CN116913369A - Protein-protein interaction prediction algorithm based on multi-scale residual error network - Google Patents
Protein-protein interaction prediction algorithm based on multi-scale residual error network Download PDFInfo
- Publication number
- CN116913369A CN116913369A CN202310245763.1A CN202310245763A CN116913369A CN 116913369 A CN116913369 A CN 116913369A CN 202310245763 A CN202310245763 A CN 202310245763A CN 116913369 A CN116913369 A CN 116913369A
- Authority
- CN
- China
- Prior art keywords
- protein
- information
- interaction prediction
- scale
- module
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 230000004850 protein–protein interaction Effects 0.000 title claims abstract description 47
- 108090000623 proteins and genes Proteins 0.000 claims abstract description 70
- 102000004169 proteins and genes Human genes 0.000 claims abstract description 69
- 239000013598 vector Substances 0.000 claims abstract description 51
- 150000001413 amino acids Chemical class 0.000 claims abstract description 31
- 230000003993 interaction Effects 0.000 claims abstract description 25
- 238000010276 construction Methods 0.000 claims abstract description 16
- 238000010606 normalization Methods 0.000 claims abstract description 16
- 238000012549 training Methods 0.000 claims abstract description 6
- 238000000034 method Methods 0.000 claims description 17
- 238000007526 fusion splicing Methods 0.000 claims description 7
- 238000012545 processing Methods 0.000 claims description 6
- 239000012636 effector Substances 0.000 claims description 3
- 238000000605 extraction Methods 0.000 claims description 3
- 230000002209 hydrophobic effect Effects 0.000 claims description 3
- 230000009467 reduction Effects 0.000 claims description 3
- 238000004590 computer program Methods 0.000 claims 2
- 201000010099 disease Diseases 0.000 abstract description 5
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 abstract description 5
- 238000009510 drug design Methods 0.000 abstract description 2
- 238000009509 drug development Methods 0.000 abstract description 2
- 230000006806 disease prevention Effects 0.000 abstract 2
- 238000013528 artificial neural network Methods 0.000 description 5
- 238000002474 experimental method Methods 0.000 description 4
- 238000011161 development Methods 0.000 description 3
- 238000010801 machine learning Methods 0.000 description 3
- 230000006916 protein interaction Effects 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 230000015556 catabolic process Effects 0.000 description 2
- 238000006731 degradation reaction Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000004543 DNA replication Effects 0.000 description 1
- 241000590002 Helicobacter pylori Species 0.000 description 1
- 101800001554 RNA-directed RNA polymerase Proteins 0.000 description 1
- 125000003275 alpha amino acid group Chemical group 0.000 description 1
- 125000000539 amino acid group Chemical group 0.000 description 1
- 230000031018 biological processes and functions Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000001413 cellular effect Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 229940079593 drug Drugs 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 229940037467 helicobacter pylori Drugs 0.000 description 1
- 230000028993 immune response Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 208000015122 neurodegenerative disease Diseases 0.000 description 1
- 239000002547 new drug Substances 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 230000011664 signaling Effects 0.000 description 1
- 241000894007 species Species 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
- 238000013518 transcription Methods 0.000 description 1
- 230000035897 transcription Effects 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
- G16B5/00—ICT specially adapted for modelling or simulations in systems biology, e.g. gene-regulatory networks, protein interaction networks or metabolic networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/082—Learning methods modifying the architecture, e.g. adding, deleting or silencing nodes or connections
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
- G16B30/00—ICT specially adapted for sequence analysis involving nucleotides or amino acids
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
- G16B40/00—ICT specially adapted for biostatistics; ICT specially adapted for bioinformatics-related machine learning or data mining, e.g. knowledge discovery or pattern finding
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Biophysics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Medical Informatics (AREA)
- Evolutionary Biology (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Software Systems (AREA)
- Biotechnology (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Molecular Biology (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Computational Linguistics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Biomedical Technology (AREA)
- Analytical Chemistry (AREA)
- Public Health (AREA)
- Epidemiology (AREA)
- Databases & Information Systems (AREA)
- Physiology (AREA)
- Bioethics (AREA)
- Proteomics, Peptides & Aminoacids (AREA)
- Chemical & Material Sciences (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention is a protein-protein interaction prediction algorithm based on a multi-scale residual error network, thereby realizing rapid prediction of whether two proteins interact or not so as to promote targeted drug development and disease prevention and control, and providing a new thought for drug design and disease prevention, comprising 1) obtaining information between amino acid pairs by utilizing a pre-training model of a protein database, and adding amino acid composition, sequence, physicochemical property and information entropy to efficiently code protein sequences; 2) Extracting the characteristics of the protein information through a multi-scale residual error network, and ensuring that the information obtained by a protein characteristic vector construction module is fully learned; 3) The multi-feature obtained by the residual network module is fused, spliced and normalized by the feature vector normalization and interaction prediction module, so that a protein-protein interaction prediction result is finally obtained; 4) And performing parameter tuning on the model, and verifying the availability of the model in a plurality of intraspecies, interspecies, diseases and network data sets.
