Tumor methylation typing and new subtype discovery method, device and program product based on multi-task learning

Abstract

The invention relates to the field of intelligent medical treatment, in particular to a tumor methylation typing and new subtype discovery method, device and program product based on multi-task learning. Comprising the following steps: S1, acquiring tumor methylation data; s2, transmitting the tumor methylation data to a first neural network and a second neural network which are trained in parallel, performing feature processing by the first neural network to obtain feature data, and performing classification by the second neural network to obtain classification types; s3, outputting the feature data and the classification type to a trained third neural network to obtain a classification predicted value, comparing the classification predicted value with a preset threshold value, when the classification predicted value is smaller than a preset threshold value interval, determining that the subtype is a new subtype, otherwise, determining that the subtype is a known type, and otherwise, determining that the subtype is a known type. New subtype detection can be carried out when tumor types are obtained, and good clinical value is achieved.

Description

Technical Field

[0001] The present application relates to the field of intelligent medical care, and specifically to a method, device, program product and computer-readable storage medium for tumor methylation typing and new subtype discovery based on multi-task learning. Background Art

[0002] Central nervous system (CNS) tumors are a highly complex and heterogeneous group of diseases, with approximately 100 different types defined by the WHO classification. These tumors range from benign to malignant, and their accurate diagnosis is crucial for the selection of treatment options and prognosis assessment for patients. DNA methylation, as an important epigenetic marker, not only reflects the developmental origin of the tumor, but can also be used to track the primary site of highly dedifferentiated metastatic cancers. DNA methylation analysis is highly robust and reproducible even with small sample sizes and poor quality.

[0003] At present, the diagnosis of brain tumors mainly relies on the following methods: 1. Traditional diagnostic methods: Histopathological diagnosis based on morphology is the current gold standard, but this method mainly relies on the subjective judgment of pathologists and has large inter-observer differences. At the same time, conventional molecular tests such as IDH mutation status, TERT promoter mutation status, BRAF mutation status, fluorescence in situ hybridization (FISH) detection and immunohistochemistry detection are also widely used in clinical practice, but these methods are often difficult to standardize and the results are not stable enough. 2. Existing machine learning methods: The machine learning methods currently used in tumor classification mainly include supervised learning and unsupervised learning. Although supervised learning methods can accurately classify known types, they cannot discover new subtypes; while unsupervised learning methods have the potential to discover new categories, they lack the ability to accurately identify known categories and have poor results. In addition, the existing active learning methods are inefficient in the sample annotation process and require a large number of experts to invest time in manual annotation. 3. Existing methylation typing methods: The current methylation typing methods are mostly focused on the prediction of central nervous system tumor types. These methods usually adopt a single analysis strategy, fail to effectively integrate multi-task learning methods, and fail to establish a reasonable mechanism for discovering new subtypes. At the same time, existing data processing solutions have obvious deficiencies in terms of automation, quality control, and standardization of result interpretation, making it difficult to support multi-center result verification.

[0004] All of the above-mentioned technical solutions have obvious limitations in practical applications, especially in terms of accuracy, repeatability, and discovery of new subtypes, which increase the uncertainty of clinical diagnostic results and affect the formulation of treatment plans. Summary of the invention

[0005] In view of the above problems, the present invention provides a tumor methylation typing and new subtype discovery method based on multi-task learning, which takes into account both known type recognition and new subtype discovery, and has good stability and repeatability. The typing method has important clinical application value, and specifically includes: S1, obtaining tumor methylation data; S2, inputting the tumor methylation data into a parallel trained first neural network and a second neural network, the first neural network performs feature processing to obtain feature data, and the second neural network performs classification to obtain a classification type; S3, inputting the feature data and the classification type into a trained third neural network to obtain a classification prediction value, and comparing the classification prediction value with a preset threshold value. When the classification prediction value is less than the preset threshold value, it is determined to be a new subtype, otherwise it is determined to be a known type.

[0006] When a new subtype is determined based on the classification prediction value, it also includes new subtype confirmation, inputting the feature data determined as the new subtype into the fourth neural network for classification to obtain a final classification result, and when the final classification result includes a new subtype, annotating the new subtype;

[0007] Optionally, the fourth neural network is a clustering network.

[0008] The first neural network includes N layers of fully connected modules, and the second neural network includes N layers of fully connected modules. Every two fully connected modules are connected by a feature exchange layer. The methylation data passes through the fully connected modules and the feature exchange layers alternately to obtain feature data in the first neural network and obtain classification types in the second neural network.

