. By quantitatively analyzing the dynamic correlation strength between various nutrients and digestibility, the method achieves the goal of accurately screening core indicator combinations from a large number of potential influencing factors.

[0017] Step S300: Construct a prediction model based on the combination of prediction factors. By establishing a multiple regression relationship between the amino acid flow characteristics of ileal digesta in laying hens and the combination of prediction factors, candidate prediction equations are obtained. Understandably, this step constructs a multiple regression model based on the combination of predictive factors obtained through screening. This model takes into account the unique patterns of amino acid flow in the ileum of laying hens. By establishing a mathematical relationship between predictive factors and digestibility, a predictive equation framework that can reflect the actual digestive physiological process is formed.

[0018] Step S400: Optimize the candidate prediction equations by comparing the determination coefficients and biological rationality of each equation at different growth stages of laying hens, and then select the optimal equation. It should be noted that this step involves multi-dimensional optimization and screening of candidate prediction equations, with a focus on evaluating the differences in predictive performance of each equation at different growth stages. This process fully considers the impact of changes in the physiological state of laying hens on the applicability of the model, ensuring that the final optimized equation has both good statistical properties and conforms to biological logic.

[0019] Step S500: Validate the model based on the optimization equation. Calculate the relative deviation between the predicted value and the in vivo measured value, and construct a validation mechanism by combining it with the correction for endogenous amino acid loss in laying hens, to obtain the target prediction model. Understandably, the validation mechanism established in this step combines prediction bias analysis with endogenous amino acid loss correction. By constructing a multi-level validation system, it effectively improves the reliability and accuracy of the model in practical applications and significantly improves the shortcomings of traditional prediction methods in failing to consider complex physiological factors.

[0020] Step S600: Predict digestibility based on the target prediction model. By inputting the predictive factor content of the feed ingredients to be tested into the target prediction model, the predicted standard ileal amino acid digestibility value of the current feed formula is obtained.

[0021] It should be noted that this step, by applying the target prediction model to predict the digestibility of actual feed formulations, realizes the transformation from theoretical model to practical application. The design of this step reflects the applicability of the method in the actual production environment and provides effective technical support for precision breeding of laying hens.

[0022] Further, step S200 includes steps S210 to S230.

[0023] Step S210: Based on the chemical composition determination value and the in vivo determination value of standard ileal amino acid digestibility, a dynamic correlation analysis is performed. By establishing a time series phase synchronization model of egg production rate fluctuation and amino acid metabolic demand in laying hens, nutritional factors that respond synchronously to physiological changes in the egg production cycle are identified. Step S220: Perform combination optimization processing based on nutritional factors, analyze the functional correlation strength between factors and construct modular grouping based on clustering coefficients, and screen out the minimum factor set with functional synergy during the peak egg production period and molting period of laying hens. Step S230: Perform biological verification processing based on the minimum factor set. By establishing a multi-level interaction network model between the ileal microbial community of laying hens and the predictive factors, a combination of predictive factors characterizing the amino acid digestion characteristics of laying hens during the egg-laying stage is obtained.

[0024] Specifically, step S210 first establishes a time-series phase synchronization model to dynamically align the egg production rate fluctuation curve of laying hens with changes in amino acid metabolic requirements. This method can capture the temporal correlation characteristics between nutritional factors and physiological cycles, identifying key nutritional factors that fluctuate synchronously with changes in the egg production cycle. Based on this, step S220 further employs a modular grouping algorithm based on clustering coefficients. By calculating the functional correlation strength between different nutritional factors, factors with synergistic effects are grouped into the same functional module. The minimum spanning tree principle in graph theory is then used to screen out the minimum set of factors that can cover the needs of key physiological stages such as peak egg production and molting in laying hens. Finally, step S230 constructs a multi-level interaction network model to analyze the correlation between the screened minimum set of factors and the ileal microbial community structure of laying hens. This network model includes multiple interaction relationships between microbial populations and nutritional factors, which can verify the biological rationality of the predictive factor combination at the system level, thus ensuring that the final predictive factor combination not only has statistical significance but also truly reflects the unique amino acid digestion and metabolism characteristics of laying hens during the egg production stage.

[0025] Further, step S300 includes steps S310 to S330.

[0026] Step S310: Perform dynamic modeling based on the combination of predictive factors. By introducing the time series phase adjustment of the amino acid flow rate of ileal digesta in the laying cycle of laying hens, a primary regression equation based on time dimension correction is constructed. Step S320: Perform structural optimization based on the primary regression equation, and establish a dynamic prediction framework based on quantile regression by analyzing the distribution characteristics of amino acid absorption efficiency at different egg-laying stages. Step S330: Apply biological constraints based on the dynamic prediction framework. By integrating the coupling relationship between the ileal microbial metabolic pathway and the amino acid digestion process of laying hens, candidate prediction equations that conform to the digestive physiological characteristics of laying hens are obtained.