Description
Technical Field
The invention relates to the technical field of protein-protein interaction prediction, in particular to a protein-protein interaction prediction algorithm based on a multi-scale residual error network.
Background
Proteins are the basic units constituting cells, and are important in biological tissues. Proteins are usually in a non-independent state in organisms, and most proteins ensure normal operation of organisms through interaction with other proteins. Thus, exploring protein-protein interactions helps to understand the mechanisms of various biological processes, such as DNA replication gene transcription, immune response, and signaling, and to facilitate targeted drug development and disease control. The interaction between proteins forms a major component of the cellular biochemical network. With the continuous enrichment of amino acid sequence information and the continuous development of machine learning technology, protein-protein interaction prediction based on a calculation method gradually becomes an economic and effective method, and provides candidate solutions for biological experiments.
With the development of machine learning and deep neural networks, an algorithm based on machine learning can be effectively combined with protein sequence information, and a qualitative leap is obtained in accuracy, so that the resource utilization rate is improved. In practical terms, the research is of essential importance for understanding life internal tissues, and has important medical application value for treating diseases and researching and developing medicines. Therefore, the protein-protein interaction prediction method based on deep learning is provided by combining residual network and protein sequence information with multi-scale system structure improvement, can be used as a simple, efficient and accurate protein-protein interaction prediction tool, and has wide practical value.
Disclosure of Invention
The invention aims to solve the difficulties and challenges faced in the field of protein-protein interaction prediction, and provides a protein-protein interaction prediction algorithm based on a multi-scale residual error network. Can be used for predicting cross-species interactions and can be used as a compact, efficient and accurate protein-protein interaction prediction tool. The technical scheme of the invention is as follows:
a protein-protein interaction prediction algorithm based on a multi-scale residual error network comprises a protein feature vector construction module, a residual error network module and a feature vector normalization and interaction prediction module;
the protein eigenvector construction module maps the effect of a single amino acid on surrounding amino acids to the effect of one residue in the protein sequence on surrounding residues by Res2 vec. Aiding the amino acids to observe context information, finding abstract features from the raw data, and thus obtaining different properties hidden in the sequence. Meanwhile, three protein sequence coding methods of pseudo amino acid composition (PseAAC), autocorrelation Descriptor (AD) and mutual information descriptor (MMI) are fused to obtain protein information in different aspects, so that the sufficiency of the protein information is ensured.
The residual network module can relieve the problems of information loss and loss during information transmission, greatly protects the integrity of information, effectively avoids the degradation problem of a deep neural network through the residual network, and reduces the possibility of information loss; meanwhile, fine-granularity information is obtained by using a multi-scale method, and finally, the information obtained by the protein feature vector construction module is fully learned.