[0009] Optionally, the feature exchange layer of the first neural network also includes fusing features of the second neural network; after the first fully connected module of the first neural network outputs features, the output features of the first fully connected module of the second neural network are obtained and fused to obtain fused features, the fused features are input to the feature exchange layer of the first neural network for processing to obtain first exchange features, the first exchange features are input to the second fully connected module of the first neural network, and the feature exchange layer repeats the above process until the Nth fully connected module.

[0010] Optionally, the feature exchange layer of the second neural network also includes fusing features of the first neural network; after the first fully connected module of the second neural network outputs features, the output features of the first fully connected module of the first neural network are obtained and fused to obtain fused features, the fused features are input to the feature exchange layer of the second neural network for processing to obtain second exchange features, the second exchange features are input to the second fully connected module of the second neural network, and the feature exchange layer repeats the above process until the Nth fully connected module.

[0011] Optionally, the fully connected module includes a batch normalization layer and an activation layer.

[0012] Optionally, the loss function of the neural network adopts one or more of the following: cross entropy loss function, hinge loss function, Huber loss function.

[0013] Optionally, the loss function of the first neural network is The calculation method is as follows:

[0014] Among them, x is the target sample, x i is any one of the m noise samples, q θ is the encoder network, p represents the probability distribution, θ represents the parameters of the neural network, and x m represents the mth noise sample, and ξ represents the target sample.

[0015] Optionally, the first neural network model or the second neural network adopts one or more of the following: convolutional neural network, residual network, Transformer.

[0016] The third neural network adopts one or more of the following: GMM, DBSCAN, K-means clustering, hierarchical clustering, Bayesian non-parametric method, infinite Gaussian mixture model.

[0017] The method also includes model updating. When a new subtype type appears, the first neural network, the second neural network, and the third neural network are updated after accumulating sample data of the new subtype type to obtain updated first neural network, second neural network, and third neural network, characteristic data is obtained based on the updated first neural network, classification type data is obtained based on the updated second neural network, and whether there is a new subtype is determined based on the updated third neural network.

[0018] Optionally, the updating method updates the neural network through incremental learning.

[0019] Optionally, the model update also includes model verification, verifying the updated model performance through historical data, and when the model performance reaches or exceeds the historical model performance, the first neural network, the second neural network, and the third neural network are updated to obtain updated first neural network, second neural network, and third neural network.

[0020] The method also includes methylation data visualization, inputting the methylation data into a first neural network model to obtain feature data, wherein the feature data includes a horizontal coordinate and a vertical coordinate of the feature data, and performing visualization processing based on the horizontal coordinate and the vertical coordinate to obtain two-dimensional visualization data.

[0021] The horizontal and vertical coordinates represent the distribution positions of the classification types to which the methylation samples belong; the horizontal and vertical coordinates are obtained based on the coordinates of similar samples;

[0022] Optionally, the two-dimensional visualization data includes distribution areas of various types;

[0023] Optionally, during the training process, the first neural network is trained by obtaining the methylation data set and the type label to obtain the horizontal coordinate and the vertical coordinate of the methylation data in each type, and then the relationship between the classification type and the horizontal coordinate and the vertical coordinate is learned to obtain a mapping relationship, and the training is repeated until the loss function remains unchanged, thereby obtaining a trained first neural network;

[0024] Optionally, when methylation sample data is newly added, the newly added methylation sample data obtains a two-dimensional horizontal coordinate and a vertical coordinate through a mapping relationship in the first neural network.

[0025] The object of the present invention is to provide a computer program product, which includes a computer program or instructions, and the computer program or instructions are executed by a processor to implement the above-mentioned tumor methylation typing and new subtype discovery method based on multi-task learning.

[0026] The object of the present invention is to provide a computer device, which includes a memory, a processor, and a computer program or instructions stored on the memory, wherein the computer program or instructions are executed by the processor to implement the above-mentioned tumor methylation typing and new subtype discovery method based on multi-task learning.

[0027] The object of the present invention is to provide a computer-readable storage medium having a computer program or instructions stored thereon, wherein the computer program or instructions are executed by a processor to implement the above-mentioned tumor methylation typing and new subtype discovery method based on multi-task learning.

[0028] Advantages of the present invention:

[0029] 1. In order to solve the problem of accurately classifying known tumor subtypes while effectively discovering new types, the present invention proposes to use multi-task learning to perform tumor classification and feature processing through parallel neural networks, and then perform cluster analysis based on tumor type and feature data to determine whether new subtypes have appeared. This solution can simultaneously perform classification and new subtype monitoring, has good stability and repeatability, and has important clinical significance.