[0027] Specifically, step S310 first uses time-series phase adjustment technology to dynamically align the changing patterns of amino acid flow rates in the ileum during the laying cycle of hens with the predictive factor data. This method effectively eliminates time-dimensional bias caused by physiological cycle fluctuations, and the constructed primary regression equation already possesses preliminary time-dynamic characteristics. Building upon this, step S320 employs quantile regression to deeply analyze the distribution characteristics of amino acid absorption efficiency at different laying stages. This method can capture the changing patterns at different quantiles in the data distribution, thereby establishing a dynamic prediction framework that adapts to the characteristics of each physiological stage of laying hens. Step S330 further integrates the coupling relationship between the ileal microbial metabolic pathway and the amino acid digestion process. By establishing a correlation model between microbial community function and digestion efficiency, biological constraints are added to the prediction equation, ensuring that the final candidate prediction equation not only has mathematical statistical significance but also truly reflects the physiological characteristics of the laying hen's digestive system. This series of processes embodies a progressive optimization approach, from time-dimensional correction to distribution characteristic analysis, and then to biological mechanism constraints, ensuring that the prediction model conforms to mathematical laws and closely reflects actual physiological processes.

[0028] In a preferred embodiment of the present invention, for any growth stage of poultry, after determining each set of candidate predictors, a primary regression equation for poultry at that growth stage is constructed based on each set of candidate predictors, as shown in Table 1.

[0029] Table 1. Primary regression equation for the digestibility of leucine in the ileum of wheat middlings raw material.

[0030] In the prediction method for multi-stage feed amino acid utilization in laying hens, the derivation of candidate prediction equations is based on a multiple linear regression model. First, the digestibility of standard ileal amino acids is screened from the chemical composition of feed ingredients using the Akaike information criterion. SID Predictive factors that have a significant impact (such as crude ash and acid detergent fiber). The selection of these factors takes into account the metabolic characteristics of the laying hen's egg production cycle. For example, during the peak egg production period, the demand for amino acids is higher, so the factors need to reflect the digestibility of energy and protein.

[0031] The general form of a multiple linear regression model is: ; In the formula, β0 is the intercept term; β i X is the regression coefficient of the i-th predictor, where i = 1, 2, ..., k; i Represents the measured value of the predictor; This is the error term.

[0032] Specifically, in Table 1, each equation represents the standard ileal leucine digestibility of the secondary wheat flour raw material. SID The predictive model for this study included crude ash (representing the inorganic content of feed, affecting digestibility), acid detergent fiber (representing cellulose content, reducing digestibility), neutral detergent fiber (representing hemicellulose and cellulose, affecting digestion), bulk density (feed density, potentially affecting digestion rate), calcium (a mineral involved in metabolic regulation), albumin (a protein component affecting amino acid supply), and total energy (total energy in feed, reflecting metabolic potential). The coefficients of these factors were obtained by fitting experimental data, where R-squared and p-values ​​were used to evaluate the goodness of fit and significance of the model.

[0033] Next, based on the primary regression equation, the model structure was optimized using quantile regression. Quantile regression is used to capture the distribution characteristics of amino acid uptake efficiency in different laying stages of hens (such as peak laying and molting), rather than just the conditional mean. This allows the model to adapt to different quantiles of the distribution (such as...). =0.25, 0.5, 0.75), thus better reflecting physiological fluctuations. The general form of the quantile regression model is: ; In the formula, This represents the standard ileal amino acid digestibility under a given predictor factor X. Quantile conditional value; β0( ) represents the quantiles The intercept term below; β1( ) represents the quantiles The coefficient of the i-th predictor represents the marginal effect of the factor at different locations in the distribution (e.g., low, medium, and high digestibility levels); the coefficient is estimated by minimizing the check function, which enables the model to capture outliers and asymmetric distributions.

[0034] Taking equation A1 in Table 1 as an example, we can extend it to a quantile regression framework. For example, for quantiles... =0.5 (median), the quantile regression equation can be written as: ; In the formula, X represents the conditional median of standard ileal amino acid digestibility under a given predictor factor X; X1 is the measured value of crude ash; X2 is the measured value of acid detergent fiber; X3 is the measured value of neutral detergent fiber; X4 is the measured value of bulk density; β0 (0.5) is the baseline digestibility at the median quantile, representing the baseline level of laying hens at a typical laying stage; β1 (0.5) is the coefficient of crude ash at the median quantile, representing the strength of the effect of inorganic content on digestibility; β2 (0.5) is the coefficient of acid detergent fiber at the median quantile, representing the inhibitory effect of cellulose on digestion; β3 (0.5) is the coefficient of neutral detergent fiber at the median quantile, representing the promoting effect of hemicellulose on microbial activity; β4 (0.5) is the coefficient of bulk density at the median quantile, representing the effect of feed density on digesta circulation.

