131 results about "Omics data" patented technology

A system and method are disclosed for the collection and aggregation of genomic, medical, and other data of interest for individuals and populations that may be of interest for analysis, research, pharmaceutical development, medical treatment, and so forth. Contributors become members of a community upon creation of an account and providing of data or files. The data is received and processed, such as to analyze, structure, perform quality control, and curate the data. Value or shares in one or more community databases are computed and attributed to each contributing member. The data is controlled to avoid identification or personalization. Third parties interested in the database information may contribute value (e.g., pay) for access and use. Value flows back to the members and to a system administrative entity.

Techniques for storing omics data that indicates long sequences of elements associated with a particular biological molecule include receiving digital omics data comprising over two kilobytes. The digital omics data is split into multiple partitions. The maximum partition size is much less than then a number of elements in a typical instance of the particular biological molecule. Each partition is encrypted. Each encrypted partition is inserted into a corresponding data packet that includes an owner field that uniquely indicates an owner of the omics data. Each data packet is uploaded into a non-centralized, peer-to-peer distributed storage network. Thus, genome data is encrypted and stored in a distributed, scalable, fully decentralized, fast, and highly secure network. The network further provides for decentralized computing, trustless validation of the genomes by way of oracles and trustless genome analysis by third party providers through smart contracts.

The invention relates to a prediction method for the survival time of a patient with breast cancer based on a deep neural network. The method comprises the following steps: 1) acquiring data, whereinthe data is clinical data and omics data, and the omics data comprises gene expression data and DNA methylation data; 2) preprocessing the data; 3) extracting the features of a data set; and 4) constructing a deep neural network. According to the invention, survival time prediction is carried out by fusing the deep neural network with the multi-omics data of breast cancer; clinical data, gene expression data and DNA methylation data of breast cancer are obtained from a TCGA database, data features are extracted, deep neural network models are constructed respectively, and then back-end fusionis performed, so the performance of survival time prediction for breast cancer is improved, and a lifetime prediction model is obtained; and the method has a remarkable effect in prediction of the survival time of patients with breast cancer.

The invention discloses a multi-omics data association relationship discovery method based on sparse matching. The method includes: preprocessing input data to improve quality of the data; selecting asuitable similarity measure according to data characteristics to calculate a similarity matrix among data features; and fusing prior information to mine potential association relationships among thedata features on the basis of a similarity network among the features. According to the method of the invention, the prior information of features of existing already-proven omics data can be fully utilized, impacts of noises on results can be reduced, uncertainty caused by data errors can be reduced, and accuracy and robustness of the results can be improved.

The invention discloses a liver cancer image omics data processing method, system and device and a storage medium. The method comprises the steps of obtaining an upper abdomen CT image of a liver cancer patient, calculating image omics data corresponding to the upper abdomen CT image, calculating a linear combination of the plurality of image omics characteristic values, and dividing the liver cancer patient into a hepatic artery chemoembolization reaction group or a hepatic artery chemoembolization reaction-free group according to a magnitude relation between the predicted value and a presetthreshold value. The used imaging omics technology can extract the tumor feature information contained in the CT image from the three-dimensional perspective in an omnibearing mode, and the defect that tumor heterogeneity is not considered in an existing method is overcome; and patients are divided into a reaction group and a non-reaction group according to scores of the imaging characteristics ofpatients, so that the clinical guiding significance is good, and a better decision reference is provided for clinical workers. The method is widely applied to the technical field of data processing.

The invention relates to a calculation method for calculating lung cancer genotyping by using a PET/CT image. Outlines of lesions are sketched layer by layer, three-dimensional volume recombination iscarried out on a two-dimensional tumor region to generate a three-dimensional interested volume, and characteristic data is extracted from the three-dimensional interested volume; once the lesion (tumor) area is determined, image features can be extracted; image omics data is directly extracted and analyzed, and a radiogenomics data model is established; a function or mathematical model is established according to the extracted characteristics of each classification of the radiogenomics; an improved random forest algorithm is established; model prediction is improved by using a Pearson correlation coefficient; and lung cancer genotyping is calculated according to deep learning, thereby improving the diagnosis efficiency of the lung cancer.