[0030] 2. The present invention visualizes the methylation typing, where the same types are clustered into a cluster and occupy one position. After adding new samples, in addition to the change in the number of classified types, the distribution position of the types in the visualized data also changes, which is not conducive to doctors observing the type attribution of the new sample classification. Therefore, the present invention maps the methylation data and the two-dimensional data to obtain the coordinates of the new sample in the two-dimensional visualized data, so that the distribution position of the type remains unchanged, and the classification of the new sample only presents a change in quantity, which helps doctors compare and obtain the type of the new sample.

[0031] 3 When a new subtype is detected, the present invention annotates the new subtype, reduces the workload of the manual sample annotation process, and improves the usability of the model.

[0032] 4. Regarding the detection of new subtypes, the present invention adds a new subtype confirmation process after distinguishing through clustering algorithm, thereby improving the accuracy and reliability of the model for new type identification.

[0033] 5. The network model of the present invention also includes a model update mechanism, which updates the model through incremental learning for newly added subtypes and samples, so that the model can complete self-optimization and iterative updates, has sustainable usability, and meets all classifications and new subtype detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0035] Figure 1 A schematic diagram of a process flow of a tumor methylation classification and new subtype discovery method based on multi-task learning provided in an embodiment of the present invention;

[0036] Figure 2 A schematic diagram of a tumor methylation classification and new subtype discovery system based on multi-task learning provided in an embodiment of the present invention;

[0037] Figure 3 A schematic diagram of a device for tumor methylation typing and new subtype discovery based on multi-task learning provided in an embodiment of the present invention;

[0038] Figure 4 A network structure diagram of tumor methylation typing and new subtype discovery based on multi-task learning provided in an embodiment of the present invention;

[0039] Figure 5 Comparison results between TADM2 and other classification algorithms provided by the embodiment of the present invention;

[0040] Figure 6 Comparison of the AUC results of the present invention and other algorithms provided in the embodiments of the present invention;

[0041] Figure 7 Comparison of the workload of TADM2 annotation provided by the embodiment of the present invention and annotation of the complete dataset;

[0042] Figure 8 A visualization diagram provided for an embodiment of the present invention;

[0043] Fig. 9 The comparison results of TAMD2 provided in the embodiment of the present invention and other visualization methods are shown. DETAILED DESCRIPTION

[0044] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0045] In some of the processes described in the specification and claims of the present invention and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this article or executed in parallel. The sequence numbers of the operations, such as S101, S102, etc., are only used to distinguish between different operations, and the sequence numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., do not represent the order of precedence, and do not limit the "first" and "second" to be different types.

[0046] Figure 1 The schematic diagram of the method for tumor methylation typing and new subtype discovery based on multi-task learning provided by the embodiment of the present invention specifically includes:

[0047] S1: Obtain tumor methylation data;

[0048] In one embodiment, the tumor includes one or more of the following: brain tumor, epithelial tumor, mesenchymal tumor, lymphoma, melanoma, breast cancer, lung cancer, colorectal cancer, prostate cancer, ovarian cancer, pancreatic cancer.

[0049] S2: inputting the tumor methylation data into a first neural network and a second neural network that are trained in parallel, wherein the first neural network performs feature processing to obtain feature data, and the second neural network performs classification to obtain classification types;

[0050] In one embodiment, the first neural network includes N layers of fully connected modules, and the second neural network includes N layers of fully connected modules. Every two fully connected modules are connected by a feature exchange layer. The methylation data passes through the fully connected modules and the feature exchange layers alternately to obtain feature data in the first neural network and obtain classification types in the second neural network.

[0051] In one embodiment, the feature exchange layer of the first neural network also includes fusing features of the second neural network; after the first fully connected module of the first neural network outputs features, the output features of the first fully connected module of the second neural network are obtained and fused to obtain fused features, the fused features are input to the feature exchange layer of the first neural network for processing to obtain first exchange features, the first exchange features are input to the second fully connected module of the first neural network, and the feature exchange layer repeats the above process until the Nth fully connected module.

[0052] In one embodiment, the feature exchange layer of the second neural network also includes fusing features of the first neural network; after the first fully connected module of the second neural network outputs features, the output features of the first fully connected module of the first neural network are obtained and fused to obtain fused features, the fused features are input to the feature exchange layer of the second neural network for processing to obtain second exchange features, the second exchange features are input to the second fully connected module of the second neural network, and the feature exchange layer repeats the above process until the Nth fully connected module.

[0053] In one embodiment, the fully connected module includes a batch normalization layer and an activation layer.

