38 results about "Subtype classification" patented technology

The invention discloses a deep learning-based lymphoma pathological image intelligent identification method. The method comprises the steps of preprocessing lymphoma pathological section image data; constructing a full convolutional neural network for segmenting a lymphatic tissue region, wherein the full convolutional neural network comprises an encoder sub-network and a decoder sub-network; constructing a lymphoma three-classification convolutional neural network under high resolution, wherein the lymphoma three-classification convolutional neural network is composed of six convolutional layers and three full-connection layers which are connected in sequence; and training the full convolutional neural network and the lymphoma three-classification convolutional neural network to finally obtain a lymphoma pathological section image classification model, and sequentially passing through the full convolutional neural network and the lymphoma three-classification convolutional neural network during testing to finally obtain a lymphoma classification result. Reliable intermediate data are provided for pathologists to judge lymphoma subtype categories, and auxiliary diagnosis referenceis provided for the pathologists to classify lymphoma subtypes by analyzing digitally scanned lymphoma pathological images, so that the pathologists are helped to quickly judge lymphoma conditions ofpatients.

Described herein are gene signatures providing prognostic, diagnostic, treatment and molecular subtype classifications of ovarian cancers through generation of ovarian cancer disease signatures (OCDSs) that account for molecular heterogeneity present in gynecological cancers. An ovarian cancer fixed signature (OCFS) is described which relates to the core programming of disease development, in addition to an ovarian cancer stem cell (OCSC) signature. Development various disease signature, suggests personalized treatment strategies focused on molecular subtypes of gynecological cancers, such as triage tests for patients.

The invention discloses a human body period expression gene identification method based on an RNN-CNN neural network fusion algorithm. The method comprises the steps of performing missing filling andtime sequence period extension on original data, then establishing a high-quality learning dataset through a data level, performing thermal graph clustering on period expression data for preliminarilyunderstanding the specific expression of a biological clock gene in a visual sense angle, then combining a recurrent neural network algorithm and a convolutional neural network algorithm for obtaining an RNN-CNN fused algorithm by means of a deep learning algorithm, and performing periodical and non-periodical subtype classification on an RNN-CNN fused algorithm classifier and common deep learning algorithms CNN and RNN, performing cross validation on a classifying result, and performing evaluation by means of an accuracy score, a recall rate and a comprehensive evaluation index.

The invention discloses a tumor immune subtype classification method and a tumor immune subtype classification system. The method comprises the following steps: acquiring a sample data set comprisingRNA-seq sequencing data and gene expression profile data of a plurality of tumor tissues; calculating microorganism abundance data, immune cell proportion data and gene expression data related to immunity in the tumor tissue corresponding to each sample in the sample data set; constructing a training sample data set by taking microorganism abundance data, immune cell proportion data and gene expression data related to immunity as classification features, expanding minority classes by adopting an SMOTE algorithm, training a random forest model, and improving the random forest model in a weighting form: increasing the weight of the minority classes. And thus, the decision tree classifier focuses on the minority class, and the classification accuracy of the minority class is improved. According to the invention, the accuracy of tumor immune subtype classification prediction can be improved, and a new target is provided for tumor immunotherapy.

The invention relates to a precise intelligent diagnosis and treatment big data system which comprises a centralized data management module, a data preprocessing module, a marker extraction module, asubtype classification module and a medicine reaction prediction module, wherein the centralized data management module is used for centrally managing clinical electronic medical history data and omics data of multiple medical institutes; the data preprocessing module is used for preprocessing the centrally managed data and used for establishing a relationship dependency net on the basis of biomedicine characteristics; the marker extraction module is used for extracting characteristics genes of patients to obtain marker sets on the basis of preprocessed data; the subtype classification moduleis used for classifying subtypes of the patients and used for confirming corresponding groups of the patients; the medicine reaction prediction module is used for establishing medicine reaction prediction models and used for predicting reactions of the patients upon different medicines according to the medicine reaction prediction models. Compared with the prior art, the system is capable of achieving effective management on medical data and medicine reaction prediction, and intelligentization can be achieved.

