A phage-specific antibacterial peptide data processing application method

Abstract

The present invention discloses a method for processing and applying bacteriophage-specific antibacterial peptide data, which includes: collecting bacteriophage polypeptide data, antibacterial peptide data, and non-antibacterial peptide data from public databases; removing redundant sequences from the antibacterial peptide data and non-antibacterial peptide data, performing homology analysis with the phage polypeptide data, and performing preprocessing to obtain the bacteriophage-specific antibacterial peptide data and non-antibacterial peptide data; constructing a positive sample data set and a negative sample data set; and applying the data, including training a neural network, applying the neural network, and obtaining positive and negative sample classification predictions for the phage-specific antibacterial peptide or non-antibacterial peptide data. The method of the present invention effectively explores the synergistic effect between phages and antibacterial peptides, achieving accurate target classification of phage-specific antibacterial peptides and non-phage-specific antibacterial peptides, thereby contributing to subsequent phage engineering improvements and the discovery of potentially highly effective antibacterial peptides.

Description

Technical Field

[0001] The present invention relates to the technical field of collaborative processing of bacteriophage data and antibacterial peptide data, and in particular to an application method for processing bacteriophage-specific antibacterial peptide data. Background Art

[0002] The misuse of antibiotics not only leads to bacterial resistance but also disrupts the normal human flora, posing a significant threat to human life. Bacteriophages grow and multiply within host cells, causing the lysis of pathogenic bacteria, reducing their density and minimizing or preventing infection or disease. Antibacterial peptides (ABPs) are a component of innate immunity and possess broad-spectrum bacterial activity. Phages and ABPs offer a distinctly different bactericidal mechanism from antibiotics and minimal side effects, making them an effective alternative to antibiotics.

[0003] Although some studies have proposed that antibacterial peptides can be used to enhance phages and have experimentally verified phage-driven antibacterial peptide expression systems, very few studies have explored the synergistic effects of the two, and there is also a lack of research on the gene fragments encoding antibacterial peptides in the phage genome, making laboratory methods time-consuming and labor-intensive.

[0004] In order to solve the above problems, it is necessary to propose a convenient and efficient data mining method to conduct in-depth exploration of the synergistic effect of bacteriophages and antibacterial peptides and improve the combined application of bacteriophages and antibacterial peptides. Summary of the Invention

[0005] In view of the above problems, the purpose of the present invention is to provide a phage-specific antibacterial peptide data processing application method, to conduct in-depth exploration of the synergistic effect of phage and antibacterial peptides, and to improve the combined application of phage and antibacterial peptides.

[0006] The embodiment of the present invention provides a phage-specific antibacterial peptide data processing application method, comprising:

[0007] S1. Collect phage peptide data from public databases;

[0008] S2. Collect antibacterial peptide data and non-antibacterial peptide data from public databases;

[0009] S3. After removing redundant sequences from the antibacterial peptide data and non-antibacterial peptide data, homology analysis is performed with the phage peptide data, and after preprocessing, phage-specific antibacterial peptide data and non-antibacterial peptide data are obtained;

[0010] S4. Use phage-specific antibacterial peptide data and non-antibacterial peptide data to construct positive sample datasets and negative sample datasets;

[0011] S5. Train a neural network based on the positive sample data set and the negative sample data set, apply the neural network, and obtain positive and negative sample classification predictions of phage-specific antibacterial peptides or non-antibacterial peptide data.

[0012] Furthermore, the process of collecting phage polypeptide data in S1 includes:

[0013] S11. Collect phage genome sequences from public databases;

[0014] S12. Use the PHANOTATE tool to identify small open reading frame sequences in phage genome sequences;

[0015] S13. Use the BioPython library to transcribe and translate the found small open reading frame sequences to obtain phage polypeptide data.

