Royal jelly freshness detection design matching method based on structure diffusion model

Abstract

The invention discloses a royal jelly freshness detection design matching method based on a structure diffusion model, and relates to the technical field of food quality detection. The invention discloses a royal jelly freshness detection design matching method based on a structure diffusion model. The method comprises the steps of sample acquisition, peptide fragment analysis, peptide fragment reverse target searching, binding energy calculation and detection design matching system construction. According to the method, protein and degradation products thereof related to royal jelly freshness can be quickly identified by virtue of a peptide fragment reverse target searching method and a mass spectrometry and database matching technology, so that the time for searching related protein markers is greatly shortened, and the freshness detection efficiency is remarkably improved; by accurately matching protein sequences and identifying degradation products, it is ensured that screened markers are closely related to the degradation process of royal jelly, and therefore scientificity and accuracy of freshness detection are improved; through utilization of royal jelly proteome information, data analysis efficiency and cost effectiveness are improved.

Description

Technical Field

[0001] The invention relates to the technical field of food quality detection, and in particular to a royal jelly freshness detection design and matching method based on a structural diffusion model. Background Art

[0002] Royal jelly is a nutrient secreted by bees. It is widely used in health products, cosmetics and food. It has high nutritional value and biologically active ingredients. However, the composition of royal jelly is easily affected by environmental conditions. Especially during storage, the protein, amino acids and other active substances in royal jelly will degrade or denature, resulting in reduced freshness, thus affecting its use effect. Therefore, how to scientifically and accurately detect the freshness of royal jelly has become an urgent problem to be solved.

[0003] At present, the royal jelly freshness detection method includes invention patent ZL202010049812.0, which provides a monoclonal antibody that recognizes the specific sensitive protein MRJP4 and develops it into an ELISA kit; CN103059135B provides a specific antibody that recognizes MRJP1 and Elisa quantitative detection; CN202110248185.8 provides a biomarker for the freshness of royal jelly and its screening method, and its principle is basically the same as the above two methods. Although these methods can realize quantitative analysis of the freshness of royal jelly compared to non-destructive testing (Raman spectroscopy, infrared spectroscopy), the whole process consumes money and costs more, especially these methods are all pre-treated by enzymatic hydrolysis, and rely on a single labeled peptide segment combined with a database to screen the marker protein, and then prepare monoclonal antibodies through immunized animal mouse experiments. Therefore, it is urgent to develop a rapid detection method and detection system for the freshness of royal jelly. Summary of the invention

[0004] Based on this, this paper proposes a royal jelly freshness detection design matching method based on the structural diffusion model. This method analyzes the protein degradation of royal jelly under different storage conditions, combines mass spectrometry analysis, proteomics database and structural prediction model, and designs a detection design matching method that can accurately identify the freshness of royal jelly, providing a new solution for royal jelly quality control.

[0005] A royal jelly freshness detection design matching method based on a structural diffusion model comprises the following steps:

[0006] Step S1: collecting royal jelly samples under different storage conditions, including royal jelly samples of the same type of fresh royal jelly and its products stored at -20°C, 4°C and 25°C for 0, 1, 3, 6, 12 and 24 months, respectively, mixing each sample with pure water according to a set ratio, extracting, centrifuging and filtering, and obtaining the supernatant of each royal jelly sample;

[0007] Step S2: analyzing the supernatant of the royal jelly samples to obtain the identified peptides in each royal jelly sample, the molecular weight data and the amino acid sequence data of the peptides, and the analysis steps include peptide separation, mass spectrometry analysis, data analysis, spectrum generation and peptide identification;

[0008] Step S3: Perform peptide reverse targeting, obtain royal jelly proteome information from the UniPort database, the proteome information contains protein sequences, match all identified peptides one by one with the royal jelly protein sequence, obtain the sequence site of each peptide, obtain the protein region according to the sequence site of the peptide, calculate the increase rate RPIR of peptides produced by degradation of each protein region, and obtain the relative degradation rate PRDR of each protein by weighting the increase rate of each peptide;