The invention belongs to the technical field of medicines, and particularly relates to a spirit inhibition medicine recommendation method and system based on mental disorder subtype classification. The technology of artificial intelligence and machine learning is adopted, and data mining and analysis are carried out on data of a mental disorder diagnosis and treatment scale, and classifying the given mental disorder. The method specifically comprises the steps of evaluating double-cluster subtype classification of the data based on the mental disorder quantity table, establishing a post-treatment evaluation prediction model by using the machine learning algorithm and the scale data before treatment, analyzing and recommending the schizophrenia medicine, preliminarily judging each schizophrenia patient and disease, dividing the subtype according to the table grading data and the patient information, respectively executing an evaluation and prediction model for each candidate medicine according to the scale index component data and the basic information of the patient, and taking the medicines which allows entering into an optimal-effect subtype after treatment and / or medicines withthe maximum reduce total value of the scale as recommended medicines. The accurate and objective disease-assisted diagnosis and treatment can be realized, and the curative effect is improved.

The invention discloses a method and a system for determining cancer network markers based on a probability model. The method comprises the following steps: utilizing a probability density function, converting gene expression data matrices of all normal samples and disease samples obtained into likelihood matrices, and constructing a normal sample distribution function according to the likelihoodmatrices of all normal samples; each element in the likelihood matrix of each disease sample is then brought into the normal sample distribution function, the set of significantly different genes foreach disease sample is determined, and the set of significantly different genes for each disease sample is mapped to proteins. Protein-protein interaction networks, and network markers for each disease sample are identified. The method or system provided by the invention can accurately and effectively acquire cancer network markers, and utilize the cancer network markers to carry out subtype classification of diseases so as to realize precise diagnosis and treatment of diseases.

The invention relates to a neutralizing monoclonal antibody 12C11 of a human nerve growth factor and its hybridoma cell strain, which belong to the immunology field. The neutralizing monoclonal antibody 12C11 of the human nerve growth factor is secreted by the hybridoma cell with a preservation number of CGMCC No.8772, the neutralizing monoclonal antibody and the human nerve growth factor have neutralization reaction effect, and the subtype classification can be a heavy chain IgG1 and a light chain Kappa. The Neutralizing monoclonal antibody has specificity and neutralizing activity, and can be used as an antagonist for the human nerve growth factor.

The invention provides a cancer subtype classification method based on multi-omics integration. The method comprises the steps of obtaining target multi-omics data of each patient in a target cancer patient group; calculating to obtain an omics similarity matrix; predicting each omics similarity matrix to obtain a predicted similarity matrix; correcting the prediction similarity matrix by utilizing the omics similarity matrix to obtain a correction matrix; performing weighted fusion to obtain a fusion matrix; and performing spectral clustering on the fusion matrix, and establishing a cancer subtype category label corresponding to the fusion matrix of each patient. The accuracy of classification and evaluation of the cancer subtypes is improved, the patients are classified through a more flexible integration method, the data analysis efficiency is improved, and convenience is provided for research on the cancer subtypes.

The invention relates to a breast cancer molecular typing method, device and system based on unsupervised learning. The method comprises the following steps: obtaining a to-be-predicted breast DCE-MRI image, and extracting regions of interest of sequence images of various specifications in the image; utilizing a molecular subtype prediction model obtained by adopting unsupervised learning training to predict and obtain molecular subtype classification probabilities corresponding to various sequence images; adopting ensemble learning fusion to obtain a final corresponding molecular subtype classification result. When the molecular typing prediction model is trained, through the thought of an unsupervised learning pre-training network and a transfer learning fine tuning network, a breast benign tumor image is fully utilized to construct an unlabeled source domain data set in the previous stage, and the feature extraction capability of the model is enhanced; and in the later stage, a target domain data set with a label is constructed by adopting the breast malignant tumor image to carry out fine tuning on the model with the pre-training weight. Compared with the prior art, the prediction precision of breast cancer molecular typing is remarkably improved.

The invention discloses a non-small cell lung cancer subtype classification system based on multi-view deep learning, which realizes classification of lung adenocarcinoma and lung squamous carcinoma based on a CT image, generates and displays image data of an examined part based on volume data generated by performing lung CT on an examined body, and obtains non-small cell lung cancer subtype classification result of the examined body. The non-small cell lung cancer pathological subtype classification system comprises an information acquisition module, a network module and a training module. According to the system, automatic lung field segmentation can be realized, interlayer information and three-dimensional information carried by a CT image are fully utilized by a multi-view model, an obtained classification model can be used as a tool for assisting a doctor through automatic feature extraction and classification training, and the system has the characteristics of high automation and high practicability.

