Screening model of antibacterial peptide and application thereof

Abstract

The invention relates to the technical field of biology, in particular to an antibacterial peptide screening model and application thereof.The screening model is constructed through the following steps that model characteristic parameters are recognized, and seven characteristics influencing the antibacterial activity of an antibacterial peptide sequence are extracted; the seven characteristics are sequence length, charge, hydrophobicity, alpha spiral tendency, amphoteric property, hydrophobic moment and L-type amino acid proportion; screening the antibacterial peptide sequences with the antibacterial activity by utilizing a machine learning model and a set threshold value, and judging that the antibacterial peptide sequences have the antibacterial activity when the comprehensive score of the seven characteristics is greater than the set threshold value. The screening model is applied to screening of probiotic strains with antibacterial / bacteriostatic activity, the probiotic strains with antibacterial / bacteriostatic activity can be successfully screened out, and the screening model is expected to have wide application prospects and excellent commercial value in the field of research and development of probiotics.

Description

Technical Field

[0001] The present invention relates to the field of biotechnology, in particular to a screening model for antimicrobial peptides and an application thereof. Background Art

[0002] Antimicrobial peptides originally refer to a class of alkaline polypeptide substances with antimicrobial activity that are induced in insects. They have a molecular weight of about 2000 to 7000 and are composed of 20 to 60 amino acid residues. At present, the use of conventional antibiotics often leads to the problem of bacterial resistance, which has seriously endangered human health. Finding new types of antibiotics is an effective way to solve the problem of drug resistance. Antimicrobial peptides have broad potential clinical application prospects due to their high antimicrobial activity, broad antimicrobial spectrum, multiple types, and the target strains are not prone to resistance mutations.

[0003] In addition, data show that antimicrobial peptides exist in probiotics. If probiotics with antibacterial or antibacterial effects can be screened out, it will help promote the development and application of probiotic products.

[0004] However, how to screen out antimicrobial peptide sequences with antimicrobial activity is a technical problem that urgently needs to be solved. Summary of the invention

[0005] The present invention aims to solve at least one of the technical problems existing in the related art. To this end, the first object of the present invention is to provide a screening model for antimicrobial peptides; the second object of the present invention is to provide an application of the screening model.

[0006] In order to achieve the first purpose, the technical solution adopted by the present invention is: The screening model for antimicrobial peptides was constructed through the following steps: S100, model feature parameter identification, extracting seven features that affect the antimicrobial activity of antimicrobial peptide sequences, the seven features are sequence length, charge, hydrophobicity, α-helix tendency, amphiphilicity, hydrophobic moment and L-type amino acid ratio; Seven parameters and their seven feature thresholds are set based on the seven features. The seven parameters and their seven feature thresholds are: Isoelectric point, greater than or equal to 9.0; hydrophobicity, average hydrophobicity index greater than or equal to 0.5; average hydrophobic moment, hydrophobic moment greater than or equal to 0.3; proportion of positively charged amino acids, greater than or equal to 15%; amphiphilicity score, greater than or equal to 0.6; sequence length, greater than or equal to 12 aa and less than or equal to 50 aa; proportion of aromatic residues, greater than or equal to 5%; Among them, the weight coefficients of isoelectric point and positively charged amino acid ratio are both 0.20, the weight coefficients of hydrophobicity and average hydrophobic moment are both 0.15, and the weight coefficients of amphiphilicity score, sequence length and aromatic residue ratio are all 0.10; S200, using the machine learning model and the set threshold, the antimicrobial peptide sequence with antimicrobial activity is screened. When the comprehensive score of the seven features is greater than the set threshold, the antimicrobial peptide sequence is determined to have antimicrobial activity. The comprehensive score calculation formula of the seven features is as follows: ; Among them, 7 is the number of features, w is the weight coefficient of the feature, and w i is the weight coefficient of the i-th feature, x is the actual value of the antimicrobial peptide sequence, and x i is the actual value of the ith antimicrobial peptide sequence, x min,i and x max,i are the minimum and maximum values ​​of the i-th feature, respectively.