[0035] Next, based on the quantile regression framework, biological constraints were introduced to integrate the coupling relationship between the ileal microbial metabolic pathway and the amino acid digestion process in laying hens. This was achieved by adding microbial-related variables or constraints to make the equations conform to physiological reality. For example, microbial metabolic pathways may affect the efficiency of amino acid digestion, so a microbial activity factor was introduced as an interaction term or additional predictor. Biological constraints can be applied through constraint optimization, such as ensuring that the predicted values ​​are within a physiologically feasible range (e.g., SID between 0-100%).

[0036] Based on the quantile regression equation, the constraint treatment can be expressed as the modified equation: ; In the formula, M represents the microbial activity factor (such as the abundance of the ileal microbial community). For microbial activity factors at quantiles The coefficient below reflects the direct enhancing effect of microbial metabolism on digestibility; The coefficient represents the interaction term, indicating the synergistic effect of predictor microbial activity.

[0037] Further, step S400 includes steps S410 to S430.

[0038] Step S410: Perform multi-stage screening based on candidate prediction equations. Analyze the differences in metabolic characteristics between the peak egg production period and molting period of laying hens by establishing a decision model based on dynamic programming, and obtain a preliminary set of screening equations. Step S420: Based on the preliminary screening of the equation set, the biological rationality is verified. The physiological fit is analyzed by constructing a phase-matching model between the laying hen amino acid metabolic pathway and the predicted equations to obtain the verified equation set. Step S430: Optimize the equations based on the verified set of equations. Establish an equation correction model by introducing a compensation mechanism for the loss of endogenous amino acids in laying hens to obtain the optimized equations.

[0039] Specifically, step S410 employs a dynamic programming algorithm to perform multi-stage screening of candidate prediction equations. This algorithm describes the metabolic characteristics of laying hens from peak egg production to molting by establishing state transition equations. The fitness of the prediction equations is evaluated at each physiological stage node, thereby selecting a subset of equations that can adapt to different physiological state changes. The state transition equation is expressed as: ; In the formula, S t S is the metabolic state index (dimensionless) of laying hens at time t, reflecting the current metabolic efficiency; t+1 The metabolic state index of laying hens at time t+1; This indicates the egg-laying stage at time t; Indicates the egg-laying stage at time t+1; κ is the feed input influence coefficient, representing the immediate effect of feed changes on metabolic state; I t η represents the feed input index at time t; η is the stage change influence coefficient, which represents the adjustment range of metabolic state during the transition of egg production stage.

[0040] Step S420 builds a phase-matching model based on this, comparing the output of the predicted equation with the actual amino acid metabolic pathways in laying hens. By calculating the phase difference between the predicted and actual physiological values ​​over time, the degree of agreement between the equation output and the biological process is assessed, ensuring that the selected equations conform to the digestive physiology of laying hens. The phase-matching model quantifies prediction bias through time series alignment, expressed as: ; In the formula, The optimal time offset (unit: days) is the offset that maximizes the cross-correlation and reflects the phase difference between the prediction and the actual value. The time offset represents the time delay or advance of the predicted value relative to the actual value; This represents the actual measurement at time t. SID value; The predicted standard ileal amino acid digestibility at time t is derived from the candidate equation output; T is the total time, representing the length of the observation period.

[0041] Step S430 further introduces an endogenous amino acid loss compensation mechanism. By establishing a quantitative relationship model between the loss amount and feed composition, the prediction equation is systematically corrected. This process effectively solves the systematic bias problem caused by neglecting endogenous losses in traditional prediction methods, enabling the final optimized equation to more accurately reflect the true amino acid digestibility and utilization rate of laying hens. The entire optimization process embodies a progressive optimization approach from mathematical screening to biological verification, and then to physiological mechanism correction, ensuring that the prediction model has both statistical reliability and conforms to physiological laws. Specifically, the quantitative relationship model between the loss amount and feed composition quantifies the impact of feed components on endogenous losses through a linear relationship, expressed as: ; In the formula, E represents the amount of endogenous amino acid loss, indicating the loss of amino acids secreted endogenously during the digestion of laying hens; The content of acid detergent fiber in feed; This refers to the neutral detergent fiber content in feed. λ represents the crude protein content of the feed; λ0 represents the baseline endogenous loss, indicating the minimum loss when there is no fiber and protein; λ1, λ2, and λ3 represent the influence coefficients of each component. In a preferred embodiment of this application, the content corresponding to the predictor factor is substituted into the standard ileal amino acid digestibility prediction equation for any growth stage to calculate the standard ileal amino acid digestibility, as shown in Table 2. The measured average value is 79.95%, with a relative deviation between 2.08% and 2.55%. Among them, the relative deviation of candidate prediction equation A03 is only 2.08%, so equation A03 is the prediction equation with the highest accuracy. However, the significance of A03 does not meet the requirements, so A04 is the optimal prediction equation.

