Label-free-based anti-aging polypeptide detection method and system

Abstract

The invention discloses an anti-aging polypeptide detection method and system based on Label-free. The method comprises the following steps: extracting protein in a sample for pretreatment; the method comprises the following steps: carrying out enzymolysis on a polypeptide sample stored in an acetic acid aqueous solution, and cutting protein into peptide fragments by using trypsin; carrying out mass spectrometry detection and analysis by using an improved Label-free technology; and performing layered search on the mass spectrum parameters by using an Ivy optimization algorithm to obtain a detection result of the anti-aging polypeptide. According to the invention, the accuracy and sensitivity of detection are improved, the cost is reduced, the interference of a marking agent is avoided, and a scientific basis is provided for research and development of anti-aging drugs and identification of biomarkers.

Description

Technical Field

[0001] The present invention relates to the field of biological detection technology, and in particular to a label-free anti-aging polypeptide detection method and system. Background Art

[0002] As functional fragments of proteins, peptides play a crucial role in drug development due to their small molecular weight and high biological activity. As research into peptide drugs deepens, accurately measuring and quantitatively analyzing peptide content and changes has become crucial, and this also applies to the study of anti-aging peptides.

[0003] Traditional peptide quantification methods often rely on labeling agents, such as isotope labeling or fluorescent labeling. Although these methods are mature, they have problems such as interference from labeling agents and high costs. Summary of the Invention

[0004] Purpose of the Invention: To address the aforementioned issues, the present invention provides a label-free anti-aging peptide detection method and system, enabling high-throughput, label-free quantitative analysis of anti-aging peptides. This technology directly compares the mass spectral signatures of peptides in a sample through mass spectrometry, aiming to improve detection accuracy and sensitivity while reducing costs and avoiding interference from labeling agents, providing a scientific basis for anti-aging drug development and biomarker identification.

[0005] Technical solution: A label-free anti-aging peptide detection method, comprising the following steps:

[0006] (1) Extract proteins from samples and perform pretreatment;

[0007] (2) enzymatically hydrolyzing the polypeptide sample stored in acetic acid aqueous solution, using trypsin to cut the protein into peptide fragments;

[0008] (3) Use improved label-free technology for mass spectrometry detection and analysis;

[0009] (4) The Ivy optimization algorithm was used to perform a hierarchical search on the mass spectrometry parameters to obtain the detection results of anti-aging peptides.

[0010] Furthermore, the pretreatment includes using a lysis buffer to lyse the sample, disrupting the cells by ultrasonication in an ice bath and removing insoluble matter by centrifugation, then removing high-abundance proteins, and storing the extracted polypeptides in an acetic acid aqueous solution.

[0011] Furthermore, the removal of high-abundance proteins is targeted at blood samples, and a multiple affinity removal column is required to specifically remove high-abundance proteins such as albumin and IgG.

[0012] Furthermore, the improved label-free technology includes peptide digestion, liquid chromatography separation, mass spectrometer ionization, mass spectrometry analysis, relative content inference, and statistical analysis; a generative adversarial network is introduced into the data preprocessing stage of the mass spectrometry analysis, in which the generator simulates the distribution of high-confidence peptide signals, and the discriminator evaluates whether the signal is a real signal based on the matching degree of ion fragments, effectively solving the problem of false positive peak identification in low signal-to-noise ratio scenarios and reducing the error rate of overlapping peak resolution.

[0013] A one-dimensional convolutional neural network model is introduced into the relative content inference, which suppresses background noise through dynamic weight allocation, significantly improving the detection sensitivity and quantitative stability of low-abundance peptides, and forming a fully automated process from signal analysis to content calculation.

