53 results about "Protein Feature" patented technology

The invention discloses an lncRNA-protein interaction prediction method based on bidirectional intention, and belongs to the technical field of bioinformatics, and the method comprises the following steps: data acquisition and preprocessing: constructing a data set; performing lncRNA feature extraction through a multilayer convolutional neural network; carrying out protein feature extraction based on an ACmix module; fusion of the lncRNA and protein bimodal features is carried out through a bidirectional intention network; performing lncRNA-protein interaction prediction on the fusion representation by using a three-layer and multi-layer perceptron network, and outputting at an output layer; and meanwhile, designing a loss function to carry out model optimization. The model is more comprehensive and accurate when the dependency relationship between the sequences is captured, so that the reliability and generalization ability of prediction are effectively improved.

The invention provides a drug target prediction method based on cross-modal attention and uncertainty evaluation, and belongs to the technical field of drug target prediction. In order to solve the technical problems that the existing drug target prediction lacks quantitative evaluation on the reliability of a prediction result and a nonlinear interaction relationship between a drug and a target is difficult to establish, the method comprises the following steps: collecting data, and fusing graph structure features and sequence features of extracted drug molecules to obtain a final code of the drug molecules; extracting amino acid sequence characteristics of the target protein, and constructing protein sequence characteristic expression; inputting the drug molecular features and the protein sequence features into a cross-modal attention module, and aligning and fusing the drug features and the protein features by using a bidirectional cross attention mechanism to obtain drug-target combined feature representation; introducing an uncertainty quantification mechanism, and outputting uncertainty estimation of a drug-target interaction prediction label and a prediction result; the method is used for predicting drug target interaction.

The invention discloses a drug sequence descriptor-based drug and target interaction prediction method, and belongs to the technical field of drug target prediction. According to the method, the descriptors of the smiles sequences of the two drugs are adopted, and after the two descriptors are added, prediction of the model on the result is more accurate due to the fact that the two descriptors contain structural information and chemical information related to tasks; for protein sequences, the characteristics of a large language model ESM-2 and the characteristics of a BERT large language model are added, and the model weights of the ESM-2 model and the BERT model are derived from tens of thousands of protein sequences, so that the representation capability of output vectors is very high; the interaction between protein features and drug features from a source sequence is realized through the designed cross attention layer with the shared weight, the prediction precision of the model is further improved, and experiments show that the model designed by the invention has a better effect on multiple indexes on multiple data sets.

The application relates to the technical field of computer-aided drug design, and provides a molecule generation method, system, medium and equipment for drug discovery, which comprises the following steps: acquiring a SMILES string of a ligand and an amino acid sequence of a target protein, respectively extracting a ligand feature vector and a protein feature vector, splicing and fusing, and then predicting a binding affinity value through a multilayer perception machine; taking the protein feature vector as a condition, generating a new molecular potential representation through a reverse denoising process of a conditional diffusion model; wherein, at each step of the reverse denoising process, a graph-level readout operation is performed on pure noise, an affinity value is predicted through the multilayer perception machine, a gradient of an affinity guidance loss is returned to a noise prediction network, and the new molecule generation is guided to a high affinity area; and the new molecular potential representation is decoded into a SMILES string through a pre-trained molecular language decoder. Novel molecules with high binding potential can be quickly generated.

Embodiments of the present application disclose a drug screening method, device and computer readable storage medium; after obtaining protein information of a target protein and molecular information of a candidate drug molecule, the target protein is a target macromolecule which can be acted on by the drug molecule, then, feature extraction is performed on the protein information and the molecular information respectively to obtain protein features of the target protein and molecular features of the candidate drug molecule, according to the protein information and the molecular information, interaction information of the target protein and the candidate drug molecule under a molecular force field is determined, then, based on the protein features, the molecular features and the interaction information, a binding strength between the target protein and the candidate drug molecule is calculated, and according to the binding strength, a target drug molecule is screened from the candidate drug molecule; the scheme can improve the accuracy of drug screening.