[0042] Table 2 Comparison of predicted and measured values ​​of ileal amino acid digestibility of middlings raw material

[0043] Further, step S500 includes steps S510 to S530.

[0044] Step S510: Perform deviation distribution analysis based on the optimization equation. Establish a deviation probability model by statistically analyzing the deviation distribution characteristics between predicted values ​​and in vivo measured values ​​at different egg-laying stages to obtain the stage deviation analysis results. Step S520: Based on the results of the phased deviation analysis, metabolic compensation treatment is carried out. Dynamic compensation and correction are performed by establishing a metabolic pathway association network between endogenous amino acid loss and digestibility deviation in laying hens, and the verification indicators after metabolic compensation are obtained. Step S530: Based on the validation index after metabolic compensation, perform model convergence determination processing, optimize model parameters by setting multi-objective convergence conditions based on biological rationality, and obtain the target prediction model.

[0045] Specifically, step S510 first systematically analyzes the prediction deviation of the optimization equation at different egg-laying stages by establishing a deviation probability model. This method uses a Gaussian mixture model to fit the deviation distribution between the predicted value and the in vivo measured value, which can accurately identify the central tendency and discrete characteristics of the deviation distribution, especially capturing the deviation patterns unique to key physiological stages such as peak egg production and molting. Based on this, step S520 constructs a metabolic pathway association network between endogenous amino acid loss and digestibility deviation. This network model uses influencing factors such as ileal microbial metabolites and digestive enzyme activity as network nodes. By calculating the correlation strength between nodes, the contribution of each factor to digestibility deviation is determined, thereby achieving targeted dynamic compensation and correction. Finally, step S530 sets multi-objective convergence conditions based on biological rationality. These conditions include the allowable range of prediction error, the physiologically feasible interval, and the limitation of model complexity. Through multiple rounds of iterative optimization, the model parameters simultaneously meet the requirements of statistical accuracy and physiological rationality, ultimately obtaining a target prediction model that has both prediction accuracy and conforms to the digestive physiological characteristics of laying hens.

[0046] Further, step S600 includes steps S610 to S630.

[0047] Step S610: Preprocess the input parameters according to the target prediction model. By synchronizing the predictor content of the feed raw materials to be tested with the current laying cycle of the laying hen, calculate the time weight coefficient of each predictor in a specific laying stage, and obtain the spatiotemporal calibration input parameter set. Step S620: Perform digestibility calculation based on the input parameter set. By establishing a dynamic response surface between the metabolic state of laying hens and the predictor, the optimal digestibility prediction value is solved based on the gradient descent method to obtain the preliminary prediction results. Step S630: Based on the preliminary prediction results, physiological compensation is performed. By constructing a feedback regulation network between the loss of endogenous amino acids in laying hens and the composition of feed, an iterative approximation algorithm is used to physiologically adapt the prediction results to obtain the predicted value of standard ileal amino acid digestibility.

[0048] Specifically, step S610 first synchronizes the predictor content of the feed ingredients to be tested with the current laying cycle of the laying hens using time series analysis. Fourier transform is then used to extract the fundamental frequency and harmonic components of the laying cycle, and the time weight coefficients of each predictor at different physiological stages are calculated to achieve spatiotemporal calibration of the input parameters. Step S620 establishes a dynamic response surface model based on the calibrated parameter set. This model uses a radial basis function neural network to construct a nonlinear mapping relationship between the predictor and metabolic state. The optimal solution is found on the multidimensional response surface using a conjugate gradient descent algorithm to obtain a digestibility prediction value that matches the current physiological state. Step S630 further constructs a feedback regulation network for endogenous amino acid loss and feed composition. This network uses parameters such as feed component decomposition rate and digesta flow rate as nodes, and iteratively optimizes using an adaptive genetic algorithm. By physiologically adapting the prediction results, a standard ileal amino acid digestibility prediction value that accurately reflects the actual digestive status of the laying hens is finally obtained.

[0049] Compared with the traditional in vivo assay method, the prediction method provided in this invention requires animal testing, expensive testing equipment, and the purchase and employment of poultry farmers to raise poultry. The standard ileal amino acid digestibility is then determined based on animal testing. Taking the determination of the standard ileal amino acid digestibility of 66 raw materials as an example, as shown in Table 3, the traditional method for determining the standard ileal amino acid digestibility costs approximately RMB 132,000 from test preparation to completion, taking 30-60 days. It also requires farm rental fees, farm management fees, labor costs for farmers and test personnel, and the purchase of equipment such as amino acid analyzers, resulting in significant consumption of manpower, financial resources, and time. However, the in vitro prediction method provided in this invention, after constructing a prediction model for the digestibility of the target standard ileal amino acids, only requires measuring 1-4 key predictive factors in the secondary flour in subsequent measurements. This allows for the direct calculation of the predicted value of the standard ileal amino acid digestibility without animal experiments. Measuring the aforementioned 66 samples (30 raw material samples + 36 chyme samples) would only cost approximately 0.11 million yuan, which is 0.83% of the cost of traditional methods. Furthermore, the measurement cycle is 1-4 days, which is 1.67%~3.33% of the cycle of traditional methods, significantly saving manpower, material resources, financial resources, and time. In the future, if this method is combined with near-infrared spectroscopy (NIRS), the detection cost will be further reduced.