[0014] Furthermore, the mass spectrometry detection and analysis using the improved Label-free technology includes:

[0015] Let the protein sequence be a string S = {s1,s2,...,s m}, the protease cleavage rule is to recognize a specific amino acid pattern, and define the cleavage site set C = {c j}, where c j is the probability of cleavage at position j in the amino acid sequence, expressed as:

[0016]

[0017] Where k is the index number of the characteristic function, indicating different chemical characteristic types; f k refers to the characteristic function based on the chemical characteristics of local amino acid residues; w k refers to the protease specificity parameter; σ is the Sigmoid function;

[0018] The peptide set after cleavage P = {p i}, of length n, satisfying:

[0019]

[0020] The migration rate of a peptide in a chromatographic column is determined by its partition coefficient between the mobile phase and the stationary phase, and its mathematical expression is as follows:

[0021]

[0022] Where C(x,t) refers to the concentration of the peptide at position x and time t on the chromatographic column; D refers to the diffusion coefficient, and v refers to the flow rate;

[0023] The peptide ionization efficiency η is related to the charge z and the desolvation energy ΔG, and its mathematical expression is as follows:

[0024]

[0025] Where β refers to the mass spectrometer parameter, ΔG0 is the threshold energy;

[0026] Generative adversarial networks are introduced in mass spectrometry analysis. By building a game framework between the generator (G) and the discriminator (D), the simulation and noise suppression of the real mass spectrometry signal distribution are achieved. The generator uses a random noise vector z~p noise As input, generate simulated mass spectrometry signal Its goal is to approximate the real peptide signal distribution preal(x); the discriminator evaluates the biological plausibility of the input signal by analyzing the ion fragment matching degree, and outputs the probability D(x)∈[0,1] that the input signal belongs to the real data. The generator and the discriminator are jointly optimized through the minimization-maximization game, and the loss function is defined as:

[0027]

[0028] The generator must simultaneously satisfy the loss function objective and the fragment matching constraint. The fragment matching S, as a core biophysical indicator, quantifies the consistency between the generated signal and the real signal in the ion fragmentation pattern. Its calculation formula is modified to the normalized dot product form:

[0029]

[0030] Where I real (m / z) and I gen (m / z) represents the ion intensity of the true signal and the generated signal at the mass-to-charge ratio m / z, respectively. The denominator is normalized by the L2 norm to eliminate the influence of the difference in the total signal amount. When S→1, it indicates that the generated signal is highly consistent with the experimentally observed peptide fragmentation pattern. Conversely, S→0 indicates that the generated signal deviates from the true biological law. This method significantly improves the reliability of peptide identification under low signal-to-noise ratio conditions by integrating adversarial learning and domain knowledge constraints.

[0031] A one-dimensional convolutional neural network model is introduced in the relative content inference, using the mass spectrometry signal sequence X∈R L As input, the convolution kernel W∈R with adaptive capability is used k Perform feature extraction; R L represents an L-dimensional real vector, R k represents a real vector of k dimensions;

[0032] The relative content inference model based on one-dimensional convolutional neural network achieves high sensitivity analysis of peptide signals through dynamic weight convolution mechanism. Different from traditional static convolution, its core innovation lies in the introduction of time-dependent dynamic weight adjustment mechanism. The model outputs the eigenvalue y t The specific operation process is expressed as:

[0033]

[0034] where w i (t) represents the dynamic weight of the i-th convolution kernel position at time step t, x t+i-1 is the ion current intensity detected by the instrument, and b is the bias term; the weight is generated in real time through the attention mechanism, which is implemented as follows: for each time step t, the query vector Q is generated through linear transformation t ∈R d , and define a learnable key vector K for each position i of the convolution kernel i ∈R d ; R d Represents a d-dimensional real vector; the weight calculation uses the scaled dot product attention form:

[0035]

[0036] This design enables the convolution kernel to dynamically adjust the receptive field weight according to the local contextual features of the input signal. Compared with the traditional convolution operation with fixed weights, it significantly improves the adaptability to the non-uniform noise distribution and low-abundance peptide signals in the mass spectrometry signal. In terms of quantitative stability optimization, the model introduces an abundance variance penalty term based on the mean square error loss function to form a composite optimization objective function:

[0037]

[0038] Where λ is the balancing hyperparameter, is the output value of the convolutional neural network, y is the true value of the polypeptide content, Var() represents the variance function, and the variance term By constraining the fluctuation of abundance estimates between adjacent scan points, the oscillation of quantitative results caused by random noise is effectively suppressed. This objective function can be used to determine the accuracy of the method, providing reliable support for subsequent research.