The invention discloses a macromolecular protein cutting method and system based on multi-source protein feature scoring, and a storage medium, and belongs to the field of computational biology and intelligent drug research and development. In order to solve the technical problem that macromolecular protein is difficult to directly input into an existing calculation model, an amino acid position cutting comprehensive score is obtained by obtaining multi-source feature data such as a protein disorder region and structural domain annotation, candidate cutting sites are screened and a final cutting scheme is determined in combination with double constraints of a structural domain hard boundary and fragment length, and the method is suitable for large-scale industrial production. And generating a protein fragment adaptive to downstream calculation. According to the method, the integrity of the protein structural domain is guaranteed, the cut fragment can be directly used as model input such as AlphaFold, the accuracy and stability of structural prediction and molecular docking are improved, the method is suitable for scenes such as virtual screening and computer-aided drug design, and the problem that an existing cutting method is lack of systematic consideration is solved.

This invention discloses a drug target affinity prediction method that integrates PPI quality and uncertainty. The method constructs a drug molecule map and a multimodal protein structure representation, and extracts multi-source features by combining the local PPI sub-map of the target protein. By calculating the protein's low-frequency level, prediction uncertainty, and PPI quality, a PPI quality-aware gating factor is generated to adaptively adjust the PPI information injection intensity, and a residual enhancement strategy is used to preserve the original protein features. Subsequently, the drug representation and the enhanced protein representation are fused using adaptive gating, and the result is input into a prediction network to output the drug-target affinity. This method effectively integrates protein function and interaction information, improves the prediction stability of low-frequency proteins and the model's generalization ability, and provides an accurate and reliable computational tool for drug screening and candidate molecule selection.

The application provides an immunogenic epitope prediction method based on antigen presentation and immunogenicity feature fusion, comprising the following steps: extracting peptide segments and type I human leukocyte antigen protein features from an immunogenic epitope database by using a constructed feature extraction module; inputting the peptide segments and type I human leukocyte antigen protein features into a pre-trained antigen presentation prediction model to obtain an antigen presentation probability based on the antigen presentation prediction model; inputting the peptide segments, type I human leukocyte antigen protein features and antigen presentation probability information into an immunogenicity prediction model to realize fusion of antigen presentation and immunogenicity features, and obtaining an immunogenicity score representing a T cell activation probability of the peptide segments based on the immunogenicity prediction model, so as to realize prediction of the immunogenic epitope. The application improves the prediction accuracy of the immunogenic epitope.

The application discloses a kind of lncRNA-protein interaction prediction methods based on bidirectional intention, belong to bioinformatics technical field, including the following steps: data acquisition and pre-processing, construct dataset;Through multilayer convolutional neural network, lncRNA feature extraction is carried out;Based on ACmix module, protein feature extraction is carried out;Through bidirectional intention network, the fusion of lncRNA and protein double mode features is carried out;LncRNA-protein interaction prediction is carried out to the fusion representation using three-layer multilayer perception machine network, and output is carried out in output layer;While designing loss function carries out model optimization.The model of the application is more comprehensive and accurate when capturing intersequence dependence, thereby effectively improving the reliability and generalization ability of prediction.

The invention belongs to the field of artificial intelligence algorithm application-drug target interaction prediction, and relates to a DTI prediction method based on LLM and RBMO. Firstly, through data collection, integration and redundancy elimination, a batch of drug target sequences with a significant positive and negative sample imbalance problem are obtained as input data; then, a K-BERT model is adopted to extract deep features represented by drug molecules SMILES, and a ProstT5 model is applied to analyze high-dimensional information of a protein sequence; and then, introducing an RBMO and KNN combined feature selection method to intelligently screen protein features, and finally selecting 513 key features with the best biological discrimination ability. And finally, combining a multi-core feature fusion mechanism with an emerging Kolmogorov-Arnold network (KAN) so as to enhance the analysis and characterization capability of the model on complex biomolecular features.