[0009] Step S4: Fit the relative degradation rate of each protein to the storage time of the royal jelly sample to obtain the relative degradation rate equation of each protein at different storage temperatures, expressed as R 2 As an evaluation parameter, a protein degradation database was constructed and the R 2 The values ​​are sorted, and the proteins with the highest ranking are taken as candidate marker proteins;

[0010] Step S5: obtaining the binding antibody information of the candidate marker protein, constructing the three-dimensional structure of the candidate marker protein using the structure prediction model, designing the binding protein based on the binding antibody information using the structure diffusion model, calculating the binding energy of each binding protein with the target marker protein using the Rosetta energy function, and constructing a binding antibody library;

[0011] Step S6: Based on the relative degradation rate equation and combined with antibody information, a royal jelly freshness detection design and matching system is constructed.

[0012] As a preferred technical solution of the present invention, supernatant analysis includes:

[0013] Peptide separation: Peptides are separated by high performance liquid chromatography according to the hydrophobicity of different peptides;

[0014] Mass spectrometry analysis: Perform mass spectrometry analysis on the separated peptides to determine the mass / charge ratio of each peptide and generate a mass spectrum of the peptides;

[0015] Data analysis: Perform data analysis on the mass spectrometry results and compare them with the known protein database to determine the amino acid sequence of the peptide;

[0016] Spectrum generation: Generate a peptide spectrum showing the mass distribution of peptides based on the results of mass spectrometry analysis;

[0017] Peptide identification: Peptide identification is performed based on the average local confidence of the peptides. When the average confidence is higher than 85%, the peptide identification is successful.

[0018] As a preferred technical solution of the present invention, the acquisition of average confidence includes:

[0019] The LDF score is calculated by comparing the similarity between the peptide to be identified and the protein database, the matching quality of the peptide spectrum is evaluated, and the LDF score is converted to a P value. The P value indicates the probability that a score obtained by a false match is greater than the set threshold. The smaller the P value, the higher the matching quality of the peptide spectrum. The P value is converted to -10*log10(P), and the score greater than the set threshold is used as the credible confidence. The proportion of credible confidence is calculated as the average confidence.

[0020] As a preferred technical solution of the present invention, obtaining the relative degradation rate PRDR of protein includes:

[0021] The calculation formula for the increase rate RPIR of peptides produced by degradation in each protein region is: Where Pet represents the content of peptide, Pet p Represents royal jelly samples stored for different periods of time under different conditions, Pet u It means royal jelly sample without any treatment;

[0022] The relative degradation rate PRDR of each protein is calculated as follows: Where n represents the total number of peptides in the target protein; PRIR K The increase rate of peptides generated by degradation of the kth region also indicates the degradation degree of this region of the protein.

[0023] As a preferred technical solution of the present invention, the acquisition of the relative degradation rate equation includes:

[0024] According to the natural degradation characteristics of royal jelly protein, a mathematical model is selected to describe the degradation process of protein over time, where the formula of the exponential model is expressed as: R(t) = Ae ―at ; Where R(t) represents the relative degradation rate at time t, A is the initial degradation rate, and k is the degradation rate constant;

[0025] Using a mathematical fitting method, the relative degradation rate data of each protein is fitted with an exponential model to obtain model parameters, and the parameters are obtained by transforming the model and solving it to obtain a relative degradation rate equation;

[0026] Using R 2 As the evaluation parameter, we can test the model’s fitting effect and select R 2 The mathematical model with the value closest to 1 obtains the relative degradation rate equation; a labeled protein degradation database is constructed, a label is assigned to each protein, and its degradation equation, parameters and evaluation indicators are recorded in the database.