The invention relates to DNA microarray chips, in particular to DNA microarray chips under human papilloma virus (HPV) high-risk and low-risk subtype classification. A solid-phase carrier substrate is adopted and matched with oligonucleotide probes of each HPV subtype so as to prepare the DNA microarray chips, which can simultaneously detect the HPV high-risk and low-risk subtypes of multiple samples at one time in high sensitivity and high specificity, and can be widely applied into qualitative and subtype detection of HPV infection.

The invention provides a dMMR subtype classification method based on a pathological image, and the method comprises the steps: dividing a labeled pathological image, obtaining a plurality of known image blocks, constructing a data set, training a deep learning network through the data set, and obtaining a region-of-interest recognition model; obtaining a target image region of a target pathological image through the region-of-interest recognition model; dividing the target image area into a plurality of target image blocks with the same size, averagely dividing all the target image blocks into a plurality of parts, and carrying out stacking and merging operation on each part of the target image blocks to obtain stacked blocks; and obtaining feature representations of all the stacked blocks, and obtaining a dMMR subtype classification result of the target pathological image according to all the feature representations. The invention further provides a dMMR subtype classification system based on the pathological image and a data processing device used for carrying out dMMR subtype classification on the target pathological image.

The present invention relates to methods for identifying the susceptibility or predisposition of an individual to schizophrenia SZ) or to clusters of symptoms associated thereof. Particularly, the tests can be performed before or after the disorders appear. More particularly, the present invention relates to the determination of epistatic effect of at least two genotype related loci. In addition, the invention provides tests for the classification of different subtypes predicting severity of illness of patients affected by or predisposed to SZ or associated clusters of symptoms.

Disclosed in the invention are a pancreatic neuroendocrine tumour susceptibility gene loci and detection methods and kits. In particular, disclosed in the invention is a pancreatic neuroendocrine tumour susceptibility gene loci based on whole genome sequencing findings, the loci is a high frequency T372R somatic mutation of a YY1 (Yin Yang1) gene. Also disclosed in the invention are methods for the detection of pancreatic neuroendocrine tumour susceptibility and pancreatic neuroendocrine tumour subtype classification, and corresponding detection kits.

The invention provides a group of long-chain non-coding RNA (lncRNA) markers for tumor immune typing, and the group of long-chain non-coding RNA markers totally comprises 49 lncRNAs. The invention also provides a tumor immune typing method based on the above lncRNA markers. According to the method, 49 lncRNAs are grouped and are combined with the expression quantity of cytotoxic T lymphocytes (CTL) to further classify patients into four groups, including an immune-active class, an immune-exclusion class, an immune-dysfunctional class and an immune-desert class by means of hierarchical clustering. The tumor immune subtype classification method provided by the invention has very strong correlation with prediction of the overall survival rate and the reactivity of anti-tumor immunotherapy, and has the characteristics of high objectivity, accuracy and repeatability. Establishment of the classification method provides a prediction index for the curative effect and the prognosis of the anti-immunotherapy of malignant tumor patients.

The invention discloses a cancer subtype classification method for performing multi-omics data integration based on a D-S evidence theory, and belongs to the technical field of multi-omics data analysis. The cancer subtype classification method comprises the steps: firstly obtaining similarity matrixes of patients under different omics backgrounds, then integrating the similarity matrixes based ona D-S evidence theory to obtain final similarity measurement, and finally performing clustering on the similarity matrixes to obtain a clustering result which is divided into different subtypes. Thecancer subtype classification method is used for dividing cancer subtypes, uses various different omics data, and is more stable and credible and has biological significance when compared with a single omics method. In addition, the cancer subtype classification method applies an integration mode based on a D-S evidence theory, so that a reasonable fusion means for data from different sources is provided. The cancer subtype classification method has a good comprehensive analysis effect on multi-omics data, and is of great significance to division of cancer subtypes.

The invention discloses a fine-grained cancer subtype classification method. The method comprises the following steps: step 1, obtaining a cell nucleus segmentation and classification result in a pathological picture; step 2, extracting an instance patch according to the cell nucleus segmentation and classification prediction result; step 3, performing instance patch convolution feature extraction; and step 4, generating picture-level features of the used instance patch through a Transform model to complete classification of cancer subtypes. The method can assist pathologists to classify cancer subtypes, and the working efficiency of the doctors is improved.