And the feature vector normalization and interaction prediction module is used for carrying out fusion splicing and feature vector normalization processing on the multiple features obtained by the residual error network module to finally obtain a protein-protein interaction prediction result.
Further, the protein characteristic vector construction module of the protein-protein interaction prediction algorithm based on the multi-scale residual network specifically comprises:
1) Using 566,995 protein sequences in the Swiss-Prot database as input, training the Res2vec model to find abstract features from the original data, thereby obtaining different hidden properties in the sequences and completing construction of any protein embedded vector;
2) Combining hydrophobic and hydrophilic effector factors of protein amino acids into the composition of classical amino acids by using pseudo amino acid composition (PseAAC) to represent internal sequence information of protein sequences, and obtaining vectors containing sequence information;
3) Extracting physicochemical properties of protein by using an Autocorrelation Descriptor (AD) to code so as to enrich sequence information and improve the accuracy of protein-protein interaction prediction;
4) Information entropy and group-based features are retrieved by k-gram using mutual information descriptors (MMIs) to obtain the joint information of amino acids. Specifically, the continuous k amino acids are regarded as a combination, and the characteristic information of the amino acids and the surrounding neighbors is obtained by adopting a sliding window method
5) And obtaining the feature vectors with the same length converted from the protein sequences with different lengths by fusion splicing the encoding methods.
Further, the residual network module of the protein-protein interaction prediction algorithm based on the multi-scale residual network specifically comprises:
1) The dimension of the feature vector is changed by utilizing the convolution block, so that the module can learn the input data more fully;
2) The model depth is increased by utilizing a plurality of identification blocks in series connection, and the learning of the data relationship is enhanced;
3) The scale block is used for representing the multi-scale features at a finer granularity level, so that the model has stronger multi-scale feature extraction capability and quite good generalization performance.
Further, the feature vector normalization and interaction prediction module of the protein-protein interaction prediction algorithm based on the multi-scale residual network specifically comprises:
1) Fusing and splicing the two protein sequence features after dimension reduction to obtain interaction features;
2) Normalizing the interaction characteristics to obtain a protein-protein interaction prediction result.
The beneficial effects of the invention are as follows:
aiming at the defects of long experimental period, low efficiency, incapability of fully considering protein sequence information and the like of the traditional experimental method, the invention provides a protein-protein interaction prediction model based on a multi-scale residual error network, which can fully acquire the association information among protein residues and also can acquire the information of amino acids. Meanwhile, vector features can be extracted from finer granularity by fusing the multi-scale convolution method, so that generalization capability and robustness of the model are improved. Can be used for predicting cross-species interactions, can help elucidate disease mechanisms, and provide new ideas for new drug design.
The invention has the following advantages:
1) Interaction information among residues is extracted by using an effective biological sequence representation method Res2vec, and characteristic information of sequences is enriched and supplemented by using PseAAC, AD, MMI. The interaction information among amino acid residues and the information characteristics such as sequence, physicochemical property, information entropy and the like are ensured to be integrated into a protein-protein interaction prediction task;
2) And the residual error network is encoded by utilizing a multi-scale system structure to obtain a feature vector, so that the receiving domain range of a network layer is increased, and the feature loss is avoided. The result shows that the protein-protein interaction prediction algorithm based on the multi-scale residual network can be an effective tool for effectively modeling protein-protein interactions, so that functions of unknown proteins are mined, and instructive suggestions are provided for biological species which are not verified through experiments. Meanwhile, the prediction result of the computer method can also be used as verification and supplement of biological experiment results.
Drawings
FIG. 1 is a schematic view of the application of the present invention
Detailed Description
The technical scheme of the invention is further described below with reference to the specific embodiments.