[0027] As a preferred technical solution of the present invention, constructing the three-dimensional structure of the candidate marker protein includes:

[0028] Obtain a structure prediction model. The initial structure prediction model is constructed based on the ESMFold model. The structural information of all royal jelly proteins is obtained and a prediction data set is formed. The prediction data set is used to train the initial structure prediction model, evaluate parameters, and adjust hyperparameters to obtain an optimized structure prediction model. An evaluation layer is added to the optimized structure prediction model. The evaluation layer matches the output result of the optimized structure prediction model with the prediction data set, adds a weight tag to the output result according to the matching result, and outputs the optimized structure prediction model with the added evaluation layer as the final structure prediction model.

[0029] The amino acid sequence of the candidate marker protein is sent to the structure prediction model to obtain the three-dimensional structure of the candidate marker protein.

[0030] As a preferred technical solution of the present invention, the design of the binding protein includes:

[0031] Obtaining the binding site information and target antigen information in the three-dimensional structure of the marker protein, using the structural diffusion model to obtain the initial antibody structure that matches the target antigen information from random noise, and evolving the initial antibody structure to generate an antibody structure that matches the binding site information in the three-dimensional structure of the marker protein;

[0032] Evolution is guided by the binding site information of the target protein, making the generated antibodies specific. By fine-tuning the binding interface between the antibody and the target protein, the antibody binding site is highly matched with the geometry and chemical properties of the marker protein.

[0033] The predicted binding energies of all generated antibody structures were obtained, and the three antibody structures with the lowest predicted binding energies were selected as binding proteins.

[0034] As a preferred technical solution of the present invention, obtaining the binding energy includes:

[0035] For the antibody structure generated by the structure prediction model, the antibody structure is matched with the antibody structure in the binding energy prediction database to obtain the binding energy preset in the binding energy prediction database for the corresponding antibody structure;

[0036] For binding proteins, the Rosetta energy function is used to calculate the binding energy of each binding protein with the target marker protein, including: obtaining the three-dimensional structure of the binding protein and the marker protein, and optimizing the structure; constructing the binding protein and the marker protein into a complex; calculating the binding energy of the complex, by separating the two parts of the complex, and calculating the individual energies of the binding protein and the marker protein respectively; calculating the total energy of the complex; the binding energy is the total energy of the complex minus the energy of the binding protein and the energy of the marker protein, optimizing the structure of the complex, and screening the structure of the complex with the lowest binding energy.

[0037] As a preferred technical solution of the present invention, the royal jelly freshness detection design matching system includes:

[0038] Input module, used for input module refers to inputting target peptide or target protein as search condition;

[0039] The acquisition module is used to obtain the relative degradation rate equation and the associated R based on the input target peptide or marker protein. 2 Values ​​and binding antibody libraries;

[0040] The recommendation module is used to screen out antibodies with the best binding energy based on the binding energy according to the antibodies in the obtained binding antibody library, and provide their binding energy values ​​and structural information.

[0041] The present invention has the following advantages:

[0042] 1. The present invention uses a peptide reverse targeting method, mass spectrometry analysis and database matching technology to quickly identify proteins and degradation products related to the freshness of royal jelly, greatly shorten the time for finding related protein markers, and significantly improve the efficiency of freshness detection; by accurately matching protein sequences and identifying degradation products, it is ensured that the screened markers are closely related to the degradation process of royal jelly, thereby improving the scientificity and accuracy of freshness detection; by utilizing royal jelly proteomic information, unnecessary experimental duplication is avoided, and the efficiency and cost-effectiveness of data analysis are improved.

[0043] 2. The present invention establishes a protein relative degradation rate data set and combines it with a structural diffusion model to develop a complete royal jelly freshness detection design and matching system. The system can not only calculate the relative degradation rate of each protein in royal jelly, but also use structure prediction and energy calculation models to design antibodies that are highly matched with the target protein, thereby achieving accurate freshness detection; based on the relative degradation rate data set, the system can comprehensively cover the degradation rules of royal jelly under different storage conditions, ensuring the comprehensiveness of the freshness detection results.