The invention provides a classification method and system of immune-related disease molecular typing and a subtype classifier, and the method comprises the steps: carrying out the molecular typing in a training set through a clustering algorithm, and obtaining a plurality of subtypes stably appearing in the training set and a marker gene of each subtype; carrying out enrichment analysis on marker genes of the subtypes, carrying out immune cell infiltration evaluation on the subtypes, and obtaining various subtype categories with stable immune characteristics according to analysis and evaluation results; comparing the treatment response rates of different subtype categories through the comparison set, and determining the subtype category needing to be recognized; constructing a support vector machine model by utilizing the feature genes obtained by screening and the optimal parameter combination; and identifying whether the to-be-classified immune-related disease data is a required subtype category or not. According to the invention, immune-related disease subtypes with stable characteristics can be identified, so that accurate drug selection and economic treatment are facilitated. The invention is applicable to molecular typing and subtype classification of various immune-related diseases, and is not limited to the embodiments herein.

The invention relates to an application of a composition of variable sites selected from an EB virus genome to preparing of an EB virus typing kit, and the EB virus typing kit based on detection of the variable sites, in particular to an application of a composition of five variable sites selected from an EB virus genome to preparing of an EB virus typing kit. The variable sites include loc613, 1oc632, loc7048, loc44531 and 1oc146628. The EB virus typing kit based on detection of the variable sites comprises a PCR amplification primer and a single-base extension primer of the composition of the five variable sites of the EB virus genome, and the variable sites include loc613, 1oc632, loc7048, loc44531 and 1oc146628. According to the EB virus typing kit, an EB virus subtype classification standard can be built, and the relative risk degree for individuals to suffer from nasopharyngeal carcinoma is estimated; the EB virus typing kit is suitable for screening of people in a high nasopharyngeal-carcinoma incidence area, and has the positive guiding significance and the important application value for tumor preventing, tumor early diagnosing and early curing and prognosis improving.

The invention belongs to the technical field of medicine, and specifically relates to a method and system for recommending antipsychotic drugs based on subtype classification of mental disorders. The present invention uses artificial intelligence and machine learning technology to classify given mental disorders through mining and analysis of mental disorder diagnosis and treatment scale data; the present invention specifically includes: bi-clustering subtype classification based on mental disorder scale evaluation data ; Use machine learning algorithms and scale data before treatment to establish a post-treatment evaluation prediction model; Drug analysis recommendation for schizophrenia: For each schizophrenia patient and initial disease judgment, use the scale scoring data and patient information to divide its subtype; according to the patient's scale index component data and the patient's basic information, the evaluation prediction model is executed for each candidate drug; the drug that can enter the superior subtype and / or the total score of the scale is reduced the most after treatment is used as Drugs are recommended. The invention can realize accurate and objective auxiliary diagnosis and treatment for diseases and improve curative effect.

The invention relates to an application of a composition of variable sites selected from an EB virus genome to preparing of an EB virus typing kit, and the EB virus typing kit based on detection of the variable sites, in particular to an application of a composition of five variable sites selected from an EB virus genome to preparing of an EB virus typing kit. The variable sites include loc613, 1oc632, loc7048, loc44531 and 1oc146628. The EB virus typing kit based on detection of the variable sites comprises a PCR amplification primer and a single-base extension primer of the composition of the five variable sites of the EB virus genome, and the variable sites include loc613, 1oc632, loc7048, loc44531 and 1oc146628. According to the EB virus typing kit, an EB virus subtype classification standard can be built, and the relative risk degree for individuals to suffer from nasopharyngeal carcinoma is estimated; the EB virus typing kit is suitable for screening of people in a high nasopharyngeal-carcinoma incidence area, and has the positive guiding significance and the important application value for tumor preventing, tumor early diagnosing and early curing and prognosis improving.

Provided herein is a consensus molecular subtype (CMS) classifier for colorectal cancer patients. Also provided are methods of using the classifier to identify a clinically beneficial therapeutic regime for each patient as well as methods of treating a patient accordingly Custom Nanostring code sets, which work on formalin-fixed, paraffin-embedded samples, are provide for use in determining the CMS for a colorectal cancer patient.