The invention discloses a drug target affinity prediction method, electronic equipment and a computer readable storage medium, the method comprises the following steps: feature extraction is carried out on input small molecule SMILES and protein sequences, and the small molecule adopts 10 molecular fingerprints of RDK, Topological, MACCS, AtomPair, ECFP4, FCFP4, FCFP6, Avalon, Layered and Pattern to construct mixed fingerprint features; compressing the high-dimensional molecular fingerprint features to vector dimensions consistent with protein features through linear mapping to realize feature balance, and splicing to form a fusion feature vector; and performing nonlinear interaction and expression enhancement on the fusion features by adopting a multi-expert hybrid module, and finally outputting an affinity prediction value between the small molecule and the protein through a linear regression layer. Through fusion of multi-source chemical fingerprints and deep protein pre-training representation, higher feature expression ability and stronger model generalization are realized; the complexity of the model is reduced through feature mapping and a lightweight multi-expert hybrid model, so that the model has better stability and expandability.

The application discloses a multi-component joint detection method, device and application for synchronous analysis of bone metabolism markers. The multi-component joint detection method comprises the following steps: pretreating a serum sample, and simultaneously extracting small molecule sterols, short peptides and protein characteristic peptide segments; adopting a reverse phase chromatography system to perform chromatographic separation on the extracts, and realizing synchronous elution of lipid-soluble substances and polar substances through a variable gradient elution program; and adopting a segmented multi-reaction monitoring mode to perform mass spectrometry detection, and simultaneously detecting high-abundance indexes and low-abundance indexes in bone metabolism markers in a same sampling period. The application realizes synchronous quantitative detection of five bone metabolism markers, solves technical problems such as fragmentation, large concentration span, incompatible chromatographic behavior and isomer interference in the prior art, and improves detection efficiency by more than 400%, and reduces detection cost by 50%-70%.

The present disclosure relates to an alternative protein material prediction device. The device includes a protein feature extractor configured to composite features of protein as sequence features, structural features, and physicochemical features; a protein graph data generator configured to generate nodes based on the sequence features and physicochemical features of the protein and generate edges between the nodes based on the structural features of the protein, thereby generating protein graph network data, and an alternative material protein predictor configured to generate an alternative protein material prediction model for predicting an alternative protein material by learning the protein graph data that reflects the composite features of the protein.

The present application relates to the field of medical informatics, and provide a paliperidone efficacy prediction method and system based on plasma proteomics, the method comprises the following steps: collecting the basic data of the target patient; selecting the protein features through the basic data to obtain the related proteins corresponding to the paliperidone efficacy; training the neural network based on the related proteins and the baseline data in the basic data to obtain the prediction model; predicting the paliperidone efficacy of the patient to be predicted through the prediction model to obtain the prediction result. The present application can predict the treatment response of schizophrenic patients to paliperidone, and significantly improve the objectivity of paliperidone use.

The invention discloses a synthetic protein design method based on protein feature engineering and a diffusion model, which comprises the following steps: firstly, searching a homologous sequence of a query sequence in a non-redundant NR database by using a P-BLAST tool and taking a target protein sequence as the query sequence; carrying out ProtParam-7 and ProtFactor-5 feature engineering on the screened homologous sequences, and carrying out ProtParam-7 and ProtFactor-5 feature engineering on the screened homologous sequences; taking data represented by the ProtParam-7 and ProtFactor-5 characteristics as original proteins, and inputting the original proteins into the diffusion model; performing multi-dimensional screening on the generated candidate proteins; evaluating the functional activity and the structural stability of the candidate protein by adopting a wet experiment, and taking the protein meeting the condition as a final generation result. According to the method, protein sequence feature extraction and diffusion model generation are combined, and multi-dimensional screening and experimental verification are carried out to ensure that the generated protein is close to the target protein in structure and function.