The invention relates to a technology for predicting clinical complications after radiotherapy of tumor patients, in particular to a method for predicting complications of normal tissue organs after tumor radiotherapy based on the multimodal image omics characteristics and the radiotherapy dosimetry characteristics. The method mainly comprises the steps: S1, a multimodal image database is established; S2, image data of endangered organs near a tumor target area are extracted; S3, the image omics characteristics of the endangered organs are extracted, and the characteristics of normal organ image data are extracted; S4, parameters of the image phenotypic characteristics of the endangered organs are extracted according to image segmentation results; S5, the image omics characteristics are analyzed; S6, parameters of the exposure doses of the endangered organs are extracted; and S7, the characteristics are collected and extracted. The complications after radiotherapy and chemotherapy of the tumor patients are predicted through the image omics data, effective treatment and intervention can be provided for the patients in time through the reliable, safe and high-precision prediction method, the complications are reduced, and thus the treatment effect on the patients and the later life quality of the patients are improved.

The embodiment of the invention provides an omics data processing method and device based on a graph neural network, equipment and a medium, and relates to the technical fields of medical treatment, artificial intelligence, cloud data and the like. The method comprises the following steps: acquiring first omics data to be processed, wherein the first omics data comprises at least two first omics characteristics; determining a first correlation between different omics features in the at least two first omics features; constructing a first graph structure corresponding to the first omics data based on the at least two first omics features and each first correlation, one node in the first graph structure representing one first omics feature, and a connecting edge in the first graph structurerepresenting the first correlation corresponding to two nodes of the connecting edge; based on the first graph structure, obtaining node features of each node in the first graph structure through a graph neural network; and obtaining a medical analysis result based on the node features of each node. According to the invention, the node features can reflect the correlation among the features, and the obtained result is more accurate.

The invention discloses an intelligent disease analysis method and system based on ultrasonic omics and deep learning, and the method comprises the steps of obtaining a plurality of pieces of ultrasonic data of a lesion site, and obtaining multi-modal ultrasonic omics data; inputting the multi-modal ultrasound omics data into a deep learning neural network, and adjusting the connection weight, theproportion convolution and the pooling layer of neurons according to the multi-modal ultrasound omics data to obtain adjusted multi-modal ultrasound omics data; and classifying the adjusted multi-modal ultrasonic omics data by using classifiers in different modes, obtaining a score of each classification through a discriminator, and obtaining prognosis judgment, curative effect evaluation and anauxiliary diagnosis result according to the score of each classification. Compared with an existing method for intelligently analyzing diseases by using single-mode ultrasonic data, the technical scheme of the invention optimizes the deep learning network from the aspects of data and model design according to the characteristics of multi-mode ultrasonic omics data, and improves the accuracy and prediction value of intelligent analysis of diseases.

The invention provides an accompanying diagnosis model based on a PDX or PDC model drug sensitivity experiment and multi-omics detection analysis, a construction method and application. The method comprises four modules of construction of a human tumor cell line model and an animal transplantation model, a drug sensitivity experiment based on the human tumor cell line model and the animal transplantation model, acquisition of multi-omics data of a human tumor sample, and mining analysis of drug sensitivity phenotype-multi-omics data. The invention provides an implementation thought and a key method of the four modules, and reasonable design of correlation and matching among the four modules, and a whole set of integration strategy of model construction, drug sensitivity experiments, omicsdata acquisition, data integration and data mining analysis is formed. According to the method, the limitation of technical means of an existing accompanying diagnosis scheme research and developmentstrategy in the links of sample planning, data acquisition, data analysis and the like is overcome, accompanying diagnosis scheme research and development can be carried out more flexibly and more widely, and meanwhile mechanism interpretation clues related to markers are obtained.