[0044] 3. The present invention simulates the natural structural evolution process of proteins through a structural diffusion model, generates biologically meaningful three-dimensional protein structures from initial random noise, and accurately predicts the three-dimensional structure of target proteins, especially for the complex protein degradation products in royal jelly, thereby improving the accuracy of structural prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 A flow chart of a royal jelly freshness detection design matching method based on a structural diffusion model of the present invention;

[0046] Figure 2 To visualize the degradation profile of the first two marker proteins associated with royal jelly freshness based on reverse targeting technology;

[0047] Figure 3 To visualize the degradation profile of the 3rd and 4th marker proteins associated with royal jelly freshness based on reverse targeting technology;

[0048] Figure 4 To visualize the degradation map of the fifth marker protein associated with royal jelly freshness based on reverse targeting technology;

[0049] Figure 5 This is the fitting diagram of the first two marker proteins highly associated with royal jelly freshness obtained based on reverse targeting technology;

[0050] Figure 6 The fitting diagrams of the third and fourth marker proteins highly associated with royal jelly freshness obtained based on reverse targeting technology;

[0051] Figure 7 This is the fitting diagram of the fifth marker protein highly associated with royal jelly freshness obtained based on reverse targeting technology;

[0052] Figure 8 To design the structural diagram of the antibody binding to the marker protein highly associated with royal jelly freshness based on the structure-diffusion model. DETAILED DESCRIPTION

[0053] In order to enable persons skilled in the art to better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0054] Example, a royal jelly freshness detection design matching method based on structural diffusion model, see Figures 1 to 8 As shown, the following steps are included:

[0055] Step S1: collecting royal jelly samples under different storage conditions, including royal jelly samples of the same type of fresh royal jelly and its products stored at -20°C, 4°C and 25°C for 0, 1, 3, 6, 12 and 24 months, respectively, mixing each sample with pure water according to a set ratio, extracting, centrifuging and filtering, and obtaining the supernatant of each royal jelly sample;

[0056] The royal jelly supernatant is filtered and purified through a special sieve to remove impurities and insoluble components;

[0057] Step S2: analyzing the supernatant of the royal jelly samples to obtain the identified peptides in each royal jelly sample, the molecular weight data and the amino acid sequence data of the peptides, and the analysis steps include peptide separation, mass spectrometry analysis, data analysis, spectrum generation and peptide identification;

[0058] Supernatant analysis includes:

[0059] Peptide separation: Peptides are separated by high performance liquid chromatography (HPIC) according to the hydrophobicity of different peptides;

[0060] Separation of peptides by high performance liquid chromatography (HPLC) includes:

[0061] Select an appropriate chromatographic column: select an appropriate chromatographic column according to the properties of the peptide, such as a reversed phase C18 column, which is suitable for the separation of most hydrophobic peptides; prepare the mobile phase: prepare a mobile phase containing an appropriate proportion of organic solvent and water, and use a mixed solution of acetonitrile-water or methanol-water, in which the proportion of acetonitrile can be gradually increased from 5% to 70%; set up a gradient elution program: set a time gradient so that the proportion of organic solvent gradually increases during the elution process, and separate the peptides according to their hydrophobicity differences; optimize the flow rate and temperature: select an appropriate flow rate (usually 1-2 ml / min) and a suitable temperature to obtain the best resolution; collect fractions: collect the eluted fractions at the desired intervals, and mix the fractions at different time periods for subsequent mass spectrometry analysis;

[0062] Mass spectrometry analysis: Perform mass spectrometry analysis on the separated peptides to determine the mass / charge ratio of each peptide and generate a mass spectrum of the peptides. Since different peptides have different amino acid compositions and sequences, their m / z values ​​in the mass spectra will also be different. By comparing the m / z values ​​obtained in the experiment with the theoretical m / z values ​​of known proteins in the database, the peptides in the sample and the proteins from which they originated can be identified. Electrospray ionization (ESI) and matrix-assisted laser desorption / ionization (MALDI) can be used for this purpose.

[0063] Data analysis: The mass spectrometry results were analyzed and compared with the known protein database to determine the amino acid sequence of the peptides. Peak Studio 8.0 was used to analyze the molecular weight and amino acid sequence of the obtained peptide data during the implementation process.