[0050] Table 3 Comparison of the prediction method of the present invention with traditional methods

[0051] Compared to traditional in vivo methods, the in vitro prediction method of this invention does not require animal testing, saving on farm rental fees, animal purchase and feeding costs, and avoiding the harm to animal welfare caused by force-feeding. It significantly increases the number of samples evaluated per unit time and shortens the evaluation cycle, thus greatly saving manpower, material resources, financial resources, and time. Furthermore, because the predictive factors screened in this invention have a wide range, the prediction accuracy of standard ileal amino acid digestibility is significantly improved. In addition, the prediction method of this invention is simple and easy to implement. Compared to traditional Excel calculation methods, by embedding the prediction equation into a poultry feed net energy prediction mini-program, it can be widely promoted through WeChat mini-programs or other mobile applications and computer software, demonstrating high practical value.

[0052] Example 2 like Figure 2 As shown in the figure, this embodiment provides a prediction system for multi-stage feed amino acid utilization in laying hens. The system includes: The acquisition module 901 is used to acquire the chemical composition determination values ​​and the in vivo determination values ​​of standard ileal amino acid digestibility of feed ingredient samples at different growth stages of laying hens. The screening module 902 is used to screen predictive factors based on the chemical composition determination value and the in vivo determination value of standard ileal amino acid digestibility. The correlation between the nutritional components of feed ingredients and the metabolic characteristics of laying hens during the egg production cycle is analyzed by the Akaike Information Criterion to obtain a combination of predictive factors. Module 903 is used to construct a prediction model based on the combination of predictive factors. By establishing a multiple regression relationship between the amino acid flow characteristics of ileal digesta in laying hens and the combination of predictive factors, candidate prediction equations are obtained. The optimization module 904 is used to optimize candidate prediction equations by comparing the coefficients of determination and biological rationality of each equation at different growth stages of laying hens, and then selecting the optimal equation. The validation module 905 is used to validate the model based on the optimization equation. It calculates the relative deviation between the predicted value and the in vivo measured value, and constructs a validation mechanism by combining the correction of endogenous amino acid loss in laying hens, thereby obtaining the target prediction model. The prediction module 906 is used to predict digestibility based on the target prediction model. By inputting the predictor content of the feed ingredients of the laying hens to be tested into the target prediction model, the predicted standard ileal amino acid digestibility value of the current feed formula is obtained.

[0053] In one specific embodiment of this application, the screening module 902 includes: The first screening unit is used to perform dynamic correlation analysis based on the measured values ​​of chemical composition and the in vivo measured values ​​of standard ileal amino acid digestibility. By establishing a time series phase synchronization model of egg production rate fluctuation and amino acid metabolic demand, nutritional factors that respond synchronously to physiological changes in the egg production cycle are identified. The second screening unit is used to perform combination optimization based on nutritional factors. By analyzing the functional correlation strength between factors and constructing modular grouping based on clustering coefficients, the minimum set of factors with functional synergy during the peak egg production period and molting period of laying hens is screened. The third screening unit is used for biological validation based on the minimum factor set. By establishing a multi-level interaction network model between the ileal microbial community of laying hens and the predictive factors, a combination of predictive factors characterizing the amino acid digestion characteristics of laying hens during the egg-laying stage is obtained.

[0054] In one specific embodiment of this application, the construction module 903 includes: The first building unit is used for dynamic modeling based on the combination of predictive factors. By introducing the time series phase adjustment of the amino acid flow rate of ileal digesta in the laying cycle of laying hens, a primary regression equation based on time dimension correction is constructed. The second building block is used to perform structural optimization based on the primary regression equation. By analyzing the distribution characteristics of amino acid absorption efficiency at different egg-laying stages, a dynamic prediction framework based on quantile regression is established. The third building block is used to apply biological constraints based on the dynamic prediction framework. By integrating the coupling relationship between the ileal microbial metabolic pathway and the amino acid digestion process of laying hens, candidate prediction equations that conform to the digestive physiological characteristics of laying hens are obtained.