The invention is suitable for the technical field of deep learning, and provides a nucleic acid aptamer generation and screening method and device, electronic equipment and a program product. The method comprises the following steps: inputting a target protein into a protein structure encoder, and extracting a first protein feature of the target protein; inputting the guiding condition and the guiding intensity coefficient into a condition diffusion model to generate at least one candidate nucleic acid aptamer; inputting each candidate nucleic acid aptamer into a nucleic acid encoder, and extracting a first nucleic acid sequence feature of each candidate nucleic acid aptamer; inputting each first nucleic acid sequence feature and the first protein feature into a semantic space alignment module, and calculating a prediction score of each candidate nucleic acid aptamer; adding each candidate nucleic acid aptamer and the predicted score thereof into a nucleic acid candidate pool; and determining the candidate nucleic acid aptamer with the highest predicted score in the nucleic acid candidate pool as the nucleic acid aptamer with the optimal binding capacity with the target protein. According to the invention, efficient generation and screening of the high-affinity nucleic acid aptamer can be realized.

The embodiment of the invention discloses a protein generation method and device, electronic equipment and a storage medium, and the method comprises the steps: firstly obtaining protein demand information which comprises attribute information and functional domain information corresponding to protein; and then, performing feature extraction on the attribute information to obtain protein feature data corresponding to the attribute information, and performing feature extraction on the functional domain information to obtain functional domain feature data matched with a functional domain corresponding to the functional domain information, so as to obtain the protein feature data according to the protein feature data and the functional domain feature data. And generating a target protein containing the functional domain corresponding to the functional domain information. According to the technical scheme, the protein with the specific functional domain can be generated, the matching degree between the generated protein and the protein demand information is improved, and the success rate of protein generation is improved.

The invention relates to the technical field of computer-aided drug design, and provides a molecule generation method and system for drug discovery, a medium and equipment, and the method comprises the steps: obtaining an SMILES character string of a ligand and an amino acid sequence of a target protein, respectively extracting a ligand feature vector and a protein feature vector, and carrying out splicing fusion to obtain a target protein feature vector; predicting a binding affinity value through a multi-layer perceptron; generating a new molecular potential representation through a reverse denoising process of a conditional diffusion model by taking the protein feature vector as a condition; wherein in each step of the reverse denoising process, performing graph-level read-out operation on pure noise, inputting a multi-layer perceptron prediction affinity value, returning the gradient of affinity guide loss to a noise prediction network, and guiding new molecule generation to face a high-affinity region; and decoding the new molecular potential representation into an SMILES character string through a pre-trained molecular language decoder. And a novel molecular structure with high binding potential can be rapidly generated.

This disclosure provides methods for identifying subjects with cancer who will not respond to or are unlikely to respond to treatment (e.g., immune checkpoint inhibitor monotherapy), wherein the cancer has altered (e.g., increased or decreased) expression levels of a set of biomarkers, and methods for providing additional cancer treatment to the subjects with cancer using said methods.

This invention relates to the field of biological aging assessment technology, specifically a method for assessing cellular aging based on proteomics data and machine learning. The method includes: constructing a cell type-specific candidate feature set based on single-cell transcriptomics data; obtaining a protein feature matrix based on plasma proteomics data; constructing independent machine learning-based cell type-specific aging prediction regression models for each cell type, using age as the response variable and the protein feature matrix as input; inputting the proteomics data of the sample to be predicted into the corresponding cell type-specific aging prediction model to obtain the predicted lifespan and the lifespan difference characterizing cellular aging for different cell types. This invention achieves quantitative assessment of the aging degree of different cell types in different organs under in vivo conditions by constructing a functional mapping bridge between single-cell transcriptomics and plasma proteomics.