[0064] Spectrum generation: Generate a peptide spectrum showing the mass distribution of peptides based on the results of mass spectrometry analysis; used to infer the primary structure of the protein;

[0065] Peptide identification: Peptide identification is performed based on the average local confidence of the peptides. When the average confidence is higher than 85%, the peptide identification is successful.

[0066] The acquisition of average confidence includes:

[0067] The LDF score is calculated by comparing the similarity between the peptide to be identified and the protein database, the matching quality of the peptide spectrum is evaluated, and the LDF score is converted to a P value. The P value indicates the probability that a score obtained by a false match is greater than the set threshold. The smaller the P value, the higher the matching quality of the peptide spectrum. The P value is converted to -10*log10(P), and the score greater than the set threshold is used as the credible confidence. The proportion of credible confidence is calculated as the average confidence.

[0068] Step S3: Perform peptide reverse targeting, obtain royal jelly proteome information from the UniPort database, the proteome information contains protein sequences, match all identified peptides one by one with the royal jelly protein sequence, obtain the sequence site of each peptide, obtain the protein region according to the sequence site of the peptide, calculate the increase rate RPIR of peptides produced by degradation of each protein region, and obtain the relative degradation rate PRDR of each protein by weighting the increase rate of each peptide;

[0069] The acquisition of protein relative degradation rate PRDR includes:

[0070] The calculation formula for the increase rate RPIR of peptides produced by degradation in each protein region is: Where Pet represents the content of peptide, Pet p Represents royal jelly samples stored for different periods of time under different conditions, Pet uIt means royal jelly sample without any treatment;

[0071] The relative degradation rate PRDR of each protein is calculated as follows: Where n represents the total number of peptides in the target protein; PRIR K The increase rate of peptides generated by degradation of the kth region also indicates the degradation degree of this region of the protein.

[0072] Step S4: Fit the relative degradation rate of each protein to the storage time of the royal jelly sample to obtain the relative degradation rate equation of each protein at different storage temperatures, expressed as R 2 As an evaluation parameter, a protein degradation database was constructed and the R 2 The top five proteins were sorted by their values, and the proteins with the highest ranking were used as candidate marker proteins. The degradation graphs and fitting graphs of the top five candidate marker proteins stored under different conditions were shown in the following figure. Figure 2 and Figure 3 As shown;

[0073] The relative degradation rate equation is obtained by:

[0074] According to the natural degradation characteristics of royal jelly protein, a mathematical model is selected to describe the degradation process of protein over time, where the formula of the exponential model is expressed as: R(t) = Ae ―at ; Where R(t) represents the relative degradation rate at time t, A is the initial degradation rate, and k is the degradation rate constant;

[0075] Using a mathematical fitting method, the relative degradation rate data of each protein is fitted with an exponential model to obtain model parameters, and the parameters are obtained by transforming the model and solving it to obtain a relative degradation rate equation;

[0076] For example, the data obtained by performing linear regression fitting on the exponential model and logarithmic transformation are: lnR = β0 + β1t; where β0 = lnA, β1 = -a; the values ​​of β0 and β1 are obtained through linear regression analysis, and the values ​​of the parameters A and a of the exponential model are calculated;

[0077] Using R 2 As the evaluation parameter, we can test the model’s fitting effect and select R 2 The mathematical model with the value closest to 1 obtains the relative degradation rate equation; a labeled protein degradation database is constructed, a label is assigned to each protein, and its degradation equation, parameters and evaluation indicators are recorded in the database.