[0007] Furthermore, in step S100, the method for identifying model characteristic parameters includes the following steps: S110, collecting antimicrobial peptide sequences from an antimicrobial peptide database, and processing the collected antimicrobial peptide sequence data to obtain an antimicrobial peptide sequence data set; S120. Feature extraction and encoding were performed on the antimicrobial peptide dataset to obtain seven features that affect the antimicrobial activity of antimicrobial peptide sequences.

[0008] Furthermore, the collected antimicrobial peptide sequence data in step S110 is processed as follows: S111. Clean the collected antimicrobial peptide sequence data by removing repeated sequences, non-standard amino acid sequences and sequences with abnormal lengths; S112. Standardize the collected antimicrobial peptide sequence data by converting all sequences to uppercase letters; S113. Interpolate or delete missing activity data to ensure the integrity and accuracy of the collected antimicrobial peptide sequence data.

[0009] Furthermore, in step S100, the method for extracting seven features affecting the antimicrobial activity of the antimicrobial peptide sequence includes variance threshold method, correlation analysis, recursive feature elimination and LASSO regularization.

[0010] Further, in step S200, the machine learning models include four types, namely, random forest, support vector machine, gradient boosting decision tree and deep neural network; Among them, the best machine learning model is the gradient boosting decision tree.

[0011] In order to achieve the second purpose, the technical solution adopted by the present invention is: An application of a screening model, wherein the screening model is the screening model for antimicrobial peptides as described in any one of the above items, and the application includes being used for screening probiotic strains.

[0012] Furthermore, the probiotics include any one of Bifidobacterium longum and Lactobacillus rhamnosus.

[0013] Furthermore, the probiotic strain includes one or more of the antimicrobial peptide sequences shown in Sequence 1 to Sequence 20.

[0014] Furthermore, the antimicrobial peptide sequence is used in the preparation of antibacterial / bacteriostatic drugs.

[0015] Furthermore, the drug also includes a pharmaceutically acceptable excipient or carrier.

[0016] The above one or more technical solutions in the embodiments of the present invention have at least one of the following technical effects: The screening model for antimicrobial peptides provided by the present invention integrates machine learning algorithms and bioinformatics methods to predict and evaluate antimicrobial peptide sequences with potential antibacterial / antibacterial activity. Preliminary results show that the model exhibits high accuracy and efficiency in identifying antimicrobial peptide sequences with antibacterial / antibacterial activity. The screening model is applied to the screening of probiotics with antibacterial / antibacterial activity, and can successfully screen out probiotic strains with antibacterial / antibacterial activity, which is expected to achieve broad application prospects and excellent commercial value in the field of probiotic research and development.

[0017] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 This is the peptide length distribution diagram provided in Example 1 of the present invention.

[0019] Figure 2 This is the amino acid frequency diagram provided in Example 1 of the present invention.

[0020] Figure 3 This is a diagram showing the antibacterial effect of Bifidobacterium longum FB1-2, FB1-3, FB1-5, FB1-9 and FB1-11 using the Oxford cup method provided in Example 3 of the present invention.

[0021] Figure 4 This is a diagram showing the antibacterial effect of Bifidobacterium longum FB1-1, FB1-4, FB1-6, FB1-7 and FB1-10 using the Oxford cup method provided in Example 3 of the present invention.

[0022] Figure 5 This is a diagram showing the antibacterial effect of Lactobacillus rhamnosus FL1-2 to FL1-6 using the Oxford cup method provided in Example 3 of the present invention.

[0023] Figure 6This is a diagram showing the antibacterial effect of Lactobacillus rhamnosus FL1-7 to FL1-9, FL1-11 and FL1-12 using the Oxford cup method provided in Example 3 of the present invention. DETAILED DESCRIPTION

[0024] In order to make the purpose, technical scheme and advantages of the present invention clearer, the technical scheme of the present invention will be clearly and completely described below in conjunction with the drawings in the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present invention. The following embodiments are used to illustrate the present invention, but cannot be used to limit the scope of the present invention.

[0025] In the following examples, the experimental methods used are conventional methods unless otherwise specified, and the materials, reagents, etc. used are all available from commercial sources unless otherwise specified.

[0026] Example 1 Construction of a screening model for antimicrobial peptides.

[0027] 1. Data collection and preprocessing.