The invention discloses a multi-omics intelligent diagnosis system based on deep learning, and the system comprises a data importing module which is used for loading multi-omics data and clinical data, and preprocesses the data; a sample similarity module which is used for constructing a multi-omics sample similarity matrix; an intelligent diagnosis training module which is used for carrying out feature representation by utilizing an automatic encoder, carrying out multi-omics feature fusion by utilizing a multi-view attention mechanism neural network and integrating a sample similarity moduleresult into a training process to finally obtain an optimal diagnosis model; and an intelligent diagnosis and prediction module which is used for carrying out intelligent diagnosis according to the multiple omics data and providing result explanation. According to the invention, the deep learning technology is combined with the multi-omics data, and the diagnosis result and interpretability of the disease are provided, so that the multi-omics intelligent diagnosis system based on deep learning is formed, the disease diagnosis capability is improved, and the interpretability of the diagnosis result is provided.

The embodiment of the invention discloses a clustering analysis method and system for multi-omics data, and the method comprises the steps: segmenting MR image information through employing a neural network, and extracting a high-throughput image hyper-parameter according to the segmentation information of each part; processing the clinical data, the demographic data and the laboratory inspection data to generate vector representations of different dimensions; performing multi-source data fusion on the high-throughput image data and the vector representations of different dimensions to obtain fused multi-source heterogeneous data; constructing a multi-source heterogeneous data set, and training and testing a multi-source graph clustering model to obtain an optimal model; inputting MR image information into the optimal model, and analyzing differences of different categories and similarities of the same category. A graph structure mode is adopted, the correlation condition between data is visually expressed, different features are captured, the model is more robust, the efficient clustering algorithm based on the graph neural network model is achieved, and the practical value is very high.

The invention discloses a sample clustering and feature recognition method based on integrated non-negative matrix factorization. The method comprises: 1, X = {X1, X2... XP} representing multi-view data composed of P different omics data matrixes of the same cancer; 2, constructing a diagonal matrix Q; 3, introducing graph regularization and sparse constraints into the integrated non-negative matrix factorization framework to obtain target functions O1 and O2; 4, solving the target function O1 to obtain a fusion feature matrix W and a coefficient matrix HI; solving the target function O2 to obtain a feature matrix WI and a fusion sample matrix H; 5, constructing an evaluation vector according to the fusion feature matrix W, and identifying common difference features according to the vector; 6, performing functional explanation on the identified common difference characteristics by using GeneCards; and 7, performing sample clustering analysis according to the fusion sample matrix. According to the method, the complementary and difference information of the multiple omics data can be fully utilized to identify the common difference characteristics, clustering analysis can be carriedout on the sample data provided by the multiple omics data, and a calculation method basis is provided for integrated research of different types of omics data.

The invention discloses a drug sensitivity prediction method and device based on multi-omics data fusion, and belongs to the field of drug sensitivity detection. The method comprises the steps that three kinds of multi-omics information including genomics data, proteomics data and metabonomics data of individual cell lines are integrated through a cell line diagram characterization module, and a cell line multilateral diagram is obtained; according to the cell line multilateral graph, multi-omics information of a cell line and potential relation between expression products of genes in a multi-omics level are fully considered, then feature extraction is conducted on the cell line multilateral graph through a cell line graph feature extraction module, and node features and edge features in the cell line multilateral graph are fully extracted to serve as cell line features; and finally, a drug sensitivity prediction module is adopted to predict the semi-inhibitory concentration of the drug according to the cell line features and the drug features extracted based on the drug feature extraction module. According to the method, the prediction accuracy of the drug sensitivity is improved on the basis of comprehensively considering genomics data, proteomics data and metabonomics data.