The algorithm is mainly based on a multi-scale residual error network and a multi-feature protein coding method, realizes accurate prediction of protein-protein interaction of protein, and can be used for predicting cross-species interaction so as to ensure high efficiency and accuracy of protein-protein interaction prediction. The protein-protein interaction prediction algorithm based on the multi-scale residual error network comprises a protein feature vector construction module, a residual error network module and a feature vector normalization and interaction prediction module, and basically comprises the following steps:
1) Protein-protein interaction data is obtained from a plurality of intra-species, inter-species and disease datasets, such as Helicobacter pylori, neurodegenerative disorders, and positive and negative samples are partitioned according to annotated protein-protein interaction labels for each sample and used for neural network training;
2) And a protein characteristic vector construction module for constructing a protein 3D structure embedding vector by training a Res2vec method. Meanwhile, three protein sequence coding methods of pseudo amino acid composition (PseAAC), autocorrelation Descriptor (AD) and mutual information descriptor (MMI) are fused to obtain protein information in different aspects, so that the sufficiency of the protein information is ensured.
3) And the residual network module is used for effectively avoiding the degradation problem of the deep neural network by adding the residual network into the neural network, so that the possibility of information loss is reduced. The information obtained by the protein feature vector construction module is fully learned.
4) And the feature vector normalization and interaction prediction module is used for carrying out fusion splicing and feature vector normalization processing on the multi-scale features obtained by the residual error network module to finally obtain a protein-protein interaction prediction result. "1" means that there is an interaction between proteins, and "0" means that there is no interaction between proteins.
Specifically, the protein characteristic vector construction module of the protein-protein interaction prediction algorithm based on the multi-scale residual network specifically comprises:
1) Training a Res2vec model by using 566995 protein sequences in a Swiss-Prot database to find abstract features from original data, so as to obtain hidden different properties in the sequences, and completing construction of any protein embedding vector;
2) The internal sequence information of the protein sequence is represented by pseudo amino acid composition (PseAAC). Specifically, hydrophobic and hydrophilic effectors of protein amino acids are combined into the composition of classical amino acids, resulting in a 31-dimensional vector containing sequence order information;
3) The physicochemical properties of the protein are extracted by using an Autocorrelation Descriptor (AD) for encoding so as to enrich the sequence information. Specifically, 7 physicochemical properties of 20 natural amino acids are encoded to obtain a feature vector with the dimension of 3 x n x 9, wherein n represents the number of the physicochemical properties;
4) Information entropy and group-based features are retrieved by k-gram using mutual information descriptors (MMIs) to obtain the joint information of amino acids. Specifically, regarding continuous k amino acids as a combination, adopting a sliding window method to obtain characteristic information of the amino acids and surrounding neighbors, and finally obtaining 119-dimensional characteristic vectors;
5) And obtaining the feature vectors with the same length converted from the protein sequences with different lengths by fusion splicing the encoding methods.
Specifically, the residual network module of the protein-protein interaction prediction algorithm based on the multi-scale residual network specifically comprises:
1) The dimension of the feature vector is changed by utilizing the convolution block, so that the module can learn the input data more fully;
2) The model depth is increased by utilizing a plurality of identification blocks in series connection, and the learning of the data relationship is enhanced;
3) The multi-scale block is used for representing multi-scale features at a finer granularity level, so that the model has stronger multi-scale feature extraction capability and quite good generalization performance.
Specifically, the feature vector normalization and interaction prediction module of the protein-protein interaction prediction algorithm based on the multi-scale residual network specifically comprises:
1) Fusing and splicing the two protein sequence features after dimension reduction to obtain interaction features;
2) And normalizing the interaction characteristics to obtain a protein-protein interaction prediction result, wherein the prediction result comprises 4 full-connection layers added with Batch Normalization and Dropout layers.
The protein-protein interaction prediction algorithm based on the multi-scale residual network consists of two parts, the first part, mapping each residue into a representation vector by means of Res2vec model trained by Swissprot dataset, and the second part, encoding the protein into a feature vector using PseAAC, AD, MMI. And splicing the feature vectors corresponding to the 2 protein sequences, and inputting the feature vectors into a residual error network. After passing through the residual network, the protein information is fully extracted, potential vectors rich in information are obtained, and finally, the feature vectors are input to the prediction module. Adding more protein information, such as structural and evolutionary information, to protein-protein interactions can further drive the study of protein-protein interactions. The research based on the algorithm is beneficial to increasing the practical economic benefit and accelerating the scientific research output, thereby promoting the vigorous development of bioinformatics.