[0039] Furthermore, the Ivy optimization algorithm improves the species growth rate of the traditional Ivy algorithm. The improved species growth rate formula is:

[0040]

[0041] Where r t+1 represents the updated position vector of individual t; r t Represents the vector before the updated position of the current individual t; x t is the position vector of individual t in the search space. g best is the historical optimal solution of all individuals in the current population; θ is the random walk weight coefficient; ∈ tis the fitness value of the individual optimal solution of the current individual i and γ is the global attraction weight coefficient.

[0042] A label-free anti-aging peptide detection system, comprising a sample processor, a mass spectrometer, and a search detector;

[0043] The sample processor extracts proteins from the sample for pretreatment; enzymatically hydrolyzes the polypeptide sample stored in an acetic acid aqueous solution, and uses trypsin to cut the protein into peptide segments;

[0044] The mass spectrometer uses an improved Label-free technology to perform mass spectrometry detection and analysis;

[0045] The search detector uses the Ivy optimization algorithm to perform a hierarchical search on mass spectrometry parameters to obtain the detection results of the anti-aging polypeptide.

[0046] Beneficial effects: (1) A generative adversarial network is introduced into the data preprocessing stage of label-free mass spectrometry analysis. The generator simulates the distribution of high-confidence peptide signals, and the discriminator evaluates whether the signal is a real signal based on the matching degree of ion fragments, effectively solving the problem of false positive peak recognition in low signal-to-noise ratio scenarios and reducing the error rate of overlapping peak analysis.

[0047] (2) A one-dimensional convolutional neural network model is introduced into the relative content inference of label-free technology. The background noise is suppressed through dynamic weight allocation, which significantly improves the detection sensitivity and quantitative stability of low-abundance peptides, forming a fully automated process from signal analysis to content calculation.

[0048] (3) The Ivy optimization algorithm improves the species growth rate formula. This improved method introduces the attraction field of the global optimal solution, so that individuals can be attracted by the global optimal solution while randomly walking, thereby balancing randomness and directionality to a certain extent and improving the global search ability and convergence efficiency of the algorithm. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 Flow chart of the method of the present invention;

[0050] Figure 2 This is the schematic diagram of the GAN model;

[0051] Figure 3 Flowchart of the lvy optimization algorithm. DETAILED DESCRIPTION

[0052] The technical solution of the present invention will be further described below with reference to the accompanying drawings.

[0053] like Figure 1 As shown, the present invention provides a technical solution:

[0054] A label-free method for detecting anti-aging peptides comprises the following steps:

[0055] (1) Extract proteins from samples and perform pretreatment;

[0056] (2) enzymatically hydrolyzing the polypeptide sample stored in acetic acid aqueous solution, using trypsin to cut the protein into peptide fragments;

[0057] (3) Use improved label-free technology for mass spectrometry detection and analysis;

[0058] (4) The Ivy optimization algorithm was used to perform a hierarchical search on the mass spectrometry parameters to obtain the detection results of anti-aging peptides.

[0059] Furthermore, the pretreatment process includes sample lysis using a lysis buffer, cell disruption by ultrasonication in an ice bath, and centrifugation to remove insoluble matter. High-abundance proteins are then removed, and the extracted peptides are stored in an aqueous acetic acid solution. This pretreatment process effectively removes interfering substances and provides a high-quality protein substrate for label-free mass spectrometry detection.

[0060] Furthermore, the removal of high-abundance proteins is targeted at blood samples, and a multiple affinity removal column is required to specifically remove high-abundance proteins such as albumin and IgG.