The invention discloses a subtype classification algorithm of intracranial hemorrhage applied to CT images based on bilinear pooling. The algorithm innovatively solves the problem of subtype classification of intracranial hemorrhage from the perspective of fine-grained classification, and is used in the CNN part. The network architecture of compact bilinear pooling has been improved, the feature expression ability of neural network has been improved, and the classification performance has been improved. Bilinear pooling has been proven to be an effective method for solving fine-grained classification problems. It treats features from two CNNs from different sources or CNNs from the same source as two different features by modeling high-order statistics. And calculate the outer product of the two features, and then perform feature fusion through the pooling operation to capture the relationship between different features, and then obtain a stronger global feature representation. The method uses image translation invariance to interactively model local paired features.

Described herein are gene signatures providing prognostic, diagnostic, treatment and molecular subtype classifications of ovarian cancers through generation of ovarian cancer disease signatures (OCDSs) that account for molecular heterogeneity present in gynecological cancers. An ovarian cancer fixed signature (OCFS) is described which relates to the core programming of disease development, in addition to an ovarian cancer stem cell (OCSC) signature. Development various disease signature, suggests personalized treatment strategies focused on molecular subtypes of gynecological cancers, such as triage tests for patients.

The invention relates to a precise intelligent diagnosis and treatment big data system which comprises a centralized data management module, a data preprocessing module, a marker extraction module, asubtype classification module and a medicine reaction prediction module, wherein the centralized data management module is used for centrally managing clinical electronic medical history data and omics data of multiple medical institutes; the data preprocessing module is used for preprocessing the centrally managed data and used for establishing a relationship dependency net on the basis of biomedicine characteristics; the marker extraction module is used for extracting characteristics genes of patients to obtain marker sets on the basis of preprocessed data; the subtype classification moduleis used for classifying subtypes of the patients and used for confirming corresponding groups of the patients; the medicine reaction prediction module is used for establishing medicine reaction prediction models and used for predicting reactions of the patients upon different medicines according to the medicine reaction prediction models. Compared with the prior art, the system is capable of achieving effective management on medical data and medicine reaction prediction, and intelligentization can be achieved.

The invention discloses a pancreaticobiliary ampulla carcinoma classification model generation method and an image classification method. The model generation method comprises the steps of constructing an initial classification model, wherein the initial classification model comprises an image preprocessing module, a cell segmentation module, a cell morphological feature extraction module and a classification module; labeling the collected bile duct cancer pathological sections and pancreatic cancer pathological sections to form a digital pathological image labeling library; dividing the digital pathological image annotation library into training set data and test set data according to a preset proportion, wherein the training set data comprises a bile duct cancer pathological section and a pancreatic cancer pathological section which are subjected to annotation processing; and training the initial classification model by adopting the training set data, and completing parameter adjustment of the initial classification model to obtain the pancreatic duct type ampullaria carcinoma classification model. According to the method, a classification model is constructed based on a full-slice digital pathological image of HE staining for the first time to perform subtype classification on pancreatic duct type peripheral ampulla cancer; that is, whether a tumor originates from a pancreatic duct (pancreatic cancer) or a bile duct (bile duct cancer) is judged.

The invention provides a subtype classification method and device for histopathologic sections, a medium and a terminal, and relates to the technical field of medical image diagnosis. The method comprises the following steps: extracting multi-scale histological characteristics of a histopathological section contained in a target image; mapping the multi-scale histological features and the pathological data features of the histopathological section to a three-dimensional feature space; in the three-dimensional feature space, determining the similarity between the multi-scale histological features and the pathological data features by using a formatter principle; and carrying out subtype classification on the histopathologic section based on the similarity. Through the scheme, the subtype classification accuracy of the histopathologic section can be improved, so that the diagnosis accuracy of a patient is improved.

The invention discloses a tumor immune subtype classification method and system, the method comprising: acquiring a sample data set including RNA-seq sequencing data and gene expression profile data of multiple tumor tissues; Corresponding microbial abundance data, immune cell ratio data, and immune-related gene expression data in the corresponding tumor tissue; use microbial abundance data, immune cell ratio data, and immune-related gene expression data as classification features to construct training sample data set, and use the SMOTE algorithm to expand the minority class, train the random forest model, and improve the random forest model in the form of weighting: increase the weight of the minority class, make the decision tree classifier focus on the minority class, and improve the minority class. class classification accuracy. The invention can improve the accuracy of tumor immune subtype classification and prediction, and provide a new target for tumor immunotherapy.