[0055] In one specific embodiment of this application, the optimization module 904 includes: The first optimization unit is used to perform multi-stage screening based on candidate prediction equations. By establishing a decision model based on dynamic programming, the metabolic characteristics of laying hens during peak egg production and molting are analyzed to obtain a preliminary set of screening equations. The second optimization unit is used to verify the biological rationality of the preliminary set of equations. It analyzes the physiological fit by constructing a phase-matching model between the laying hen amino acid metabolic pathway and the predicted equations, and obtains the verified set of equations. The third optimization unit is used to optimize the set of equations after verification. By introducing a compensation mechanism for the loss of endogenous amino acids in laying hens, an equation correction model is established to obtain the optimized equations.

[0056] In one specific embodiment of this application, the verification module 905 includes: The first verification unit is used to perform deviation distribution analysis based on the optimization equation. By statistically analyzing the deviation distribution characteristics between predicted values ​​and in vivo measured values ​​at different egg-laying stages, a deviation probability model is established to obtain the stage deviation analysis results. The second verification unit is used to perform metabolic compensation treatment based on the results of the phased deviation analysis. It dynamically compensates and corrects the metabolic pathways associated with the loss of endogenous amino acids in laying hens and the deviation in digestibility by establishing a network, and obtains the verification indicators after metabolic compensation. The third verification unit is used to determine the model convergence based on the verification indicators after metabolic compensation. It optimizes the model parameters by setting multi-objective convergence conditions based on biological rationality to obtain the target prediction model.

[0057] In one specific embodiment of this application, the prediction module 906 includes: The first prediction unit is used to preprocess the input parameters according to the target prediction model. By synchronizing the content of the predictor factors of the feed raw materials to be tested with the current laying cycle of the laying hens, the time weight coefficient of each predictor factor in a specific laying stage is calculated to obtain the spatiotemporal calibration input parameter set. The second prediction unit is used to calculate digestibility based on the input parameter set. By establishing a dynamic response surface between the metabolic state of laying hens and the prediction factors, the optimal digestibility prediction value is solved based on the gradient descent method to obtain the preliminary prediction results. The third prediction unit is used to perform physiological compensation processing based on the preliminary prediction results. By constructing a feedback regulation network between the loss of endogenous amino acids in laying hens and the feed composition, an iterative approximation algorithm is used to make physiological adaptive corrections to the prediction results, and the predicted value of standard ileal amino acid digestibility is obtained.

[0058] Example 3 Corresponding to the above method embodiments, this embodiment also provides a device for predicting the amino acid utilization rate of multi-stage feed for laying hens. The device for predicting the amino acid utilization rate of multi-stage feed for laying hens described below and the method for predicting the amino acid utilization rate of multi-stage feed for laying hens described above can be referred to in correspondence.

[0059] Figure 3 This is a block diagram illustrating a predictive device 800 for multi-stage feed amino acid utilization in laying hens, according to an exemplary embodiment. Figure 3 As shown, the predictive device 800 for the amino acid utilization rate of multi-stage feed for laying hens may include: a processor 801 and a memory 802. The predictive device 800 may also include one or more of a multimedia component 803, an I / O interface 804, and a communication component 805.

[0060] The processor 801 controls the overall operation of the multi-stage feed amino acid utilization rate prediction device 800 for laying hens to complete all or part of the steps in the aforementioned method for predicting the multi-stage feed amino acid utilization rate for laying hens. The memory 802 stores various types of data to support the operation of the multi-stage feed amino acid utilization rate prediction device 800. This data may include, for example, instructions for any application or method operating on the multi-stage feed amino acid utilization rate prediction device 800, as well as application-related data such as contact data, sent and received messages, images, audio, video, etc. The memory 802 can be implemented using 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. The multimedia component 803 may include a screen and an audio component. The screen may be, for example, a touchscreen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signals may be further stored in the memory 802 or transmitted via the communication component 805. The audio component also includes at least one speaker for outputting audio signals. I / O interface 804 provides an interface between processor 801 and other interface modules, such as keyboards, mice, and buttons. These buttons can be virtual or physical. Communication component 805 is used for wired or wireless communication between the multi-stage feed amino acid utilization prediction device 800 for laying hens and other devices. Wireless communication includes Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G, or 4G, or a combination thereof. Therefore, the corresponding communication component 805 may include a Wi-Fi module, a Bluetooth module, or an NFC module.

[0061] In an exemplary embodiment, a predictive device 800 for the multi-stage feed amino acid utilization rate of laying hens may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the aforementioned predictive method for the multi-stage feed amino acid utilization rate of laying hens.

[0062] In another exemplary embodiment, a computer-readable storage medium including program instructions is also provided. When executed by a processor, these program instructions implement the steps of the above-described method for predicting the amino acid utilization rate of multi-stage feed for laying hens. For example, the computer-readable storage medium may be the memory 802 including the program instructions, which may be executed by a processor 801 of a device 800 for predicting the amino acid utilization rate of multi-stage feed for laying hens to complete the above-described method for predicting the amino acid utilization rate of multi-stage feed for laying hens.