[0061] Furthermore, the improved label-free technology includes peptide digestion, liquid chromatography separation, mass spectrometer ionization, mass spectrometry analysis, relative content inference, and statistical analysis; a generative adversarial network is introduced into the data preprocessing stage of the mass spectrometry analysis, in which the generator simulates the distribution of high-confidence peptide signals, and the discriminator evaluates whether the signal is a real signal based on the matching degree of ion fragments, effectively solving the problem of false positive peak identification in low signal-to-noise ratio scenarios and reducing the error rate of overlapping peak resolution.

[0062] A one-dimensional convolutional neural network model is introduced into the relative content inference, which suppresses background noise through dynamic weight allocation, significantly improving the detection sensitivity and quantitative stability of low-abundance peptides, and forming a fully automated process from signal analysis to content calculation.

[0063] Furthermore, the mass spectrometry detection and analysis using the improved Label-free technology includes:

[0064] Let the protein sequence be a string S = {s1,s2,...,s m}, the protease cleavage rule is to recognize a specific amino acid pattern, and define the cleavage site set C = {c j}, where c j is the probability of cleavage at position j in the amino acid sequence, expressed as:

[0065]

[0066] Where k is the index number of the characteristic function, indicating different chemical characteristic types; f k Refers to the characteristic function based on the chemical characteristics of local amino acid residues; w k refers to the protease specificity parameter; σ is the Sigmoid function;

[0067] The peptide set after cleavage P = {p i}, of length n, satisfying:

[0068]

[0069] The migration rate of a peptide in a chromatographic column is determined by its partition coefficient between the mobile phase and the stationary phase, and its mathematical expression is as follows:

[0070]

[0071] Where C(x,t) refers to the concentration of the peptide at position x and time t on the chromatographic column; D refers to the diffusion coefficient, and v refers to the flow rate;

[0072] The peptide ionization efficiency η is related to the charge z and the desolvation energy ΔG, and its mathematical expression is as follows:

[0073]

[0074] Where β refers to the mass spectrometer parameter, ΔG0 is the threshold energy;

[0075] Generative adversarial networks are introduced into mass spectrometry analysis. By building a game framework between the generator and the discriminator, the simulation and noise suppression of the real mass spectrometry signal distribution are achieved. The generator uses a random noise vector z~p noise As input, generate simulated mass spectrometry signal Its goal is to approximate the real peptide signal distribution preal(x); the discriminator evaluates the biological plausibility of the input signal by analyzing the ion fragment matching degree, and outputs the probability D(x)∈[0,1] that the input signal belongs to the real data. The generator and the discriminator are jointly optimized through the minimization-maximization game, and the loss function is defined as:

[0076]

[0077] The generator must simultaneously satisfy the loss function objective and the fragment matching constraint. The fragment matching S, as a core biophysical indicator, quantifies the consistency between the generated signal and the real signal in the ion fragmentation pattern. Its calculation formula is modified to the normalized dot product form:

[0078]

[0079] Where I real (m / z) and I gen (m / z) represents the ion intensity of the true signal and the generated signal at the mass-to-charge ratio m / z, respectively. The denominator is normalized by the L2 norm to eliminate the influence of the difference in the total signal amount. When S→1, it indicates that the generated signal is highly consistent with the experimentally observed peptide fragmentation pattern. Conversely, S→0 indicates that the generated signal deviates from the true biological law. This method significantly improves the reliability of peptide identification under low signal-to-noise ratio conditions by integrating adversarial learning and domain knowledge constraints.