[0078] Step S5: Obtain the binding antibody information of the candidate marker protein, and use the structure prediction model to construct the three-dimensional structure of the candidate marker protein. The three-dimensional structure of the candidate marker protein is as follows: Figure 4As shown; the structural diffusion model is used to design binding proteins based on the binding antibody information, and the Rosetta energy function is used to calculate the binding energy of each binding protein with the target marker protein to construct a binding antibody library; the binding energy of a target marker protein in the binding antibody library with the corresponding binding protein is shown in Table 1;

[0079] Table 1 Binding energy of target proteins highly related to royal jelly freshness and binding proteins

[0080]

[0081] Constructing the three-dimensional structure of candidate marker proteins includes:

[0082] Obtaining a structure prediction model: The initial structure prediction model is built based on the ESMFold model to obtain the structural information of all royal jelly proteins and form a predicted structure data set;

[0083] The structure prediction is based on the protein's FASTA format sequence, which is uploaded to ESMfold for structure prediction and outputs one or more PDB files containing the predicted three-dimensional structure of the protein;

[0084] The design of binding proteins includes:

[0085] Obtain the binding site information and target antigen information in the three-dimensional structure of the marker protein, use the structural diffusion model to obtain the initial antibody structure that matches the target antigen information from random noise, and use ProteinMPNN to optimize and design the initial antibody structure. At the same time, generate an antibody structure complex that matches the three-dimensional structure of the marker protein based on ESMfold;

[0086] The predicted binding energies of all generated antibody structures were obtained, and the three antibody structures with the lowest predicted binding energies were selected as candidate binding antibody proteins.

[0087] The specific steps of the implementation process include:

[0088] Generation of initial structure: The diffusion model based on RFdiffusion is usually used to generate binding protein structures from scratch. First, the diffusion model starts with a random noise structure and then gradually evolves it into a three-dimensional protein structure that conforms to the laws of physicochemistry. The model generates an antibody structure that matches the target protein based on the binding site information of the target protein. Through the reverse diffusion process, a reasonable structure is gradually constructed from the random distribution, and finally an antibody structure that matches the target antigen is generated.

[0089] Evaluation of binding energy: The generated antibody structure is further evaluated by calculating the binding energy, which is a key parameter to measure the binding stability between the antibody and the target protein. The structural diffusion model can screen multiple generated antibody structures and select those with the lowest binding energy for subsequent verification; predict which antibody structures are most suitable for binding to the marker protein, thus providing a basis for subsequent antibody engineering;

[0090] Structural validation and iteration: Use computational tools (such as Rosetta, etc.) to further validate and optimize the generated antibody structure. If the generated structure fails to achieve the expected binding effect, the diffusion model can be used again to regenerate an improved antibody structure by adjusting the guide signal or input data;

[0091] The acquisition of binding energy includes:

[0092] For the antibody structure generated by the structure prediction model, match it with the antibody structure in the binding energy prediction database to obtain the binding energy preset in the binding energy prediction database for the corresponding antibody structure;

[0093] For binding proteins, the Rosetta energy function is used to calculate the binding energy of each binding protein with the target marker protein, including: obtaining the three-dimensional structure of the binding protein and the marker protein, and optimizing the structure; constructing the binding protein and the marker protein into a complex; calculating the binding energy of the complex, by separating the two parts of the complex, and calculating the individual energies of the binding protein and the marker protein respectively; calculating the total energy of the complex; the binding energy is the total energy of the complex minus the energy of the binding protein and the energy of the marker protein, optimizing the structure of the complex, and screening the structure of the complex with the lowest binding energy.

[0094] Step S6: Based on the relative degradation rate equation and combined with antibody information, a royal jelly freshness detection design and matching system is constructed.

[0095] Royal jelly freshness detection design and matching system includes:

[0096] Input module, used for input module refers to inputting target peptide or target protein as search condition;

[0097] The acquisition module is used to obtain the relative degradation rate equation and the associated R based on the input target peptide or marker protein. 2 Values ​​and binding antibody libraries;

[0098] The recommendation module is used to screen out antibodies with the best binding energy based on the binding energy according to the antibodies in the obtained binding antibody library, and provide their binding energy values ​​and structural information.

[0099] It should be understood that those skilled in the art can make improvements or changes based on the above description, and all such improvements and changes should fall within the scope of protection of the appended claims of the present invention. Parts not described in detail in this specification belong to the prior art known to those skilled in the art.