[0063] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for predicting the amino acid utilization rate of multi-stage feed for laying hens, characterized in that, include: Chemical composition and in vivo digestibility of standard ileal amino acids were determined from feed ingredients samples at different growth stages of laying hens. Predictive factors were screened based on the measured values ​​of the chemical composition and the measured values ​​of the standard ileal amino acid digestibility in vivo. The correlation between the nutritional components of feed ingredients and the metabolic characteristics of laying hens during the egg production cycle was analyzed using the Akaike Information Criterion to obtain a combination of predictive factors. A prediction model was constructed based on the combination of prediction factors. By establishing a multiple regression relationship between the amino acid flow characteristics of ileal digesta in laying hens and the combination of prediction factors, candidate prediction equations were obtained. The candidate prediction equations were optimized by comparing the coefficients of determination and biological rationality of each equation at different growth stages of laying hens, and the optimized equations were selected. The model was validated based on the optimization equation. The relative deviation between the predicted value and the in vivo measured value was calculated, and a validation mechanism was constructed by combining the correction of endogenous amino acid loss in laying hens to obtain the target prediction model. Digestibility is predicted based on the target prediction model. By inputting the predictive factor content of the feed ingredients for laying hens to be tested into the target prediction model, the predicted standard ileal amino acid digestibility value of the current feed formula is obtained.

2. The method for predicting the amino acid utilization rate of multi-stage feed for laying hens according to claim 1, characterized in that, Predictive factor screening was performed based on the measured values ​​of the chemical composition and the in vivo measured values ​​of the standard ileal amino acid digestibility, including: Based on the measured values ​​of the chemical components and the measured values ​​of the standard ileal amino acid digestibility in vivo, dynamic correlation analysis was performed. By establishing a time series phase synchronization model of egg production rate fluctuations and amino acid metabolic requirements in laying hens, nutritional factors that respond synchronously to physiological changes in the egg production cycle were identified. Based on the nutritional factors, a combination optimization process was performed. By analyzing the functional correlation strength between factors and constructing a modular grouping based on the clustering coefficient, the minimum set of factors with functional synergy during the peak egg production period and molting period of laying hens was screened. Biological validation was performed based on the minimum factor set. By establishing a multi-level interaction network model between the ileal microbial community of laying hens and the predictive factors, a combination of predictive factors characterizing the amino acid digestion characteristics of laying hens during the egg-laying stage was obtained.

3. The method for predicting the amino acid utilization rate of multi-stage feed for laying hens according to claim 1, characterized in that, Constructing a prediction model based on the combination of predictor factors includes: Dynamic modeling is performed based on the combination of predictive factors. By introducing the time-series phase adjustment of the amino acid flow rate of ileal digesta during the laying cycle of laying hens, a primary regression equation based on time dimension correction is constructed. Based on the primary regression equation, structural optimization was performed, and a dynamic prediction framework based on quantile regression was established by analyzing the distribution characteristics of amino acid absorption efficiency at different egg-laying stages. Biological constraints were applied based on the dynamic prediction framework, and candidate prediction equations that conform to the digestive physiology of laying hens were obtained by integrating the coupling relationship between the ileal microbial metabolic pathway and the amino acid digestion process.

4. The method for predicting the amino acid utilization rate of multi-stage feed for laying hens according to claim 1, characterized in that, Optimization is performed based on the candidate prediction equation, including: Based on the candidate prediction equations, a multi-stage screening process was carried out. By establishing a decision model based on dynamic programming, the differences in metabolic characteristics between the peak egg production period and the molting period of laying hens were analyzed, and a preliminary set of screening equations was obtained. The biological rationality of the preliminary set of equations was verified. The physiological fit was analyzed by constructing a phase-matching model between the laying hen amino acid metabolic pathway and the predicted equations, and the verified set of equations was obtained. The validated set of equations was optimized by introducing a compensation mechanism for the loss of endogenous amino acids in laying hens to establish an equation correction model, resulting in optimized equations.

5. The method for predicting the amino acid utilization rate of multi-stage feed for laying hens according to claim 1, characterized in that, Model validation is performed based on the optimization equations, including: Based on the optimization equation, deviation distribution analysis was performed. By statistically analyzing the deviation distribution characteristics between predicted values ​​and in vivo measured values ​​at different egg-laying stages, a deviation probability model was established to obtain the stage deviation analysis results. Based on the results of the phased deviation analysis, metabolic compensation was performed. Dynamic compensation and correction were carried out by establishing a metabolic pathway association network between endogenous amino acid loss and digestibility deviation in laying hens, and validation indicators after metabolic compensation were obtained. The model convergence determination process is performed based on the validation index after metabolic compensation. The model parameters are optimized by setting multi-objective convergence conditions based on biological rationality to obtain the target prediction model.