[0016] Furthermore, in said S13:

[0017] The transcription process includes: using the phage DNA chain as a template to synthesize an expected complementary mRNA coding chain, performing transcription operations on each small open reading frame sequence to obtain mRNA coding information;

[0018] The translation process includes: translating the mRNA coding information formed after transcription into the corresponding amino acid sequence according to the codons in the codon table to obtain phage polypeptide data.

[0019] Furthermore, in said S2:

[0020] The process of collecting antibacterial peptide data includes:

[0021] S2A1, collect peptide sequences with antibacterial activity from the CAMPR4 database;

[0022] S2A2, retaining the experimentally verified polypeptide sequence with antibacterial activity;

[0023] S2A3. Use the CD-HIT tool to remove redundant sequences with a similarity of 90% and sequences with a length of less than 5 amino acids in the peptide sequence data to obtain antibacterial peptide data.

[0024] The process for collecting data on non-antibacterial peptides includes:

[0025] S2B1, collect peptide sequences with non-antibacterial activity from the UniProt database;

[0026] S2B2. Use the CD-HIT tool to remove redundant sequences in peptide sequences with non-antibacterial activity and sequences with a length of less than 5 amino acids to obtain non-antibacterial peptide data.

[0027] Furthermore, the process of obtaining phage-specific antibacterial peptide data and non-antibacterial peptide data in S3 includes:

[0028] S31, using the BLASTp tool to compare the phage polypeptide data with the antibacterial peptide data and the non-antibacterial peptide data, respectively, to obtain preliminary sequences of the phage-specific antibacterial peptides and non-antibacterial peptides;

[0029] S32. Remove repetitive sequences, sequences longer than 100 amino acids, and sequences containing non-natural amino acids from the preliminary sequences of antibacterial peptides and non-antibacterial peptides, respectively, to obtain phage-specific antibacterial peptide and non-antibacterial peptide data.

[0030] Furthermore, the positive sample dataset and the negative sample dataset in S4 are divided into a training set, a validation set and a test set respectively through a stratified sampling strategy.

[0031] Furthermore, the neural network application in S5 includes the steps of:

[0032] S51. Use the ESM-2 protein pre-training model to convert the amino acid letter representations in the phage-specific antibacterial peptide and non-antibacterial peptide sequences into embedded feature representations in space;

[0033] S52, performing a convolution operation on the embedded features using a convolutional neural network (CNN) with feature extraction function;

[0034] S53. The convolutional feature representation is input into the fully connected layer (FCN) to perform positive and negative sample classification prediction on the phage-specific antibacterial peptide or the non-phage-specific antibacterial peptide.

[0035] Beneficial effects of the present invention:

[0036] 1. The present invention utilizes public databases and proposes an effective data collection method to collect phage polypeptide data, antibacterial peptide data, and non-antibacterial peptide data. Combined with bioinformatics analysis tools, the present invention analyzes and finds important gene coding fragments, explores the synergistic effect between phages and antibacterial peptides, and achieves accurate target classification of phage-specific antibacterial peptides / non-phage-specific antibacterial peptides, thereby contributing to the subsequent engineering improvement of phages and the discovery of potentially highly effective antibacterial peptides.

[0037] 2. The present invention uses phage-specific antibacterial peptide data and non-antibacterial peptide data to construct positive sample data sets and negative sample data sets, conducts neural network training and application, obtains positive and negative sample classification predictions of phage-specific antibacterial peptide or non-antibacterial peptide data, and verifies the accuracy of phage-specific antibacterial peptide / non-phage-specific antibacterial peptide target classification. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 Schematic diagram of the process of the phage-specific antibacterial peptide data processing application method of the present invention;

[0039] Figure 2 This is a schematic diagram of the principle structure of the present invention;

[0040] Figure 3 This is a schematic diagram of the neural network training application process of the present invention. DETAILED DESCRIPTION

[0041] The following describes embodiments of the present invention in detail. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar symbols throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended only to explain the present invention and are not to be construed as limiting the present invention.

[0042] In the existing technology, there is a lack of exploration of the synergistic effect of bacteriophages and antibacterial peptides, and there is also a lack of exploration of the gene fragments encoding antibacterial peptides in the bacteriophage genome.