[0028] All antimicrobial peptide sequences were collected from the Antimicrobial Peptide Database, URL: https: / / aps.unmc.edu / home. The database contains 3940 peptides, including 3146 natural antimicrobial peptides (AMPs) from six kingdoms of life (383 bacteriocins / peptide antibiotics from bacteria, 5 from archaea, 8 from protists, 29 from fungi, 250 from plants, and 2463 from animals), 190 predicted and 314 synthetic AMPs. The database has established strict registration standards to ensure that the included antimicrobial peptides have reliable experimental evidence. In addition, the database regularly collects and updates the latest discovered antimicrobial peptide sequences, and is currently recognized as a highly professional and comprehensive database of antimicrobial peptide sequences.

[0029] The data preprocessing steps include: data cleaning by removing repeated sequences, non-standard amino acid sequences and sequences with abnormal lengths; converting all sequences to uppercase letters to achieve sequence standardization; interpolating or deleting missing activity data to complete the processing of missing values; and randomly dividing the data set into training set, validation set and test set in a ratio of 8:1:1.

[0030] In addition, data visualization analysis was performed on the dataset to create two key histograms: peptide length distribution and amino acid frequency, e.g. Figure 1 and Figure 2The peptide length distribution diagram shows the number distribution of peptides of different lengths. The statistical results show that the length of antimicrobial peptides is generally concentrated in the range of 30aa to 50aa. The amino acid frequency diagram shows the frequency of occurrence of various amino acids in all peptide sequences. The statistical results show that the three most frequently occurring amino acids are: glycine (G): 12179 times, lysine (K): 10027 times, and leucine (L): 8708 times.

[0031] 2. Data feature extraction.

[0032] In order to fully describe the properties of antimicrobial peptides, the following features were extracted: sequence composition features such as amino acid composition, dipeptide composition, etc.; physicochemical characteristics such as hydrophobicity, charge, isoelectric point and molecular weight; structural characteristics such as secondary structure prediction and hydrophobic helix moment.

[0033] The results showed that antimicrobial peptide sequences have the following characteristics: first, these peptides are usually small, with a length of less than 20 amino acid residues; second, they usually carry a positive charge of at least +1. In addition, antimicrobial peptides are rich in hydrophobic and non-polar amino acids, and the secondary structure is generally an α-helical structure. Another important feature is that the proportion of L-type amino acids is more conducive to antimicrobial activity than D-type amino acids.

[0034] In the screening model of antimicrobial peptides, the present invention sets characteristic thresholds of seven key parameters based on data analysis and feature importance assessment to improve screening efficiency and accuracy. These parameters and their characteristic thresholds are: isoelectric point (pI ≥ 9.0), hydrophobicity (average hydrophobicity index ≥ 0.5), average hydrophobic moment (hydrophobic moment ≥ 0.3), proportion of positively charged amino acids (≥ 15%), amphiphilic score (≥ 0.6), sequence length (12aa ≤ length ≤ 50aa), and proportion of aromatic residues (≥ 5%). The setting of these screening characteristic thresholds is based on the statistical differences between antimicrobial peptides and non-antimicrobial peptides in each parameter, as well as the contribution of each parameter to the model prediction, thereby ensuring that the screened peptide sequences have higher antimicrobial potential.

[0035] Use Python to construct and screen antimicrobial peptide sequences. 3. Model construction and evaluation.

[0036] Feature extraction and encoding were performed on 3167 antimicrobial peptide sequences. Seven features were extracted, including sequence length, charge, hydrophobicity, α-helix tendency, amphiphilicity, hydrophobic moment, and L-type amino acid ratio. In addition, One-hot encoding and BLOSUM62 matrix encoding were used, each of which added 20 feature dimensions, bringing the total number of features to 47. Feature selection methods such as variance thresholding, correlation analysis, recursive feature elimination (RFE), and LASSO regularization were used to finally retain the 30 most representative features.

[0037] Based on these features, four machine learning models were constructed: random forest (RF), support vector machine (SVM), gradient boosted decision tree (GBDT), and deep neural network (DNN). The model performance was evaluated using five-fold cross validation, and the results showed that the GBDT model performed best, with an accuracy of 0.93 and an F1 score of 0.92; the RF accuracy was 0.92 and the F1 score was 0.91; the DNN accuracy was 0.91 and the F1 score was 0.90; although the SVM model performed relatively poorly, it still achieved an accuracy of 0.89 and an F1 score of 0.88.