6. A predictive system for amino acid utilization rate in multi-stage feed for laying hens, characterized in that, include: The acquisition module is used to acquire the chemical composition values ​​and in vivo intestinal amino acid digestibility values ​​of feed ingredient samples from different growth stages of laying hens. The screening module is used to screen predictive factors based on the measured values ​​of the chemical components and the measured values ​​of the standard ileal amino acid digestibility in vivo. The module also uses the Akaike Information Criterion to analyze the correlation between the nutritional components of feed ingredients and the metabolic characteristics of laying hens during the egg production cycle, and then performs dimensionality reduction mining to obtain a combination of predictive factors. The construction module is used to construct a prediction model based on the combination of prediction factors. By establishing a multiple regression relationship between the amino acid flow characteristics of ileal digesta in laying hens and the combination of prediction factors, candidate prediction equations are obtained. The optimization module is used to optimize the candidate prediction equations by comparing the coefficients of determination and biological rationality of each equation at different growth stages of laying hens, and then selecting the optimal equation. The verification module is used to verify the model according to the optimization equation. It calculates the relative deviation between the predicted value and the in vivo measured value, and constructs a verification mechanism by combining the correction of endogenous amino acid loss in laying hens to obtain the target prediction model. The prediction module is used to predict digestibility based on the target prediction model. By inputting the predictor content of the feed ingredients for laying hens to be tested into the target prediction model, the predicted standard ileal amino acid digestibility value of the current feed formula is obtained.

7. The prediction system for multi-stage feed amino acid utilization in laying hens according to claim 6, characterized in that, The filtering module includes: The first screening unit is used to perform dynamic correlation analysis based on the measured values ​​of the chemical components and the measured values ​​of the standard ileal amino acid digestibility in vivo. By establishing a time series phase synchronization model of egg production rate fluctuations and amino acid metabolic requirements, nutritional factors that respond synchronously to physiological changes in the egg production cycle are identified. The second screening unit is used to perform combination optimization processing based on the nutritional factors. By analyzing the functional correlation strength between factors and constructing modular grouping based on clustering coefficients, the minimum set of factors with functional synergy during the peak egg production period and molting period of laying hens is screened. The third screening unit is used to perform biological verification processing based on the minimum factor set. By establishing a multi-level interaction network model between the ileal microbial community of laying hens and the predictive factors, a combination of predictive factors characterizing the amino acid digestion characteristics of laying hens during the egg-laying stage is obtained.

8. The prediction system for multi-stage feed amino acid utilization in laying hens according to claim 6, characterized in that, The building module includes: The first construction unit is used to perform dynamic modeling processing based on the combination of predictive factors. By introducing the time series phase adjustment of the ileal digestive amino acid flow rate during the laying cycle of laying hens, a primary regression equation based on time dimension correction is constructed. The second building unit is used to perform structural optimization processing based on the primary regression equation, and to establish a dynamic prediction framework based on quantile regression by analyzing the distribution characteristics of amino acid absorption efficiency at different egg-laying stages. The third building block is used to perform biological constraint processing based on the dynamic prediction framework. By integrating the coupling relationship between the ileal microbial metabolic pathway and the amino acid digestion process of laying hens, candidate prediction equations that conform to the digestive physiological characteristics of laying hens are obtained.

9. The prediction system for multi-stage feed amino acid utilization in laying hens according to claim 6, characterized in that, The optimization module includes: The first optimization unit is used to perform multi-stage screening based on the candidate prediction equations. By establishing a decision model based on dynamic programming, the metabolic characteristics of laying hens during peak egg production and molting are analyzed to obtain a preliminary set of screening equations. The second optimization unit is used to perform biological rationality verification processing based on the preliminary set of equations. It analyzes the physiological consistency by constructing a phase-matching model between the laying hen amino acid metabolic pathway and the predicted equations, and obtains the verified set of equations. The third optimization unit is used to perform optimization processing based on the verified set of equations. By introducing a compensation mechanism for the loss of endogenous amino acids in laying hens, an equation correction model is established to obtain the optimized equations.

10. The prediction system for multi-stage feed amino acid utilization in laying hens according to claim 6, characterized in that, The verification module includes: The first verification unit is used to perform deviation distribution analysis based on the optimization equation, and to establish a deviation probability model by statistically analyzing the deviation distribution characteristics between the predicted values ​​and the in vivo measured values ​​at different egg-laying stages, so as to obtain the stage deviation analysis results. The second verification unit is used to perform metabolic compensation processing based on the results of the phased deviation analysis. It dynamically compensates and corrects the metabolic pathways associated with the loss of endogenous amino acids in laying hens and the deviation in digestibility by establishing a network, and obtains the verification indicators after metabolic compensation. The third verification unit is used to perform model convergence determination processing based on the verification index after metabolic compensation, and to optimize the model parameters by setting multi-objective convergence conditions based on biological rationality to obtain the target prediction model.

Images