The foregoing is illustrative of the present invention and is not to be construed as limiting the scope of the invention. Various changes and modifications to the present invention may be made by one skilled in the art after reading the teachings herein, and such equivalent changes and modifications are intended to fall within the scope of the invention as defined in the appended claims.
Claims (6)
1. A protein-protein interaction prediction algorithm based on a multi-scale residual network, comprising the steps of:
1) Protein feature vector construction, namely, taking a protein sequence as input, extracting hidden information among amino acid composition, sequence, physicochemical property, information entropy and amino acid pairs contained in protein, and converting the protein sequences with different lengths into feature vectors with fixed lengths;
2) The residual network module acquires sequence characteristics, acquires protein characteristic information with finer granularity by utilizing a multi-scale frame, and connects a plurality of identification blocks in series to increase the depth of the model, so that the study of the data relationship is further enhanced;
3) Feature vector normalization and interaction prediction are carried out, fusion splicing is carried out on features of two proteins with predicted interactions, and feature vector normalization processing is carried out, so that a protein-protein interaction prediction result is finally obtained.
2. The protein-protein interaction prediction algorithm based on the multi-scale residual network according to claim 1, wherein the protein-protein interaction prediction algorithm based on the multi-scale residual network comprises a protein feature vector construction module, a residual network module and a feature vector normalization and interaction prediction module. The protein characteristic vector construction module codes the protein by extracting hidden information among amino acid composition, sequence, physicochemical property, information entropy and amino acid pairs contained in the protein; the residual network module acquires protein characteristic information with finer granularity through a multi-scale frame, and a plurality of identification blocks are connected in series to increase the depth of the model, so that the learning of the data relationship is further enhanced; the feature vector normalization and interaction prediction module is used for finally obtaining a protein-protein interaction prediction result through fusion splicing and vector normalization processing of features of two proteins for predicting interaction.
3. The protein feature vector construction module of the protein-protein interaction prediction algorithm based on the multi-scale residual network according to claim 2 specifically comprises:
1) Using 566,995 protein sequences in the Swiss-Prot database as input, training the Res2vec model to find abstract features from the original data, thereby obtaining different hidden properties in the sequences and completing construction of any protein embedded vector;
2) Combining hydrophobic and hydrophilic effector factors of protein amino acids into the composition of classical amino acids by using pseudo amino acid composition (PseAAC) to represent internal sequence information of protein sequences, and obtaining vectors containing sequence information;
3) Extracting physicochemical properties of protein by using an Autocorrelation Descriptor (AD) to code so as to enrich sequence information and improve the accuracy of protein-protein interaction prediction;
4) Information entropy and group-based features are retrieved by k-gram using mutual information descriptors (MMIs) to obtain the joint information of amino acids. Specifically, the continuous k amino acids are regarded as a combination, and the characteristic information of the amino acids and the surrounding neighbors is obtained by adopting a sliding window method
5) And obtaining the feature vectors with the same length converted from the protein sequences with different lengths by fusion splicing the encoding methods.
4. The residual network module of a multi-scale residual network-based protein-protein interaction prediction algorithm according to claim 2, in particular comprising:
1) The dimension of the feature vector is changed by utilizing the convolution block, so that the module can learn the input data more fully;
2) The model depth is increased by utilizing a plurality of identification blocks in series connection, and the learning of the data relationship is enhanced;
3) The scale block is used for representing the multi-scale features at a finer granularity level, so that the model has stronger multi-scale feature extraction capability and quite good generalization performance.