[0080] A one-dimensional convolutional neural network model is introduced in the relative content inference, using the mass spectrometry signal sequence X∈R L As input, the convolution kernel W∈R with adaptive capability is used k Perform feature extraction; R L represents an L-dimensional real vector, R k represents a real vector of k dimensions;

[0081] The relative content inference model based on one-dimensional convolutional neural network achieves high sensitivity analysis of peptide signals through dynamic weight convolution mechanism. Different from traditional static convolution, its core innovation lies in the introduction of time-dependent dynamic weight adjustment mechanism. The model outputs the eigenvalue y t The specific operation process is expressed as:

[0082]

[0083] where w i (t) represents the dynamic weight of the i-th convolution kernel position at time step t, x t+i-1 is the ion current intensity detected by the instrument, and b is the bias term; the weight is generated in real time through the attention mechanism, which is implemented as follows: for each time step t, the query vector Q is generated through linear transformation t ∈R d , and define a learnable key vector K for each position i of the convolution kernel i ∈R d ; R d Represents a d-dimensional real vector; the weight calculation uses the scaled dot product attention form:

[0084]

[0085] This design enables the convolution kernel to dynamically adjust the receptive field weight according to the local contextual features of the input signal. Compared with the traditional convolution operation with fixed weights, it significantly improves the adaptability to the non-uniform noise distribution and low-abundance peptide signals in the mass spectrometry signal. In terms of quantitative stability optimization, the model introduces an abundance variance penalty term based on the mean square error loss function to form a composite optimization objective function:

[0086]

[0087] Where λ is the balancing hyperparameter, is the output value of the convolutional neural network, y is the true value of the polypeptide content, Var() represents the variance function, and the variance term By constraining the fluctuation of abundance estimates between adjacent scan points, the oscillation of quantitative results caused by random noise is effectively suppressed. This objective function can be used to determine the accuracy of the method, providing reliable support for subsequent research.

[0088] The Ivy optimization algorithm simulates the growth of ivy and possesses both global and local search capabilities, enabling rapid identification of optimal solutions while maintaining population diversity. This paper utilizes this algorithm to optimize an objective function. Its inputs include peptide mass spectrometry data, control parameters, and the objective function, and its outputs are optimized peptide detection parameters and population fitness information. This algorithm can improve the accuracy and sensitivity of label-free detection of anti-aging peptides.

[0089] S2.1: Initialization: Randomly initialize the population. Each individual represents an ivy plant, and its position in the search space represents the potential value of the problem's decision variable.

[0090] I y =I min +rand(1,D)*(I max -I min ),i=1,...,Npop

[0091] Where, I max Refers to the highest point in the search space; I min Refers to the lowest point in the search space; D is the dimension of the uniformly distributed random number in the interval [0,1]. Substitute the initialized individual position vector into the objective function The initial objective function value of each individual was calculated to measure the accuracy and stability of the model in predicting the peptide content under the corresponding parameter combination.

[0092] S2.2: Coordinated Species Growth: Simulate the growth rate of ivy as a function of time. Let G V is the growth rate, φ is the growth rate, is the correction coefficient for deviation growth. Then the growth rate of the discrete time system at time (t+1) is It can be expressed as:

[0093]

[0094] Where, is the growth rate of the discrete-time system at time t; rand refers to a random real number in the interval [0,1]; N(1,D) represents a random vector of dimension D, which is a random number from a standard Gaussian distribution.

[0095] S2.3: Obtain sunlight for growth: Simulate the behavior of plants obtaining sunlight.

[0096]

[0097] S2.4 Propagation and Evolution: Member I y Roam around the search space to the nearest I yi 、The most important neighbor I best This is equivalent to trying to follow the best member of the entire population I best , find a better optimal solution. Find the most suitable data and improve the accuracy of model data.

[0098]

[0099] S2.5: Survivor selection: Simulate the alternating stages of climbing and expansion in the life of ivy, and use decision-making methods to determine the growth direction of ivy. By comparing the quality of new and old solutions, the better solution is selected as the member of the next generation population. Current members Growth rate The new value of is calculated by the following formula. By comparing the data of anti-aging peptides, more accurate correlation data is obtained.

[0100]

[0101] While the traditional Ivy algorithm is highly effective in improving the accuracy of the objective function, its "coordinated and orderly species growth" step suffers from a problem: the growth rate formula for the discrete-time system moment is overly dependent on the random real number μ_t and the standard Gaussian distributed random number ε_t. This results in a high degree of randomness in the growth direction of individuals, potentially causing them to randomly wander around the search space without order, increasing the risk of falling into a local optimal solution. Therefore, this patent incorporates the attraction field of the global optimal solution into the species growth process, so that individuals are not only affected by their own random wandering but also attracted by the global optimal solution. This balances randomness and directionality to a certain extent, improving the algorithm's global search capability and convergence efficiency.

[0102] The Ivy optimization algorithm improves the species growth rate of the traditional Ivy algorithm. The improved species growth rate formula is:

[0103]

[0104] Where r t+1 represents the updated position vector of individual t; r t Represents the vector before the updated position of the current individual t; x t is the position vector of individual t in the search space. g best is the historical optimal solution of all individuals in the current population; θ is the random walk weight coefficient; ∈ t is the fitness value of the individual optimal solution of the current individual i and γ is the global attraction weight coefficient.

Claims

1. A label-free anti-aging polypeptide detection method, characterized in that: The following steps are involved: (1) Extract proteins from samples and perform pretreatment; (2) enzymatically hydrolyzing the polypeptide sample stored in acetic acid aqueous solution, using trypsin to cut the protein into peptide fragments; (3) Use improved label-free technology for mass spectrometry detection and analysis; (4) The Ivy optimization algorithm was used to perform a hierarchical search on the mass spectrometry parameters to obtain the detection results of anti-aging peptides.

2. The label-free anti-aging polypeptide detection method according to claim 1, characterized in that: The pretreatment includes using a lysis buffer to lyse the sample, disrupting the cells by ultrasonication in an ice bath and removing insoluble matter by centrifugation, then removing high-abundance proteins, and storing the extracted polypeptides in an acetic acid aqueous solution.

3. The label-free anti-aging polypeptide detection method according to claim 2, characterized in that: The removal of high-abundance proteins is targeted at blood samples, and requires the use of a multiple affinity removal column to specifically remove high-abundance proteins such as albumin and IgG.

4. The label-free anti-aging polypeptide detection method according to claim 1, characterized in that: The improved label-free technology includes peptide digestion, liquid chromatography separation, mass spectrometer ionization, mass spectrometry analysis, relative content inference, and statistical analysis; a generative adversarial network is introduced in the data preprocessing stage of the mass spectrometry analysis, in which the generator simulates the distribution of high-confidence peptide signals, and the discriminator evaluates whether the signal is a real signal based on the matching degree of ion fragments; a one-dimensional convolutional neural network model is introduced in the relative content inference, and background noise is suppressed through dynamic weight allocation.

5. The label-free anti-aging polypeptide detection method according to claim 1, characterized in that: The mass spectrometry detection and analysis using the improved Label-free technology includes: Let the protein sequence be a string S = {s1,s2,...,s m }, the protease cleavage rule is to recognize a specific amino acid pattern, and define the cleavage site set C = {c j }, where c j is the probability of cleavage at position j in the amino acid sequence, expressed as: Where k is the index number of the characteristic function, indicating different chemical characteristic types; f k refers to the characteristic function based on the chemical characteristics of local amino acid residues; w k refers to the protease specificity parameter; σ is the Sigmoid function; The peptide set after cleavage P = {p i }, of length n, satisfying: The migration rate of a peptide in a chromatographic column is determined by its partition coefficient between the mobile phase and the stationary phase, and its mathematical expression is as follows: Where C(x,t) refers to the concentration of the peptide at position x and time t on the chromatographic column; D refers to the diffusion coefficient, and v refers to the flow rate; The peptide ionization efficiency η is related to the charge z and the desolvation energy ΔG, and its mathematical expression is as follows: Where β refers to the mass spectrometer parameter, ΔG0 is the threshold energy; Generative adversarial networks are introduced in mass spectrometry analysis. The generator uses random noise vector z~p noise As input, generate simulated mass spectrometry signal Its goal is to approximate the real peptide signal distribution preal(x); the discriminator evaluates the biological plausibility of the input signal by analyzing the ion fragment matching degree, and outputs the probability D(x)∈[0,1] that the input signal belongs to the real data. The generator and the discriminator are jointly optimized through the minimization-maximization game, and the loss function is defined as: Among them, the generator needs to satisfy both the loss function objective and the fragment matching constraint. The calculation formula of the fragment matching S is modified to the normalized dot product form: Where I real (m / z) and I gen (m / z) represents the ion intensity of the real signal and the generated signal at the mass-to-charge ratio m / z, respectively. The denominator is normalized by the L2 norm to eliminate the influence of the difference in the total signal amount. When S→1, it indicates that the generated signal is highly consistent with the experimentally observed peptide fragmentation pattern; conversely, S→0 indicates that the generated signal deviates from the true biological law. A one-dimensional convolutional neural network model is introduced in the relative content inference, using the mass spectrometry signal sequence X∈R L As input, the convolution kernel W∈R with adaptive capability is used k Perform feature extraction; R L represents an L-dimensional real vector, R k Represents a k-dimensional real vector; introduces a time-dependent weight dynamic adjustment mechanism, and the model outputs the eigenvalue y t The specific operation process is expressed as: where w i (t) represents the dynamic weight of the i-th convolution kernel position at time step t, x t+i-1 is the ion current intensity detected by the instrument, and b is the bias term; the weight is generated in real time through the attention mechanism, which is implemented as follows: for each time step t, the query vector Q is generated through linear transformation t ∈R d , and define a learnable key vector K for each position i of the convolution kernel i ∈R d ; R d Represents a d-dimensional real vector; weight calculation uses scaled dot product attention form; In terms of quantitative stability optimization, the model introduces an abundance variance penalty term based on the mean square error loss function to form a composite optimization objective function: Where λ is the balancing hyperparameter, is the output value of the convolutional neural network, y is the true value of the polypeptide content, Var() represents the variance function, and the variance term Volatility is estimated by constraining the abundances between adjacent scan points.

6. The label-free anti-aging polypeptide detection method according to claim 1, characterized in that: The Ivy optimization algorithm improves the species growth rate of the traditional Ivy algorithm. The improved species growth rate formula is: Where r t+1 represents the updated position vector of individual t; r t Represents the vector before the updated position of the current individual t; x t is the position vector of individual t in the search space; g best is the historical optimal solution of all individuals in the current population; θ is the random walk weight coefficient; ∈ t is the fitness value of the individual optimal solution of the current individual i and γ is the global attraction weight coefficient.

7. A label-free anti-aging polypeptide detection system, characterized in that: including a sample processor, a mass spectrometer, and a search detector; The sample processor extracts proteins from the sample for pretreatment; enzymatically hydrolyzes the polypeptide sample stored in an acetic acid aqueous solution, and uses trypsin to cut the protein into peptide segments; The mass spectrometer uses an improved Label-free technology to perform mass spectrometry detection and analysis; The search detector uses the Ivy optimization algorithm to perform a hierarchical search on mass spectrometry parameters to obtain the detection results of the anti-aging polypeptide.

8. The label-free anti-aging polypeptide detection system according to claim 7, characterized in that: The improved label-free technology includes peptide digestion, liquid chromatography separation, mass spectrometer ionization, mass spectrometry analysis, relative content inference, and statistical analysis; a generative adversarial network is introduced in the data preprocessing stage of the mass spectrometry analysis, in which the generator simulates the distribution of high-confidence peptide signals, and the discriminator evaluates whether the signal is a real signal based on the matching degree of ion fragments; a one-dimensional convolutional neural network model is introduced in the relative content inference, and background noise is suppressed through dynamic weight allocation.

9. The label-free anti-aging polypeptide detection system according to claim 7, characterized in that: The mass spectrometry detection and analysis using the improved Label-free technology includes: Let the protein sequence be a string S = {s1,s2,...,s m }, the protease cleavage rule is to recognize a specific amino acid pattern, and define the cleavage site set C = {c j }, where c j is the probability of cleavage at position j in the amino acid sequence, expressed as: Where k is the index number of the characteristic function, indicating different chemical characteristic types; f k refers to the characteristic function based on the chemical characteristics of local amino acid residues; w k refers to the protease specificity parameter; σ is the Sigmoid function; The peptide set after cleavage P = {p i }, of length n, satisfying: The migration rate of a peptide in a chromatographic column is determined by its partition coefficient between the mobile phase and the stationary phase, and its mathematical expression is as follows: Where C(x,t) refers to the concentration of the peptide at position x and time t on the chromatographic column; D refers to the diffusion coefficient, and v refers to the flow rate; The peptide ionization efficiency η is related to the charge z and the desolvation energy ΔG, and its mathematical expression is as follows: Where β refers to the mass spectrometer parameter, ΔG0 is the threshold energy; Generative adversarial networks are introduced in mass spectrometry analysis. The generator uses random noise vector z~p noise As input, generate simulated mass spectrometry signal Its goal is to approximate the real peptide signal distribution preal(x); the discriminator evaluates the biological plausibility of the input signal by analyzing the ion fragment matching degree, and outputs the probability D(x)∈[0,1] that the input signal belongs to the real data. The generator and the discriminator are jointly optimized through the minimization-maximization game, and the loss function is defined as: Among them, the generator needs to satisfy both the loss function objective and the fragment matching constraint. The calculation formula of the fragment matching S is modified to the normalized dot product form: Where I real (m / z) and I gen (m / z) represents the ion intensity of the real signal and the generated signal at the mass-to-charge ratio m / z, respectively. The denominator is normalized by the L2 norm to eliminate the influence of the difference in the total signal amount. When S→1, it indicates that the generated signal is highly consistent with the experimentally observed peptide fragmentation pattern; conversely, S→0 indicates that the generated signal deviates from the true biological law. A one-dimensional convolutional neural network model is introduced in the relative content inference, using the mass spectrometry signal sequence X∈R L As input, the convolution kernel W∈R with adaptive capability is used k Perform feature extraction; R L represents an L-dimensional real vector, R k Represents a k-dimensional real vector; introduces a time-dependent weight dynamic adjustment mechanism, and the model outputs the eigenvalue y t The specific operation process is expressed as: where w i (t) represents the dynamic weight of the i-th convolution kernel position at time step t, x t+i-1 is the ion current intensity detected by the instrument, and b is the bias term; the weight is generated in real time through the attention mechanism, which is implemented as follows: for each time step t, the query vector Q is generated through linear transformation t ∈R d , and define a learnable key vector K for each position i of the convolution kernel i ∈R d ; R d Represents a d-dimensional real vector; weight calculation uses scaled dot product attention form; In terms of quantitative stability optimization, the model introduces an abundance variance penalty term based on the mean square error loss function to form a composite optimization objective function: Where λ is the balancing hyperparameter, is the output value of the convolutional neural network, y is the true value of the polypeptide content, Var() represents the variance function, and the variance term Volatility is estimated by constraining the abundances between adjacent scan points.

10. The label-free anti-aging polypeptide detection system according to claim 7, characterized in that: The Ivy optimization algorithm improves the species growth rate of the traditional Ivy algorithm. The improved species growth rate formula is: Where r t+1 represents the updated position vector of individual t; r t Represents the vector before the updated position of the current individual t; x t is the position vector of individual t in the search space; g best is the historical optimal solution of all individuals in the current population; θ is the random walk weight coefficient; ∈ t is the fitness value of the individual optimal solution of the current individual i and γ is the global attraction weight coefficient.

Images