The application discloses a conformation prediction method. The method comprises the following steps: acquiring a conformation prediction model, and acquiring structure data of a complex to be predicted; wherein the structure data of the complex to be predicted comprises structure data of a ligand and structure data of a target protein; inputting the structure data of the ligand and the structure data of the target protein into the conformation prediction model to acquire ligand features and target protein features; performing position embedding, cross attention and connection operation on the ligand features and the target protein features to obtain a connection result matched with the ligand features and the target protein features; and predicting a target binding conformation of the structure data of the ligand and the structure data of the target protein according to the connection result. The technical scheme of the embodiment of the application provides a new conformation prediction method, learns important features such as shapes between ligands and target proteins, improves the prediction performance of the binding conformation, and helps to find the optimal binding conformation.

The application discloses a kind of methods for predicting drug-target interaction, model, equipment and medium, method includes: generating drug-protein feature vector subgraph;Graph attention sampling is applied, and the center node of subgraph is updated;Adaptive iterative optimization of feature vector is carried out using NDLS;While using graph SAGE to update the feature vector of center node, obtain node multi-source information;Through double-layer GTN, node multi-source information is aggregated, and output prediction value.The application uses NDLS to carry out adaptive iterative optimization of node feature, uses graph SAGE to carry out deep sampling aggregation to neighborhood node, and pays attention to the receptive field of different levels of network node;More relevant node information is effectively aggregated from neighborhood;The method of the application considers the receptive field of different levels of network node, by fusing respective complementary advantages, so that more relevant node information is effectively aggregated from neighborhood, and the prediction ability of DTI prediction model is further improved.

The application provides a protein coding method based on a position sequence matrix, which comprises the following steps: classifying amino acid sequences according to the dipole and volume of the side chain of the amino acid; constructing a sequence matrix and a position matrix; each element in the sequence matrix is used for indicating the frequency of two-by-two combination of all amino acids in the protein sequence; the position matrix is used for indicating the position information of any two groups of amino acids in the protein sequence; the diagonal line and the value above the diagonal line of the sequence matrix are selected together to code the amino acid sequence data, and the reciprocal of the one-dimensional sequence length is added as a component of the sequence matrix coding to distinguish the amino acid sequence length, so that the amino acid sequence data is coded into a feature vector. The position information and the sequence information of the protein sequence data are combined to code the protein sequence, so that the protein feature information can be fully obtained, the accuracy of protein interaction recognition is improved, and the robustness of the protein interaction prediction algorithm is enhanced.

The application discloses a kind of molecular-protein reaction prediction classification methods based on the improvement of Transformer network: (1) respectively to molecular and protein feature coding, sequence data is converted into machine recognizable feature vector dataset, wherein molecule uses adjacency matrix representation;(2) the dataset is divided into training set, verification set, test set;(3) training set is input to the improved Transformer network and is trained, and the feature vector of data is learned using bidirectional attention mechanism.(4) molecular feature vector and protein feature vector after training are spliced into Transformer network, input to full connection layer, and classification result is obtained.The molecular-protein reaction prediction classification method of the improved Transformer network proposed in the application can effectively improve the molecular-protein reaction prediction classification accuracy.

The invention relates to the technical field of data processing, in particular to a compound activity cliff classification method and device, electronic equipment and a storage medium, and the method comprises the steps: firstly obtaining a reference compound and a target compound determined corresponding to a target protein; respectively extracting compound characteristics corresponding to the reference compound and the target compound, and extracting protein characteristics of the target protein; under at least one specified affinity evaluation dimension, classification results are obtained according to fusion features of the two compound features and the protein features; wherein the at least one specified affinity evaluation dimension is randomly determined in preset N affinity evaluation dimensions. Therefore, a classification result under any specified affinity evaluation dimension can be obtained, and the fusion features can comprehensively represent the internal relation between the to-be-classified compound and the target protein, so that the interpretability of the classification prediction result is enhanced.