[0038] Therefore, a GBDT model was constructed to predict the antimicrobial activity of antimicrobial peptide sequences. First, data preprocessing was performed: the antimicrobial peptide sequences and their antimicrobial activity labels were loaded from CSV files, and multiple features such as sequence composition, physicochemical properties, structural characteristics, and encoding information were extracted. Next, the dataset was randomly divided into a training set and a test set in a ratio of 8:2 to ensure the reliability of model evaluation. The Gradient Boosting Classifier in the machine learning algorithm (scikit-learn) library was used for model training, with parameters set to 100 estimators, a learning rate of 0.1, and a maximum depth of 3 to balance the accuracy and complexity of the model. The model performance was evaluated on the test set by calculating the accuracy and generating a classification report, which included indicators such as precision, recall, and F1 score, which fully reflected the classification ability of the model.

[0039] 4. Setting of screening threshold.

[0040] Based on the Youden index method, the optimal threshold for screening antimicrobial peptide sequences was determined to be 0.65. When the prediction probability (score) is greater than 0.65, the antimicrobial peptide sequence is determined to have antimicrobial activity. Using this threshold for prediction, the model's accuracy on the test set reached 0.94, and the F1 score was 0.93, with improvements in all indicators.

[0041] In the screening model of antimicrobial peptides, in order to improve the accuracy and efficiency of screening, the present invention selected seven key features: isoelectric point (pI), hydrophobicity, average hydrophobic moment, positive charge ratio, amphipathicity, sequence length, and aromatic residue ratio. Based on data analysis and feature importance assessment, we assigned a weight coefficient to each feature to reflect its relative contribution in the prediction of antimicrobial activity. Among them, the weight coefficients of isoelectric point and positive charge ratio are both 0.20, emphasizing their key role in the interaction between peptides and bacterial membranes; the weight coefficients of hydrophobicity and average hydrophobic moment are both 0.15, reflecting the influence of hydrophobic properties on the ability of peptides to penetrate cell membranes; the weight coefficients of amphipathicity, sequence length, and aromatic residue ratio are all 0.10, reflecting their auxiliary contribution to antimicrobial activity.

[0042] In order to evaluate the antibacterial potential of peptide sequences, the present invention constructed a comprehensive scoring model by multiplying the normalized values ​​of the above seven features with their corresponding weight coefficients and summing them. First, each feature was normalized to the minimum and maximum values ​​so that its value range was mapped to between 0 and 1. Then, the comprehensive score was calculated using the following formula: ; Among them, n is the number of features, w is the weight coefficient of the feature, and w i is the weight coefficient of the i-th feature, x is the actual value of the antimicrobial peptide sequence, and x i is the actual value of the ith antimicrobial peptide sequence, x min,i and x max,i are the minimum and maximum values ​​of the ith feature, respectively. The calculated comprehensive score is compared with the set threshold (0.65) to determine whether the antimicrobial peptide sequence has antimicrobial activity. When the comprehensive score is greater than the set threshold, the antimicrobial peptide sequence is determined to have antimicrobial activity.

[0043] Use Python to set the comprehensive feature threshold for screening antimicrobial peptide sequences.

[0044] Example 2 The antimicrobial peptide sequences of 20 probiotic strains were screened using the antimicrobial peptide screening model.

[0045] First, the peptide sequences of 20 probiotic strains were determined using 4D-DIA quantitative proteomics technology, including protein extraction, enzymatic hydrolysis, liquid chromatography-mass spectrometry tandem (LC-MS / MS) analysis and bioinformatics analysis, as described below: 1. Protein extraction

[0046] The sample was ground into powder in liquid nitrogen, and an appropriate amount of powder was transferred to a 1.5 ml centrifuge tube, and an appropriate amount of SDT lysis solution (4% SDS, 100 mM Tris-HCL, 10 ml 1 M DTT, 1 mM PMSF and 2 mM EDTA) was added. After oscillation and mixing, it was boiled at 95 ° C for 15 min, ultrasonically lysed on ice for 10 min, and the supernatant was centrifuged to obtain a protein solution. 4 times the volume of frozen acetone was added to the protein solution, precipitated at -20 ° C overnight, and centrifuged at 4 ° C to retain the precipitate. The precipitate was washed with cold acetone, and the precipitate was redissolved in 8 M urea, and the total protein concentration was determined by BCA protein quantitative analysis.