5. The feature vector normalization and interaction prediction module of a multi-scale residual network-based protein-protein interaction prediction algorithm according to claim 2, in particular comprising:
1) Fusing and splicing the two protein sequence features after dimension reduction to obtain interaction features;
2) Normalizing the interaction characteristics to obtain a protein-protein interaction prediction result.
6. A computer device comprising a memory, a graphics card, a central processing unit, said memory storing a computer program, characterized in that the central processing unit implements the steps of the method according to any one of claims 1 to 5 when said computer program is executed.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310245763.1A CN116913369A (en) | 2023-03-15 | 2023-03-15 | Protein-protein interaction prediction algorithm based on multi-scale residual error network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310245763.1A CN116913369A (en) | 2023-03-15 | 2023-03-15 | Protein-protein interaction prediction algorithm based on multi-scale residual error network |
Publications (1)
Publication Number | Publication Date |
---|---|
CN116913369A true CN116913369A (en) | 2023-10-20 |
Family
ID=88357103
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202310245763.1A Pending CN116913369A (en) | 2023-03-15 | 2023-03-15 | Protein-protein interaction prediction algorithm based on multi-scale residual error network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN116913369A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117198426A (en) * | 2023-11-06 | 2023-12-08 | 武汉纺织大学 | Multi-scale medicine-medicine response interpretable prediction method and system |
-
2023
- 2023-03-15 CN CN202310245763.1A patent/CN116913369A/en active Pending
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117198426A (en) * | 2023-11-06 | 2023-12-08 | 武汉纺织大学 | Multi-scale medicine-medicine response interpretable prediction method and system |
CN117198426B (en) * | 2023-11-06 | 2024-01-30 | 武汉纺织大学 | Multi-scale medicine-medicine response interpretable prediction method and system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112256828B (en) | Medical entity relation extraction method, device, computer equipment and readable storage medium | |
CN113936735A (en) | Method for predicting binding affinity of drug molecules and target protein | |
CN113178229B (en) | Deep learning-based RNA and protein binding site recognition method | |
CN112562784B (en) | Protein function prediction method combining multitask learning and self-attention mechanism | |
CN109785901A (en) | A kind of protein function prediction technique and device | |
CN112289372B (en) | Protein structure design method and device based on deep learning | |
CN116913369A (en) | Protein-protein interaction prediction algorithm based on multi-scale residual error network | |
Hu et al. | Protein language models and structure prediction: Connection and progression | |
Yang et al. | Modality-DTA: multimodality fusion strategy for drug–target affinity prediction | |
CN114420211A (en) | Attention mechanism-based RNA-protein binding site prediction method | |
CN116417093A (en) | Drug target interaction prediction method combining transducer and graph neural network | |
KR102138165B1 (en) | Method for providing identity analyzing service using standard genome map database by nationality, ethnicity, and race | |
Du et al. | Improving protein domain classification for third-generation sequencing reads using deep learning | |
Kawano et al. | Seq2seq fingerprint with byte-pair encoding for predicting changes in protein stability upon single point mutation | |
CN117153246A (en) | Bitter peptide prediction method based on peptide language model | |
CN116386724A (en) | Method and device for predicting protein interaction, electronic device and storage medium | |
CN116705146A (en) | Multi-view enzyme function prediction method taking molecular structure and sequence mining into consideration | |
CN110915140B (en) | Method for encoding and decoding quality values of a data structure | |
CN117524318B (en) | New antigen heterogeneous data integration method and device, equipment and storage medium | |
CN113656539B (en) | Cross-modal retrieval method based on feature separation and reconstruction | |
US20230253076A1 (en) | Local steps in latent space and descriptors-based molecules filtering for conditional molecular generation | |
JP7490168B1 (en) | Method, device, equipment, and medium for mining biosynthetic pathways of marine nutrients | |
Li et al. | Deepacr: Predicting anti-crispr with deep learning | |
CN116562305B (en) | Aspect emotion four-tuple prediction method and system | |
CN115101119B (en) | Isochrom function prediction system based on network embedding |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication |