39 results about "Gene interaction" patented technology

The invention discloses a bidirectional correlation analysis method for gene interaction of plants and pathogenic bacteria and application, and belongs to the technical field of biology. Comprising the following steps: (1) obtaining genetic variation information of a plant, genetic variation information of pathogenic bacteria and an anti-infection phenotype of the plant to the pathogenic bacteria; (2) carrying out forward whole genome correlation analysis, and locating pathogenic bacteria effect factors from plant disease-resistant genes; (3) carrying out reverse whole genome correlation analysis, and locating a plant disease-resistant gene from a pathogenic bacterium effect factor; and (4) disease-resistant genes obtained by forward whole genome correlation analysis and reverse whole genome correlation analysis are the same as the effect factors, and the disease-resistant genes and the effect factors are gene pairs interacting in plants and pathogenic bacteria. The method provided by the invention can efficiently, accurately and simultaneously screen out the interaction gene pair of the plant disease-resistant gene and the pathogenic bacterium effect factor, provides powerful support for analyzing a resistance mechanism exerted by the plant disease-resistant gene, and lays a foundation for cultivating a new variety of durable and broad-spectrum disease-resistant plants.

The invention discloses an accurate report automatic analysis and generation system based on gene detection data, and belongs to the technical field of bioinformatics and artificial intelligence. The system comprises a variation intelligent labeling module, a heterogeneous knowledge graph reasoning module, a scene adaptive report generation module and an intelligent quality control and feedback module; a three-layer heterogeneous knowledge graph containing gene-disease association, drug-gene interaction and clinical guide decision is constructed, variation-disease association reasoning is performed by adopting a relational graph attention network, scene adaptive generation of report content is realized in combination with a semantic slot filling mechanism, and the four modules form a closed-loop collaborative system through deep coupling. Gene detection data can be automatically converted into a clinical diagnosis report, and the report generation time is shortened from 48 hours to 30 minutes or less.

The invention discloses a mouse multi-gene collaborative insertion and knockout system based on multi-data coupling, and relates to the technical field of bioengineering, and the mouse multi-gene collaborative insertion and knockout system comprises a multi-gene collaborative targeting vector, a high-fidelity Cas9 protein expression unit, a homologous recombination template library, a multi-modal efficiency prediction module, a dual-fluorescence Cre report unit and an off-target inhibition element. According to the method, vector silencing is avoided by connecting the nuclear matrix attachment region in series with the gRNA expression cassette, and the multi-gene collaborative editing efficiency is improved to gt by combining homologous arm optimization design and a machine learning driven efficiency prediction model; 80%; cas9 fusion deaminase and a glycosylase inhibitor are combined with miRNA-mediated off-target inhibition, so that off-target sites of a whole genome are reduced to be less than or equal to 1 / sample; real-time monitoring of tissue specific expression is achieved by means of a double-fluorescence Cre reporting system, space-time regulation and control application is supported, and a high-reliability tool is provided for complex gene interaction research.

This invention provides a high-throughput method for constructing dual sgRNA libraries and its applications. The method utilizes high-throughput microarray synthesis technology to prepare a set of DNA fragments containing multiple dual sgRNA expression cassettes in a single step. After amplification, these fragments are assembled with a target vector containing a first promoter and a second gRNA backbone sequence in a first round of directed assembly to obtain a preliminary recombinant plasmid set. Then, linearized enzyme digestion and homologous recombination technology are used to insert a fragment containing a transcription termination sequence and a complete second promoter to complete the construction of the dual sgRNA expression unit. Finally, transformation and amplification yield the dual sgRNA plasmid library. This invention avoids the high error rate and high cost of long-chain oligonucleotide synthesis by utilizing microarray synthesis and simplifies the operation process through two rounds of directed assembly, significantly improving the throughput, fidelity, and efficiency of library construction. It is applicable to the construction of genome-wide dual sgRNA libraries, providing an efficient and reliable technical platform for high-throughput gene function screening, drug target discovery, and gene interaction research based on CRISPR.

This invention discloses a gene regulation inference method guided by topological data analysis and gene network embedding. It combines TDA and GNN to enhance the inference capability of gene regulation networks. By capturing the topological structure of the gene regulation network graph through TDA features, the model's ability to model gene expression is enhanced. The TDA features and GAT embedding representations are effectively integrated through gating fusion. This fusion mechanism enables the model to adaptively adjust node embeddings based on global topological characteristics, which not only improves the accuracy of gene interaction representation but may also enhance the accuracy of regulatory relationship prediction. The traditional GAT architecture is extended through a four-layer graph attention mechanism. Each layer uses residual connections to alleviate the gradient vanishing problem and improve training stability. In addition, independent multilayer perceptron branches are designed for transcription factors and target gene embeddings. This deep architecture can achieve more expressive feature transformations and capture subtle patterns in gene regulation networks.

The application provides a HBV inhibitor screening method based on a molecule-gene interaction constraint graph convolution network. In view of the limitation of a traditional drug discovery method in processing complex biological data, the application is based on a constructed compound library verified by anti-HBV in-vitro activity, a plurality of gene targets associated with the corresponding compound and an interaction network thereof, a graph data processing capacity of a graph convolution network model is used, and a molecule-gene interaction constraint graph convolution network model is constructed. The model combines an interaction matrix of a target protein corresponding to the gene, a gene feature matrix and a compound activity label, and effectively predicts the biological activity category of the compound. The specific steps include data processing, graph data generation, graph convolution network model training, hyperparameter optimization and model evaluation. The model parameter AUC value is 0.97, and the model effect is good. The application provides a new path and idea for virtual screening of anti-HBV drugs, and has potential application value.

The present invention discloses a method, device, terminal, and medium for predicting synthetic lethal gene pairs based on a graph convolutional neural network. The synthetic lethal gene pair prediction method includes: obtaining protein structural features based on protein structure data; obtaining protein sequence features based on protein sequence data; obtaining protein functional features based on a protein-protein interaction network; merging and standardizing the protein structural features, sequence features, and functional features to obtain the gene features of the primary protein-producing genes; obtaining interactions between genes, and training a synthetic lethal gene pair prediction model based on a graph convolutional neural network using the gene interactions and gene features; obtaining a final feature representation for each gene based on the trained synthetic lethal gene pair prediction model, and predicting whether two genes form a synthetic lethal gene pair based on the final feature representation. This method improves the efficiency of feature extraction and the ability to predict gene interactions.

The invention belongs to the field of plant molecular genetics and crop breeding, and provides a rapid identification method and a creation method of an extremely early rice variety. According to the invention, a specific genotype combination causing the rice to present an extremely early-maturing phenotype is clear and verified for the first time, i.e., the three genes Ghd7, Ghd7.1 and Ghd8 are represented as function deletion alleles, and meanwhile, the Hd1 gene is represented as a functional alleles. Based on the specific combination, genotype identification is performed on a plurality of known extremely early rice varieties in northeast and southern China, and results are completely identical. Furthermore, a novel rice material with the genotype combination is created through a gene editing technology, and a typical strain with an extremely early heading stage is successfully obtained, so that the accuracy and reliability of the combination are verified on the molecular level. According to the invention, the problem of inaccurate prediction caused by gene interaction complexity in the prior art is solved, and a powerful molecular tool is provided for early-maturing breeding of rice.

The invention discloses a whole genome gene-gene and gene-environment interaction detection method based on an interpretable genetic information neural network, which comprises the following steps of: 1, training a deep neural network model by using genotype data and phenotype data to obtain a trained deep neural network model; 2, inputting genome data of a sample, and predicting a phenotype value of the sample by using the trained deep neural network model; and step 3, calculating the xPI value of each genome variation site, and carrying out whole genome association analysis and genetic site detection through the xPI values. By constructing the deep neural network model and combining the genotype data and phenotype data of the sample, the target character is accurately predicted. Through trained network analysis, the xAI-GWAS not only quantifies the importance of each genotype data, but also can detect the additive effect, the gene interaction effect and the gene-environment interaction effect of the genotype data.

This disclosure describes a machine learning (ML) framework, including a graph convolutional neural network (GCN), for identifying gene expression signatures associated with cancer driver events. The model is trained to identify the TP53 mutation status of cancer samples from gene expression, utilizing a comprehensive, curated graph structure of gene interactions. Quantitative scores are generated to rank the severity of driver events in each sample. Very high AUC results for unknown data across several tumor types are achieved in this method. A strong correlation with protein function exists. The Signature in Transcriptome Associated with Mutant Proteins (STAMP) model can also predict driver events in many combinations of key oncogenes / pathways and several tumor types, based on well-established annotations from the literature. Thus, the STAMP model can identify and quantify driver events, which may lead to improved targeted therapy selection and prioritization in cancer patients.

A computerized system and method are provided for assessing a number of gene-gene interactions between the HLA and IRF5 gene regions. At least one computing device enrolls subjects in a registry, including SLE patients having met classification criteria for SEE and Sjogren's patients having met AECG criteria. Moreover, at least one computing device can perform genotyping for the subjects and healthy control subjects, for submission to a genotyping platform. Further, at least one computing device can develop HLA risk factor models for each of a plurality of stages, and perform statistical analysis for each of the plurality of stages.

The application discloses an immune pre-pneumonia evaluation method and system based on a machine learning algorithm, and relates to the technical field of immune evaluation.The application constructs a machine learning algorithm model based on multiple specific HLA subtypes, captures the correlation between gene interaction and immune pathways through a time sequence network, combines multi-task loss and adversarial training to enhance robustness, and realizes high-precision pre-risk evaluation of immune pneumonia caused by immunotherapy.The scheme solves the problems of the prior art, such as dependence on lagging symptoms, insufficient generalization ability of prediction markers, and sensitivity to data noise, has good biological interpretability and clinical practicability, can be extended to other immune side reaction prediction, and provides reliable support for personalized treatment decisions.

The application provides a biotin-labeled exogenous circular DNA and a construction method and application in protein interaction thereof, and belongs to the technical field of gene interaction. The application provides a construction method of the biotin-labeled exogenous circular DNA, wherein a target sequence is amplified by using a biotin-labeled primer to obtain linear DNA containing a preset restriction enzyme cutting site; the obtained linear DNA is subjected to single enzyme cutting treatment by using a corresponding restriction endonuclease; and the DNA after enzyme cutting is subjected to a self-ligation reaction to generate closed circular DNA. The preparation process of the biotin-labeled circular DNA is stable and controllable, and has high repeatability; the biotin labeling of the circular DNA can be combined with a streptavidin system to realize high-affinity and specific enrichment of a DNA-protein complex; and the circular DNA can be stably delivered in cells and maintain the structural integrity of the circular DNA, thereby providing an effective means for truly reflecting the interaction between the circular DNA and the protein in a physiological environment.

A pear transcription factor that inhibits anthocyanin biosynthesis PyERF4.1 gene, its recombinant expression vector and application, wherein the PyERF4.1 nucleotide sequence of the gene is shown in SEQ ID No.1. In the present invention, the transcription factor PyERF4.1 was overexpressed and knocked out in pear callus and tomato fruits to verify its function of inhibiting anthocyanin biosynthesis, and PyERF4.1 was PyERF3, PyMYB114, and PybHLH3 co-transformed with the PyERF4.1 gene into pear fruits and strawberry fruits, resulting in reduced anthocyanin accumulation in pear fruits and strawberry fruits. Through biological function verification, it is shown that the PyERF3, PyMYB114, and PybHLH3 gene cloned in the present invention interacts with the PyERF4.1 gene to inhibit the function of anthocyanin biosynthesis in pear peel.

The invention relates to the technical field of bioinformatics, in particular to a single-cell RNA sequencing cell type annotation method and system based on deep learning. According to the technical scheme, the method comprises the following steps: integrating scRNA-seq, epigenetic and proteome data, and mapping the data to a shared feature space through cross-modal alignment; dynamically adjusting a similarity threshold value to generate a cell dynamic relation graph; designing a hybrid network architecture, and combining a dynamic graph neural network to extract local topological features and capture global gene interaction with a lightweight Transform; performing self-supervised pre-training by using cross-modal contrast learning to generate general cell characterization; adaptive target data is transferred and learned through an adapter module, and fine tuning is supervised in combination with dynamic focus loss; federal learning security aggregation is realized based on differential privacy and homomorphic encryption, and model generalization is improved while privacy is protected. According to the method, the scRNA-seq, epigenetic data and proteome data are integrated, and multi-modal data complement each other, so that the cell characteristics can be described more accurately, and the annotation accuracy is remarkably improved.

The invention relates to the technical field of biomedicine, in particular to an endometrial cancer prognosis model and a construction method thereof, and the technical key points are as follows: the model is a gene-gene interaction model, gene-gene interaction is taken as a variable, a COX-ph model is used for screening prognosis independent correlation variables, and weights are given; the subject working curve is used for model prediction of the probability of the lifetime of the patient of 1-5 years; the method comprises the following steps: carrying out RNA-seq sequencing on an endometrial cancer tumor tissue, and constructing a stable prediction model (PMID: 35436725) by taking a product of expression quantities of two standardized genes as measurement of gene-gene interaction. According to the invention, RNA-seq sequencing is carried out on tumor tissues of endometrial cancer, so that the gene expression quantity and gene-gene interaction are effectively evaluated, and the detection cost can be remarkably reduced while a robust prediction result is provided for a patient; moreover, the data scale in the scheme of the invention is relatively large, so that a guarantee can be provided for the robustness of the model.

The present invention discloses a gene identification method based on a masked graph autoencoder, which comprises the following steps: 1. obtaining gene interaction data and omics feature data and performing preprocessing; 2. masking the processed data, specifically including two branches, a node masking module and an edge masking module; 3. inputting the masked network into a graph autoencoder for training, which learns the network embedding representation by reconstructing the nodes and edges of the network; 4. obtaining a low-dimensional embedding of features through the trained encoder, and finally using a logistic regression classifier to classify genes. The present invention simultaneously focuses on the node information and structural information of the graph by masking the nodes and edges in the network respectively, and reduces the dependence of the feature training model on label information in a self-supervised learning manner, thereby accurately classifying genes.

The invention provides an HBV inhibitor screening method based on a molecule-gene interaction constraint graph convolutional network. Aiming at the limitation of a traditional drug discovery method in processing complex biological data, the method is based on a constructed anti-HBV in-vitro activity verification compound library, a plurality of gene targets associated with corresponding compounds and an interaction network of the gene targets, and utilizes the graph data processing capability of a graph convolutional network model to determine the anti-HBV in-vitro activity verification compound library. And a molecule-gene interaction constraint graph convolutional network model is constructed. The model combines an interaction matrix of a target protein corresponding to a gene, a gene feature matrix and a compound activity tag to effectively predict the biological activity category of the compound. The method comprises the specific steps of data processing, graph data generation, graph convolutional network model training, hyper-parameter optimization and model evaluation. The model parameter AUC value is 0.97, and the model effect is good. The invention provides a new path and thought for virtual screening of anti-HBV drugs, and has potential application value.

The invention discloses a Kolmogov-Arnod network-based whole genome prediction method, which comprises the following steps of: obtaining a reference genome and genotype data to be predicted, annotating single nucleotide polymorphism (SNP) in the genotype data to be predicted based on the reference genome to obtain a genotype matrix of a gene level, the method comprises the following steps: acquiring a gene level genotype matrix, performing dimension reduction processing on the gene level genotype matrix by using a gene-principal component analysis (Gene-PCA) method to obtain low-dimensional representation of high-dimensional genotype data, inputting the obtained gene level genotype matrix into a pre-trained gene expression level prediction model to obtain a predicted gene expression profile, and predicting the gene expression level. And inputting the obtained low-dimensional representation of the high-dimensional genotype data and the predicted gene expression profile into a pre-trained whole genome prediction model to obtain a predicted phenotype value corresponding to the genotype data to be predicted. According to the method, the technical problem that the prediction precision is insufficient due to the fact that an existing linear statistical method cannot describe gene interaction and nonlinearity can be solved.

The invention discloses a targeted senile degenerative bone disease key lesion regulation factor mRNA therapy recommendation evaluation model method, and the model comprises a data input layer which receives bone disease genetic association and regulation data, including GWAS summary data, space transcriptome and single cell transcriptome data; the intervention target priority ordering module is used for integrating the obtained genetic evidence, regulation evidence and network evidence and carrying out priority ordering on intervention targets; and the intervention target identification network module is used for predicting a key intervention target according to an intervention target network generated by acquiring a gene interaction relationship from the pathway, and obtaining a potential mRNA intervention therapy. The genetic evidence comprises a genetic risk site set annotated through multi-modal regulation genomics data; the regulation evidence comprises functional genomics data related to diseases; the network evidence comprises a high-credibility protein interaction relationship. And the intervention target identification network module is used for analyzing an intervention target network and further comprises disturbance removal analysis and regulation and control hierarchy analysis.

The invention discloses a cloned hematopoietic multi-phenotype genetic big data driven cancer suppression target gene prediction method, which comprises the following steps: acquiring a cloned hematopoietic multi-phenotype genetic big data set which comprises a plurality of different cloned hematopoietic phenotypes, genetic regulation evidence and gene interaction evidence; a candidate target gene-clone hematopoietic multi-phenotype prediction matrix is constructed by analyzing genetic regulation evidence and gene interaction evidence, and the prediction matrix comprises candidate target gene data under different clone hematopoietic phenotypes and is used for evaluating the importance of each candidate target gene in different clone hematopoietic phenotypes; according to the prediction matrix, in combination with genetic regulation evidence and gene interaction evidence, performing priority ranking on candidate target genes to obtain cloned hematopoietic genetic target gene leader recommendation, and obtaining a leader recommendation target gene list; and obtaining a genetic target gene recommendation map according to the leader recommendation target gene list, and comparing the genetic target gene recommendation map with an immune disease map to obtain a list of target genes capable of being intervened.

A method for determining gene to gene interaction based upon gene modification is disclosed. The method includes: (i) extracting a cooperative network implementing an unsupervised regression tree; (ii) classifying a plurality of gene interactions into one of up-regulation category and / or down-regulation category; (iii) creating a gene expression profile matrix based upon gene data and resulting expression weighting functions; (iv) receiving a request to modify a gene and / or an expression of the gene; and (v) simulating an interaction of the modification within the gene expression profile matrix to provide a weighted direction graph.

The present invention discloses a bidirectional association analysis method for plant-pathogen gene interactions and its application, belonging to the field of biotechnology. The method comprises the following steps: (1) obtaining plant genetic variation information, pathogen genetic variation information, and the plant's resistance-susceptibility phenotype to the pathogen; (2) performing forward whole-genome association analysis to locate the pathogen effector from the plant's disease-resistance genes; (3) performing reverse whole-genome association analysis to locate the plant's disease-resistance genes from the pathogen effector; (4) the disease-resistance genes and effector obtained from the forward whole-genome association analysis and the reverse whole-genome association analysis are identical, and the two are the gene pairs interacting between the plant and the pathogen. The method provided by the present invention can efficiently and accurately screen out the gene pairs interacting between plant disease-resistance genes and pathogen effector factors simultaneously, providing strong support for analyzing the resistance mechanism exerted by plant disease-resistance genes and laying the foundation for breeding new plant varieties with long-lasting and broad-spectrum disease resistance.

A method and device for predicting gene interaction relationships based on a graph neural network, the method comprising: preprocessing the transcriptional expression data of each experimental group, including standardizing the gene ID, and representing the gene transcriptional expression data of different experimental groups with a unified gene ID; obtaining raw data for subsequent analysis; calculating gene similarity according to each sample group, and obtaining similarity data between genes in the sample group; selecting gene pairs that meet set conditions as the basis for constructing a graph data structure by statistically analyzing and screening the data of all sample groups; further processing the raw data to ensure that all sample groups contain only genes defined in the graph structure; then, standardizing the data of these sample groups so that the data dimensions of all sample groups are consistent; and performing model training on the generated graph data group to ultimately obtain edge weight data in the graph structure, which can accurately reflect the interaction relationship between genes. The present invention can be used to analyze gene expression data, predict the interaction relationship between genes, and perform unsupervised learning in the absence of labeled data.