[0047] 2. Enzyme hydrolysis.

[0048] Take equal amounts of protein solution according to protein concentration, fill the volume to 200ul with 8M urea, then reduce with 10mM DTT at 37°C for 45min, alkylate with 50mM iodoacetamide (IAM) in a dark room at room temperature for 15min, add pre-cooled acetone 4 times the volume of protein solution, and precipitate at -20°C for 2h. After centrifugation, the protein precipitate was air-dried and resuspended in 200ul 25mM ammonium bicarbonate solution and 3ul trypsin (Promega), and digested overnight at 37°C. After digestion, the peptides of each sample were desalted on a C18 column, concentrated by vacuum centrifugation and re-dissolved in 0.1% (v / v) formic acid.

[0049] 3. LC-MS / MS analysis.

[0050] The proteins were separated using a Bruker NanoElute high performance liquid chromatograph, and the operation was performed according to the instructions of the Bruker NanoElute high performance liquid chromatograph.

[0051] Proteins were quantitatively analyzed using a Bruker timsTOF Pro2 mass spectrometer, which was operated according to the instructions of the Bruker timsTOF Pro2 mass spectrometer.

[0052] The mass spectrometry data were analyzed by bioinformatics using DIA-NN software.

[0053] 4. Screen probiotics with antibacterial activity.

[0054] The probiotic samples were analyzed using the screening model of antimicrobial peptides provided by the present invention. When the comprehensive score of the screening model was greater than 0.65, the peptide sequence was determined to have antimicrobial activity. 20 probiotic strains were screened to contain antimicrobial peptide sequences with antibacterial / antimicrobial activity. The prediction results of the antimicrobial peptide sequences with antimicrobial activity in the 20 probiotic strains are shown in Table 1.

[0055] Table 1 Antimicrobial peptide sequences with antimicrobial activity in 20 probiotic strains

[0056] Example 3 Evaluation of the antibacterial / antimicrobial effects of probiotics.

[0057] The Oxford cup method was used to evaluate the antibacterial / antimicrobial activity of probiotics against pathogenic bacteria. First, the OD600 of the pathogenic bacteria solution was adjusted to 0.1-0.2 using LB broth medium, and 200 μL of the solution was spread on the agar plate. Then, the Oxford cups were placed on the agar plate in sequence and the whole solution of probiotics grown to the logarithmic phase was added to each Oxford cup well. MRS liquid medium was used as a control. After culturing in a 37°C constant temperature incubator for 18-24 hours, the obvious inhibition zone was observed.

[0058] The experimental results are as follows Figures 3 to 6 As shown, Bifidobacterium longum FB1-1 to FB1-7 and FB1-9 to FB1-12 have antibacterial activity against Klebsiella pneumoniae and Staphylococcus aureus, respectively, and Lactobacillus rhamnosus FL1-2 to FL1-9, FL1-11 and FL1-12 have antibacterial activity against Bacillus cereus and Listeria monocytogenes. The experimental results show that the screening model for antimicrobial peptides provided by the present invention can be used for the screening of probiotic strains containing antimicrobial peptide sequences with antibacterial / antibacterial activity.

[0059] The creativity of the technical solution provided by the present invention is mainly reflected in the following aspects: First, it integrates machine learning algorithms and bioinformatics methods to achieve comprehensive analysis of probiotic genome and proteomic data; Second, through multi-dimensional feature engineering and multi-model comparison, the accuracy and reliability of prediction are improved; Third, the screening model of antimicrobial peptides is applied to the screening of probiotics with antibacterial and / or antibacterial activity, which is conducive to promoting the development and application of probiotic products; Fourth, this method realizes high-throughput screening, greatly improving the efficiency of screening strains with potential antibacterial / antibacterial activity from a large number of probiotics; Fifth, the present invention combines computer screening with experimental verification, and verifies the actual application effect of the screening model through the Oxford cup method.