The invention relates to a drug-target affinity prediction method and system based on gene ontology guidance and multi-modal attention. The method comprises the following steps: complementing gene ontology GO function annotation of target protein; gO feature representation is extracted, sequence features are extracted through an ESM-2 model, and fusion is carried out through a gating mechanism to obtain protein features; using a Molformer model to obtain drug sequence features, converting an SMILES sequence of a drug molecule into a graph structure, using Transformer to extract topological structure features, and fusing to obtain drug features; a double-branch multi-head cross attention mechanism is used to realize bidirectional interaction of protein features and drug features, a Mama layer is used to perform deep feature extraction, and then prediction is performed. By extracting protein features and drug features, complex interaction between a drug and a target spot can be effectively captured; and a head cross attention mechanism is combined with a Mama layer to carry out deep feature extraction, so that the feature expression capability and the prediction performance are enhanced.

A computer-implemented method includes identifying, in a set of previously aligned nucleotide and polypeptide sequences characteristic of a candidate protein, which can be referred to as target residues; identifying at least one candidate region consisting of at least one pair of target residues identified in the first step that include target residues located at a determined distance in space and being exposed at the surface of the candidate protein; determining the advantageous chemical interactions between each residue and / or between each pair of residues within the candidate protein, from the 2D and / or 3D structure of the candidate protein that are hydrophobic bonds and / or hydrogen bonds and / or saline bridges and / or negative-repulsion and / or positive-repulsion bonds; selecting the residues linked by said advantageous chemical interactions that are at a distance of at most 10 angstroms; and selecting at least one therapeutic target region from among the candidate regions that include the selected residues.

The application discloses a method for evaluating the curative effect of traditional Chinese medicine in colorectal cancer by fusing network representation and deep learning. The application accurately identifies disease core driver genes through single-cell transcriptome differential analysis and protein-protein interaction network topology centrality index, simultaneously extracts the structural characteristics of traditional Chinese medicine ingredients by graph isomorphism network, constructs traditional Chinese medicine-compound isomorphism network, and extracts the isomorphism topological representation of traditional Chinese medicine by network representation algorithm. The one-dimensional convolutional neural network is used to deeply mine the protein sequence nature semantic features of disease gene sequence, so that the protein features and the end dimension of traditional Chinese medicine are aligned. Finally, the cross-modal joint vector is constructed through feature splicing, and the intervention probability output by the full connection neural network is used to realize the scoring of the candidate drug. The application establishes a direct and quantitative connection between the traditional Chinese medicine intervention space and the single-cell pathological mechanism, realizes the end-to-end evaluation of the anti-colorectal cancer efficacy of traditional Chinese medicine, and can be applied to the fields of traditional Chinese medicine screening and individualized drug administration scheme making.

The protein mutation stability change prediction method based on integrated convolutional neural network model and regression hierarchical training belongs to the technical field of protein stability change prediction. First, a training set and a test set are constructed, and four data sets are collected and arranged. Then the following four steps are carried out: first, the data is enhanced based on antisymmetry, and the sample data of stable and unstable mutations in the training set is balanced; second, protein features are extracted; third, a regression hierarchical sampling strategy is used to train the model; fourth, an integrated model based on multiple CNN sub-models is constructed for protein mutation stability change prediction. The present application first introduces spatial neighborhood evolutionary information; for the first time, a regression hierarchical sampling strategy is proposed and successfully applied in the training process of the model, effectively improving the prediction ability of the model for extreme ΔΔG; the trained CNN sub-models are combined, and the information of all samples in the training set is fully utilized, which is conducive to enhancing the generalization ability of the model.

The application discloses a drug target affinity prediction method based on a graph virtual node, which introduces virtual nodes and virtual edges in a drug molecule structure graph, takes a graph transformer as a graph feature extractor, takes virtual node features as drug representation, then inputs protein and drug features into a feature fusion module, adds an attention mechanism and a gated skip connection mechanism in feature fusion, captures potential interaction while fusing feature information of different hidden layers, and realizes higher-precision affinity prediction. The application can solve the technical problems of long time consumption and high cost of a drug target affinity prediction method in traditional machine learning, and the technical problems of other existing deep learning that can only transmit neighbor node features and cannot consider global information of drug features, and the technical problem of directly splicing drug and protein features and failing to capture the interaction relationship between the two.