Claims

1. A royal jelly freshness detection design and matching method based on a structural diffusion model, characterized in that: The following steps are involved: Step S1: collecting royal jelly samples under different storage conditions, including royal jelly samples of the same type of fresh royal jelly and its products stored at -20°C, 4°C and 25°C for 0, 1, 3, 6, 12 and 24 months, respectively, mixing each sample with pure water according to a set ratio, extracting, centrifuging and filtering, and obtaining the supernatant of each royal jelly sample; Step S2: analyzing the supernatant of the royal jelly samples to obtain the identified peptides in each royal jelly sample, the molecular weight data and the amino acid sequence data of the peptides, and the analysis steps include peptide separation, mass spectrometry analysis, data analysis, spectrum generation and peptide identification; Step S3: Perform peptide reverse targeting, obtain royal jelly proteome information from the UniPort database, the proteome information contains protein sequences, match all identified peptides one by one with the royal jelly protein sequence, obtain the sequence site of each peptide, obtain the protein region according to the sequence site of the peptide, calculate the increase rate RPIR of peptides produced by degradation of each protein region, and obtain the relative degradation rate PRDR of each protein by weighting the increase rate of each peptide; Step S4: Fit the relative degradation rate of each protein to the storage time of the royal jelly sample to obtain the relative degradation rate equation of each protein at different storage temperatures, expressed as R 2 As an evaluation parameter, a protein degradation database was constructed and the R 2 The values ​​are sorted, and the proteins with the highest ranking are taken as candidate marker proteins; Step S5: obtaining the binding antibody information of the candidate marker protein, constructing the three-dimensional structure of the candidate marker protein using the structure prediction model, designing the binding protein based on the binding antibody information using the structure diffusion model, calculating the binding energy of each binding protein with the target marker protein using the Rosetta energy function, and constructing a binding antibody library; Step S6: Based on the relative degradation rate equation and combined with antibody information, a royal jelly freshness detection design and matching system is constructed.

2. A royal jelly freshness detection design and matching method based on a structural diffusion model according to claim 1, characterized in that: Supernatant analysis includes: Peptide separation: Peptides are separated by high performance liquid chromatography according to the hydrophobicity of different peptides; Mass spectrometry analysis: Perform mass spectrometry analysis on the separated peptides to determine the mass / charge ratio of each peptide and generate a mass spectrum of the peptides; Data analysis: Perform data analysis on the mass spectrometry results and compare them with the known protein database to determine the amino acid sequence of the peptide; Spectrum generation: Generate a peptide spectrum showing the mass distribution of peptides based on the results of mass spectrometry analysis; Peptide identification: Peptide identification is performed based on the average local confidence of the peptides. When the average confidence is higher than 85%, the peptide identification is successful.

3. A royal jelly freshness detection design and matching method based on a structural diffusion model according to claim 2, characterized in that: The acquisition of average confidence includes: The LDF score is calculated by comparing the similarity between the peptide to be identified and the protein database, the matching quality of the peptide spectrum is evaluated, and the LDF score is converted to a P value. The P value indicates the probability that a score obtained by a false match is greater than the set threshold. The smaller the P value, the higher the matching quality of the peptide spectrum. The P value is converted to -10*log10(P), and the score greater than the set threshold is used as the credible confidence. The proportion of credible confidence is calculated as the average confidence.

4. A royal jelly freshness detection design and matching method based on a structural diffusion model according to claim 1, characterized in that: The acquisition of protein relative degradation rate PRDR includes: The calculation formula for the increase rate RPIR of peptides produced by degradation in each protein region is: Where Pet represents the content of peptide, Pet p Represents royal jelly samples stored for different periods of time under different conditions, Pet u It means royal jelly sample without any treatment; The relative degradation rate PRDR of each protein is calculated as follows: Where n represents the total number of peptides in the target protein; PRIR K The increase rate of peptides generated by degradation of the kth region also indicates the degradation degree of this region of the protein.

5. A royal jelly freshness detection design and matching method based on a structural diffusion model according to claim 1, characterized in that: The relative degradation rate equation is obtained by: According to the natural degradation characteristics of royal jelly protein, a mathematical model is selected to describe the degradation process of protein over time, where the formula of the exponential model is expressed as: R(t) = Ae ―at ; Where R(t) represents the relative degradation rate at time t, A is the initial degradation rate, and k is the degradation rate constant; Using a mathematical fitting method, the relative degradation rate data of each protein is fitted with an exponential model to obtain model parameters, and the parameters are obtained by transforming the model and solving it to obtain a relative degradation rate equation; Using R 2 As the evaluation parameter, we can test the model’s fitting effect and select R 2 The mathematical model with the value closest to 1 obtains the relative degradation rate equation; a labeled protein degradation database is constructed, a label is assigned to each protein, and its degradation equation, parameters and evaluation indicators are recorded in the database.

6. A royal jelly freshness detection design and matching method based on a structural diffusion model according to claim 1, characterized in that: Constructing the three-dimensional structure of candidate marker proteins includes: Obtain a structure prediction model. The initial structure prediction model is constructed based on the ESMFold model. The structural information of all royal jelly proteins is obtained and a prediction data set is formed. The prediction data set is used to train the initial structure prediction model, evaluate parameters, and adjust hyperparameters to obtain an optimized structure prediction model. An evaluation layer is added to the optimized structure prediction model. The evaluation layer matches the output result of the optimized structure prediction model with the prediction data set, adds a weight tag to the output result according to the matching result, and outputs the optimized structure prediction model with the added evaluation layer as the final structure prediction model. The amino acid sequence of the candidate marker protein is sent to the structure prediction model to obtain the three-dimensional structure of the candidate marker protein.

7. A royal jelly freshness detection design and matching method based on a structural diffusion model according to claim 1, characterized in that: The design of binding proteins includes: Obtaining the binding site information and target antigen information in the three-dimensional structure of the marker protein, using the structural diffusion model to obtain the initial antibody structure that matches the target antigen information from random noise, and evolving the initial antibody structure to generate an antibody structure that matches the binding site information in the three-dimensional structure of the marker protein; Evolution is guided by the binding site information of the target protein, making the generated antibodies specific. By fine-tuning the binding interface between the antibody and the target protein, the antibody binding site is highly matched with the geometry and chemical properties of the marker protein. The predicted binding energies of all generated antibody structures were obtained, and the three antibody structures with the lowest predicted binding energies were selected as binding proteins.

8. A royal jelly freshness detection design and matching method based on a structural diffusion model according to claim 7, characterized in that: The acquisition of binding energy includes: For the antibody structure generated by the structure prediction model, match it with the antibody structure in the binding energy prediction database to obtain the binding energy preset in the binding energy prediction database for the corresponding antibody structure; For binding proteins, the Rosetta energy function is used to calculate the binding energy of each binding protein with the target marker protein, including: obtaining the three-dimensional structure of the binding protein and the marker protein, and optimizing the structure; constructing the binding protein and the marker protein into a complex; calculating the binding energy of the complex, by separating the two parts of the complex, and calculating the individual energies of the binding protein and the marker protein respectively; calculating the total energy of the complex; the binding energy is the total energy of the complex minus the energy of the binding protein and the energy of the marker protein, optimizing the structure of the complex, and screening the structure of the complex with the lowest binding energy.

9. A royal jelly freshness detection design and matching method based on a structural diffusion model according to claim 1, characterized in that: Royal jelly freshness detection design and matching system includes: Input module, used for input module refers to inputting target peptide or target protein as search condition; The acquisition module is used to obtain the relative degradation rate equation and the associated R based on the input target peptide or marker protein. 2 Values ​​and binding antibody libraries; The recommendation module is used to screen out antibodies with the best binding energy based on the binding energy according to the antibodies in the obtained binding antibody library, and provide their binding energy values ​​and structural information.

Images