[0060] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A screening model for antimicrobial peptides, characterized in that: Build it through the following steps: S100, model feature parameter identification, extracting seven features that affect the antimicrobial activity of antimicrobial peptide sequences, the seven features are sequence length, charge, hydrophobicity, α-helix tendency, amphiphilicity, hydrophobic moment and L-type amino acid ratio; Seven parameters and their seven feature thresholds are set based on the seven features. The seven parameters and their seven feature thresholds are: Isoelectric point, greater than or equal to 9.0; hydrophobicity, average hydrophobicity index greater than or equal to 0.5; average hydrophobic moment, hydrophobic moment greater than or equal to 0.3; proportion of positively charged amino acids, greater than or equal to 15%; amphiphilicity score, greater than or equal to 0.6; sequence length, greater than or equal to 12 aa and less than or equal to 50 aa; proportion of aromatic residues, greater than or equal to 5%; Among them, the weight coefficients of isoelectric point and positively charged amino acid ratio are both 0.20, the weight coefficients of hydrophobicity and average hydrophobic moment are both 0.15, and the weight coefficients of amphiphilicity score, sequence length and aromatic residue ratio are all 0.10; S200, using the machine learning model and the set threshold, screen the antimicrobial peptide sequence with antimicrobial activity. When the comprehensive score of the seven features is greater than the set threshold, the antimicrobial peptide sequence is determined to have antimicrobial activity. The comprehensive score calculation formula of the seven features is as follows: ； Among them, 7 is the number of features, w is the weight coefficient of the feature, and w i is the weight coefficient of the i-th feature, x is the actual value of the antimicrobial peptide sequence, and x i is the actual value of the ith antimicrobial peptide sequence, x min,i and x max,i are the minimum and maximum values ​​of the i-th feature, respectively.

2. The antimicrobial peptide screening model according to claim 1, characterized in that: In step S100, the method for identifying model characteristic parameters includes the following steps: S110, collecting antimicrobial peptide sequences from an antimicrobial peptide database, and processing the collected antimicrobial peptide sequence data to obtain an antimicrobial peptide sequence data set; S120. Feature extraction and encoding were performed on the antimicrobial peptide dataset to obtain seven features that affect the antimicrobial activity of antimicrobial peptide sequences.

3. The antimicrobial peptide screening model according to claim 2, characterized in that: The collected antimicrobial peptide sequence data in step S110 are processed as follows: S111. Clean the collected antimicrobial peptide sequence data by removing repeated sequences, non-standard amino acid sequences and sequences with abnormal lengths; S112. Standardize the collected antimicrobial peptide sequence data by converting all sequences to uppercase letters; S113. Interpolate or delete missing activity data to ensure the integrity and accuracy of the collected antimicrobial peptide sequence data.

4. The antimicrobial peptide screening model according to claim 1, characterized in that: In step S100, the method for extracting seven features affecting the antimicrobial activity of antimicrobial peptide sequences includes variance threshold method, correlation analysis, recursive feature elimination and LASSO regularization.

5. The antimicrobial peptide screening model according to claim 1, characterized in that: In step S200, the machine learning models include four types, namely random forest, support vector machine, gradient boosting decision tree and deep neural network; Among them, the best machine learning model is the gradient boosting decision tree.

6. An application of a screening model, characterized in that: The screening model is the antimicrobial peptide screening model according to any one of claims 1 to 5, and the application includes screening for probiotic strains.

7. The use of the screening model according to claim 6, characterized in that: The probiotics include any one of Bifidobacterium longum and Lactobacillus rhamnosus.

8. The use of the screening model according to claim 6, characterized in that: The probiotic strain includes one or more of the antimicrobial peptide sequences shown in Sequence 1 to Sequence 20.

9. The use of the screening model according to claim 8, characterized in that: The application of the antimicrobial peptide sequence in the preparation of antimicrobial / bacteriostatic drugs.

10. The use of the screening model according to claim 9, characterized in that: The drug also includes pharmaceutically acceptable excipients or carriers.

Images