[0043] In view of the above problems, the present invention provides a phage-specific antibacterial peptide data processing application method. Figure 1 A schematic flow chart of a method for processing bacteriophage-specific antibacterial peptide data provided by an embodiment of the present invention, the method comprising:

[0044] S1. Collect phage peptide data from public databases.

[0045] like Figure 2 As shown, the public database can be used to collect phage genome sequences from the public database of the National Center for Bioinformation (NCBI) of the United States; then, the PHANOTATE tool is used to identify small open reading frames in the phage genome sequence, which have the potential to encode polypeptides; finally, the BioPython toolkit is used to perform two steps of transcription and translation on the found small open reading frame sequences, and finally the polypeptide sequence data of the phage is obtained.

[0046] The transcription process includes: using the phage DNA chain as a template to synthesize an expected complementary mRNA coding chain, transcribing each small open reading frame sequence to obtain mRNA coding information and ensure accurate transmission of information.

[0047] The translation process includes: translating the mRNA coding information formed after transcription into the corresponding amino acid sequence according to the codons in the codon table to obtain phage polypeptide data.

[0048] After two steps of transcription and translation, the coding information in the phage DNA chain is converted into polypeptide sequence information.

[0049] S2. Collect antibacterial peptide data and non-antibacterial peptide data from public databases.

[0050] To collect antibacterial peptide data, peptide sequences with antibacterial activity can be collected from the CAMPR4 database, and only experimentally verified antibacterial peptides are retained. Then, the CD-HIT tool is used to remove redundant sequences with a similarity of 90% and sequences with a length of less than 5 amino acids in the data to obtain antibacterial peptide data.

[0051] For the collection of non-antibacterial peptide data, since there is currently no experimentally verified non-antibacterial peptide data available in public databases, polypeptide sequences with non-antibacterial activity can be collected from the Universal Protein Resource Database (UniProt) database; analogous to the antibacterial peptide data collection and processing method, the CD-HIT tool is used to remove redundant sequences in polypeptide sequences with non-antibacterial activity, as well as sequences with a length of less than 5 amino acids, to obtain non-antibacterial peptide data.

[0052] S3. Analyze the phage peptide data, the antibacterial peptide data, and the non-antibacterial peptide data respectively, and obtain phage-specific antibacterial peptide data and non-antibacterial peptide data after preprocessing.

[0053] Studies have shown that antibacterial peptides can be used to enhance bacteriophages, and that antibacterial peptides may be derived from the genomic sequence of phage hemolysin. Therefore, as important phage elements, antibacterial peptides can effectively modify the corresponding phage genome sequence, thereby improving the bacteriophage's sterilization ability.

[0054] The homology between the phage and the antibacterial peptide can be determined by performing similarity analysis on them. The homology columns can be aligned using the BLASTp tool (for example, setting the query coverage percentage to 50 and the expected alignment value to the default value of 10) to obtain the preliminary sequence of the phage-specific antibacterial peptide.

[0055] The phage and non-antibacterial peptides can be analyzed for similarity to determine their homology, or the BLASTp tool can be used for similarity comparison to obtain the preliminary sequence of the phage-specific non-antibacterial peptide.

[0056] like Figure 2As shown, the phage-specific antibacterial peptide data and non-antibacterial peptide data after sequence alignment are preprocessed. When illegal data appears, the data quality will be reduced, so the data needs to be screened out. Peptide sequences that are too long have weak antibacterial activity and may be toxic, so they need to be removed. Non-natural amino acids usually do not represent important biological information and may also introduce interference factors. They need to be removed in order to better understand the biological processes in natural organisms. To address the above problems, the processing steps include removing repeated sequences, removing sequences longer than 100 amino acids, and removing sequences containing non-natural amino acids, etc., to obtain high-quality phage-specific antibacterial peptide and non-antibacterial peptide data.