The present disclosure describes a machine learning (ML) framework including a graph convolutional neural network (GCN) for identifying gene expression features associated with cancer driven events. The model is trained to identify the TP53 mutation status of a cancer sample from gene expression using a comprehensive selected gene interaction map structure. A quantitative score is generated to rank the severity of the drive events in each sample. By means of the method, an extremely high AUC result is achieved on unseen data of multiple tumor types. And the gene has strong correlation with protein functions. Based on the annotations established in the literature, a transcriptome feature (STAMP) model associated with the mutant protein can also predict driving events of a variety of important cancer gene / pathway combinations and a variety of tumor types. Therefore, the STAMP model can identify and quantify driving events, which can provide a new way for cancer patients to improve selection and priority ranking of targeted therapy.

The invention discloses a method for constructing a chromosome three-dimensional structure based on fusion gene data and application of the method, and belongs to the technical field of genomes. According to the method, internal interaction of fusion genes and interaction between chromosomes are analyzed, hierarchical structures of second, third and fourth stages of the chromosomes are determined by calculating the frequency of the interaction, and a three-dimensional structure is constructed according to the hierarchical structures of the second, third and fourth stages of the chromosomes; the method is of great significance in exploring gene interaction and determining the relation between the chromosome spatial position and the disease.

The invention provides a drug gene association prediction method and system based on multi-view comparative learning, and relates to the crossing field of biological medicine and artificial intelligence technology. The method comprises the following steps: constructing a drug-gene interaction view and a semantic view; performing neighborhood information aggregation on each view, and updating vector representation of nodes in each view after arrangement; fusing the vector representations of the nodes in each view through inner product operation, and adding to obtain a prediction score; the vector representation of each view node is subjected to comparative learning, so that the model discrimination capability is improved; calculating a BPR loss function, and optimizing model parameters; and forming a multi-task joint optimization strategy training model taking a prediction task as a main task and a contrast learning task as an auxiliary task, and completing the training if the loss of each task tends to be stable. According to the method, through multi-view comparative learning, the potential relation between known drugs and genes is deeply explored, so that the accuracy of drug-gene correlation prediction is improved.

According to the present invention, a system: acquires a dataset composed of gene expression data of a sample belonging to either of two groups; generates a gene network representing interactions between a plurality of genes by referring to a pathway database; selects one gene from among the plurality of genes; extracts a path by following the gene network downstream with the selected gene as a starting point; calculates path scores by executing enrichment analysis by using, as inputs, a list of genes included in the path and the dataset; generates a ranked gene list by ranking the plurality of genes on the basis of the path scores of the plurality of genes; and calculates drug scores by executing the enrichment analysis by using, as inputs, a list of target genes of drugs stored in a drug database and the ranked gene list.

The invention provides a wheat germ processing scheme recommendation method and system combined with deep learning. The wheat germ processing scheme recommendation method comprises the following steps: firstly, acquiring multi-dimensional basic data (including variety characteristics, raw material quality and microstructure image data) of wheat germs and processing scene demand parameters (including processing product application scenes, processing capacity adaptation and processing resource constraint data); performing feature gene extraction processing on the data to obtain a wheat germ feature gene set and a scene demand feature gene set; inputting the sequence into a pre-trained depth scheme evolution model, and generating a processing link feature gene sequence through a gene interaction layer; constructing a plurality of processing scheme prototypes based on the processing link characteristic gene sequence; and finally, performing double-circulation adaptive optimization on the processing scheme prototype and the feature gene set to obtain a target processing recommendation scheme, and transmitting the target processing recommendation scheme to a processing execution terminal, thereby realizing scientific and accurate wheat germ processing scheme recommendation.

Target genes are identified with high accuracy and the effectiveness of drugs is evaluated. [Solution] The system acquires a dataset consisting of gene expression data of samples belonging to one of two groups, references a pathway database to generate a gene network representing the interactions of multiple genes, selects one gene from the multiple genes, traces the gene network downstream starting from the selected gene to extract a path, calculates a path score by performing enrichment analysis using as input a list of genes included in the path and the dataset, ranks the multiple genes based on the path scores of the multiple genes to generate a ranked gene list, and calculates a drug score by performing enrichment analysis using as input a list of genes that are targets of drugs stored in a drug database and the ranked gene list.