The invention discloses a method for identifying litchi varieties by using a wide targeted metabonomics technology. The method comprises the following steps: (1) obtaining a known litchi variety endogenous small molecule metabolite peak area database through litchi sample preparation, litchi sample detection and analysis, litchi sample qualitative and quantitative analysis and metabonomics data processing and analysis, screening to obtain 69 common difference significant metabolites, and generating a heatmap; and (2) obtaining the peak area of the endogenous small molecule metabolite of the unknown variety, combining the peak area with the peak area database of the known variety, carrying out multivariate statistical analysis, carrying out clustering analysis on the screened common differential metabolite, and judging the known variety which is firstly gathered with the variety to be identified as the variety to be identified or the similar variety. According to the method, the endogenous small molecule metabolites are obtained by utilizing a wide targeted metabonomics analysis technology, common difference significance metabolites of litchis are obtained through multivariate statistical analysis, and a visual map is generated and can be used for litchi variety identification.

The present invention provides a platform that enables codeless generation of online mass spectrometry data exploration portals. The platform facilitates upload of generic spreadsheets containing processed mass spectrometry results (e.g., peak tables) and enables on-the-fly hierarchical organization of data. Following data upload, platform users can select individual visualizations to add to their custom web portal from a menu of options. Based on these selections, a complete webpage is constructed with all associated functionality embedded. These custom web portals can then be shared with collaborators and other researchers at the discretion of the creator via a developed user permissions sharing scheme.

The invention discloses a drug sensitivity prediction method and device based on multi-omics similarity guidance, and the method comprises the steps: a cell line diagram constructed based on multi-omics data of a cell line can fully integrate four types of multi-omics information of genomics data, transcriptomics data, proteomics data and metabonomics data of an individual cell line; compared with an existing cell line characterization mode, more kinds of omics information can be contained, and meanwhile the potential relation between products expressed by the cell line in the multi-omics level is fully considered; on the basis, a drug sensitivity prediction model for predicting the drug sensitivity based on the cell line diagram adopts a multi-omics similarity guiding mode, multi-omics similarity among individuals can be fully considered while multi-omics information of individual specificity is efficiently extracted, huge biological priori knowledge can be provided, and the drug sensitivity prediction model can be used for predicting the drug sensitivity based on the cell line diagram. Therefore, more accurate drug sensitivity prediction is realized.

The invention discloses a miRNA-disease association prediction method and device based on graph neural network fusion multi-view information, and the method integrates miRNA-disease-related multi-omics data to construct a plurality of views, not only considers a plurality of homogeneous similarity networks, but also considers a heterogeneous bipartite network, extracts node features on all views in combination with a graph neural network and multi-view learning, captures dependency between global features and local features through a discriminator, and can better capture the complex nonlinear relation between miRNA and diseases.

The invention discloses a cancer survival analysis system based on multiple tasks and multiple modes, and the system comprises a data importing module which comprises a data loading module and a data preprocessing module; a system training module, comprising a representation learning module, a multi-omics and pathological map fusion module, a multi-task prediction module and a loss function module; and a system prediction module, used for carrying out survival analysis and grading according to the multi-omics data and the pathological map of the cancer patient. According to the invention, the deep learning technology and the multi-task learning technology are combined with the multi-omics and pathological map data, so that the complementary characteristics among the multi-modal data can be captured, and the sharing relevance among multiple tasks can also be captured, thereby forming a cancer survival analysis system based on multiple tasks and multiple modalities, and providing automatic survival analysis and cancer grading results.

The invention belongs to the technical field of medical equipment, and particularly relates to an endogastric biopsy Raman image auxiliary diagnosis method and system based on artificial intelligence. The artificial intelligence technology is applied to gastroscope stimulated Raman scattering endoscopic biopsy tissue image auxiliary diagnosis for the first time, and the method comprises the step: after histopathology image information is obtained by means of the stimulated Raman scattering microimaging technology, adopting the image classification and image omics data analysis based on a deep learning neural network and machine learning, and constructing an endogastric biopsy Raman image auxiliary diagnosis system. The system comprises a stomach tissue Raman image data preprocessing module, a neural network model, an algorithm module for training the neural network model, a neural network fine tuning module and a test module; compared with an existing traditional endoscope diagnosis and treatment system, the system has the advantages that real-time, rapid and intelligent diagnosis support in the endoscope examination process is achieved, and pathologists do not need to conduct explanation.