[0057] S4. Use phage-specific antibacterial peptide data and non-antibacterial peptide data to construct positive sample datasets and negative sample datasets.

[0058] After obtaining the data of phage-specific antibacterial peptides and non-antibacterial peptides, in order to better verify the accuracy of the acquisition of phage-specific antibacterial peptides, a neural network can be used to accurately classify phage-specific antibacterial peptides or non-phage-specific antibacterial peptides.

[0059] Phage-specific antibacterial peptide data and non-antibacterial peptide data were used to construct positive sample data sets and negative sample data sets. The positive sample data sets and negative sample data sets were divided into data sets using a stratified sampling strategy, and the proportions of training sets, validation sets and test sets for each category were approximately equal.

[0060] S5. Train a neural network based on the positive sample data set and the negative sample data set, apply the neural network, and obtain positive and negative sample classification predictions of phage-specific antibacterial peptides or non-antibacterial peptide data.

[0061] The neural network is trained, verified and tested based on the training set, validation set and test set obtained from the positive sample data and negative sample data to obtain a trained neural network.

[0062] Use the trained neural network to perform neural network applications, such as Figure 3 Shown, including:

[0063] Feature extraction,The protein universal pre-trained model (ESM-2) is applied to the amino acid representation of,phage-specific antibacterial peptide data to generate an embedded feature vector,representation in space.

[0064] Feature processing: A convolutional neural network (CNN) with feature extraction function performs convolution operation on the embedded features to achieve feature extraction of antibacterial peptide data.

[0065] Target classification: A fully connected neural network (FCN) with classification function is used as a classifier to perform target classification of phage-specific antibacterial peptides / non-antibacterial peptides.

[0066] Experimental data:

[0067] This experiment uses a dataset of phage-specific antibacterial peptides or non-antibacterial peptides that was collected, processed, and analyzed. The dataset contains 655 phage-specific antibacterial peptides and 2,959 phage-specific non-antibacterial peptides.

[0068] 80% of each dataset is used as the training set and 20% as the test set.

[0069] The evaluation index is accuracy (ACC Accuracy), and the calculation formula is:

[0070]

[0071] Among them, TP (True Positive) represents the number of true positives, that is, the number of true phage-specific antibacterial peptide sequences that are correctly predicted; TN (True Negative) represents the number of true negatives, that is, the number of true phage-specific non-antibacterial peptide sequences that are correctly predicted; FP (False Positive) is the number of false positives, that is, the number of phage-specific non-antibacterial peptide sequences that are originally phage-specific antibacterial peptide sequences; FN (False Negative) is the number of false negatives, that is, the number of phage-specific non-antibacterial peptide sequences that are originally predicted to be phage-specific antibacterial peptides.

[0072] Table 1 Dataset model application experimental results

[0073]

[0074] Table 1 shows the prediction results on the test set (consisting of 132 positive samples and 592 negative samples) using this evaluation metric. Using five-fold cross-validation on the training set, the average accuracy on the validation set was 98% ± 1%, and the accuracy on the test set reached 98.5%.

[0075] From the results in the table, it can be found that this method can provide accurate and real data samples, and can accurately reflect the diversity and complexity of bacteriophages and antibacterial peptides.

[0076] The present invention uses the above-mentioned method to collect phage polypeptide data, antibacterial peptide data, and non-antibacterial peptide data from public databases; the phage polypeptide data and the antibacterial peptide data and non-antibacterial peptide data are analyzed and preprocessed separately to obtain phage-specific antibacterial peptide data and non-antibacterial peptide data. Then, using the phage-specific antibacterial peptide data and non-antibacterial peptide data, positive sample data sets and negative sample data sets are constructed, and a neural network is applied to obtain positive and negative sample classification predictions of the phage-specific antibacterial peptide or non-antibacterial peptide data, thereby verifying the accuracy of the phage-specific antibacterial peptide or non-antibacterial peptide data.

[0077] 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 various embodiments of the present invention.

Claims

1. A bacteriophage-specific antibacterial peptide data processing application method, characterized in that: include: S1. Collect phage peptide data from public databases; S2. Collect antibacterial peptide data and non-antibacterial peptide data from public databases; S3. After removing redundant sequences from the antibacterial peptide data and non-antibacterial peptide data, homology analysis is performed with the phage peptide data, and after preprocessing, phage-specific antibacterial peptide data and non-antibacterial peptide data are obtained; S4. Use phage-specific antibacterial peptide data and non-antibacterial peptide data to construct positive sample datasets and negative sample datasets; S5. Train a neural network based on the positive sample data set and the negative sample data set, apply the neural network, and obtain positive and negative sample classification predictions of phage-specific antibacterial peptides or non-antibacterial peptide data.

2. The method for processing and applying bacteriophage-specific antibacterial peptide data according to claim 1, characterized in that: The process of collecting phage polypeptide data in S1 includes: S11. Collect phage genome sequences from public databases; S12. Use the PHANOTATE tool to identify small open reading frame sequences in phage genome sequences; S13. Use the BioPython library to transcribe and translate the found small open reading frame sequences to obtain phage polypeptide data.

3. The method for processing and applying bacteriophage-specific antibacterial peptide data according to claim 2, characterized in that: In said S13: The transcription process includes: using the phage DNA chain as a template to synthesize an expected complementary mRNA coding chain, performing transcription operations on each small open reading frame sequence to obtain mRNA coding information; The translation process includes: translating the mRNA coding information formed after transcription into the corresponding amino acid sequence according to the codons in the codon table to obtain phage polypeptide data.

4. The method for processing and applying bacteriophage-specific antibacterial peptide data according to claim 1, characterized in that: In S2: The process of collecting antibacterial peptide data includes: S2A1, collect peptide sequences with antibacterial activity from the CAMPR4 database; S2A2, retaining the experimentally verified polypeptide sequence with antibacterial activity; S2A3, using the CD-HIT tool to remove redundant sequences with a similarity of 90% and sequences with a length of less than 5 amino acids from the peptide sequence data to obtain antibacterial peptide data; The process for collecting data on non-antibacterial peptides includes: S2B1, collect peptide sequences with non-antibacterial activity from the UniProt database; S2B2. Use the CD-HIT tool to remove redundant sequences in peptide sequences with non-antibacterial activity and sequences with a length of less than 5 amino acids to obtain non-antibacterial peptide data.

5. The method for processing and applying bacteriophage-specific antibacterial peptide data according to claim 4, characterized in that: The process of obtaining phage-specific antibacterial peptide data and non-antibacterial peptide data in S3 includes: S31, using the BLASTp tool to compare the phage polypeptide data with the antibacterial peptide data and the non-antibacterial peptide data, respectively, to obtain preliminary sequences of the phage-specific antibacterial peptides and non-antibacterial peptides; S32. Remove repetitive sequences, sequences longer than 100 amino acids, and sequences containing non-natural amino acids from the preliminary sequences of antibacterial peptides and non-antibacterial peptides, respectively, to obtain phage-specific antibacterial peptide and non-antibacterial peptide data.

6. The method for processing and applying bacteriophage-specific antibacterial peptide data according to claim 1, characterized in that: The positive sample dataset and the negative sample dataset in S4 are divided into a training set, a validation set, and a test set respectively through a stratified sampling strategy.

7. The method for processing and applying bacteriophage-specific antibacterial peptide data according to claim 1, characterized in that: The neural network application in S5 includes the following steps: S51. Use the ESM-2 protein pre-training model to convert the alphabetical representations in phage-specific antibacterial peptide and non-antibacterial peptide sequences into embedded feature representations in space. S52, performing a convolution operation on the embedded features using a convolutional neural network (CNN) with a feature extraction function; S53. The convolutional feature representation is input into the fully connected layer (FCN) to perform positive and negative sample classification prediction on the phage-specific antibacterial peptide or the non-phage-specific antibacterial peptide.

Images