[0054] In one embodiment, the loss function of the neural network adopts one or more of the following: cross entropy loss function, hinge loss function, Huber loss function.

[0055] In one embodiment, the loss function of the first neural network is The calculation method is as follows:

[0056] Among them, x is the target sample, x i is any one of the m noise samples, q θ is the encoder network, p represents the probability distribution, θ represents the parameters of the neural network, and x m represents the mth noise sample, and ξ represents the target sample.

[0057] In one embodiment, the first neural network model or the second neural network adopts one or more of the following: convolutional neural network, residual network, Transformer.

[0058] S3: Input the feature data and classification type into a trained third neural network to obtain a classification prediction value, compare the classification prediction value with a preset threshold, and when the classification prediction value is less than the preset threshold interval, determine it as a new subtype, otherwise determine it as a known type.

[0059] In one embodiment, when a new subtype is determined based on the classification prediction value, it also includes new subtype confirmation, and the feature data determined as the new subtype is input into the fourth neural network for classification to obtain the final classification result. When the final classification result includes a new subtype, the new subtype is annotated.

[0060] In one embodiment, the fourth neural network is a clustering network.

[0061] In one embodiment, the third neural network adopts one or more of the following: GMM, DBSCAN, K-means clustering, hierarchical clustering, Bayesian non-parametric method, infinite Gaussian mixture model.

[0062] In one embodiment, the method also includes model updating. When a new subtype type appears, the first neural network, the second neural network, and the third neural network are updated after accumulating sample data of the new subtype type to obtain updated first neural network, second neural network, and third neural network, characteristic data is obtained based on the updated first neural network, classification type data is obtained based on the updated second neural network, and whether there is a new subtype is determined based on the updated third neural network.

[0063] In one embodiment, the updating method updates the neural network through incremental learning.

[0064] In one embodiment, the model update also includes model verification, which verifies the updated model performance through historical data. When the model performance reaches or exceeds the historical model performance, the first neural network, the second neural network, and the third neural network are updated to obtain updated first neural network, second neural network, and third neural network.

[0065] In a specific embodiment, the present invention proposes a tumor methylation classification model TADM2 (Tumor subtype Annotation and Discovery based on Multi-task learning of DNA Methylation), which mainly includes: a network architecture, such as Figure 4 As shown:

[0066] Input layer: receives methylation data (dimension: 32000);

[0067] Dimensionality reduction network path: 3-layer fully connected layer architecture (1000→500→250→100→2); each layer uses BatchNorm and ReLU activation functions, and the Dropout rate is 0.2 to prevent overfitting;

[0068] Classification network path: 3-layer fully connected layer architecture (1000→500→250→100→95); using the same normalization and activation function settings;

[0069] Cross-stitch Unit implementation (feature exchange layer): set a cross unit between every two main layers (fully connected architecture); implementation formula: Shared Task A = αAA(Task A) + αBA(Task B)

[0070] Shared Task B = αAB(Task A) + αBB(Task B)

[0071] Among them, αAA, αBA, αAB, αBB are learnable parameters;

[0072] Loss function design for dimensionality reduction network (first neural network): InfoNCE contrastive learning loss;

[0073] Among them, x is the target sample, x i is m noise samples, q θ is the encoder network.

[0074] The loss function of the classification network (second neural network), cross entropy classification loss, is used in the supervised learning branch to calculate the difference between the predicted category and the true label.

[0075] Mechanism of new subtype discovery:

[0076] 1. Rating distribution modeling: Use a single-component Gaussian mixture model (GMM) to build a rating distribution model for each known category, calculate the probability density of new samples, and convert it into a p-value.

[0077] 2. Adaptive clustering algorithm: parameter optimization range eps: 0.1-2.5, step size 0.1; min_samples: 1-9;

[0078] Optimization goal: maximize the silhouette score, clustering condition: form at least 2 clusters.

[0079] Self-optimizing system: 1. Data processing flow.

[0080] Input: methylation data matrix; preprocessing: data standardization; feature extraction: obtain low-dimensional representation through the trained model; classification prediction: obtain category prediction results and confidence.

[0081] 2. New subtype confirmation process.

[0082] Anomaly detection: Screening abnormal samples based on p-value; Cluster analysis: Using optimized DBSCAN parameters for clustering; Manual confirmation: Experts review clustering results.

[0083] 3. Model update mechanism.

[0084] Triggering conditions: accumulation of a certain number of newly annotated samples (new subtype annotated samples); Update method: incremental learning to maintain model stability; Verification mechanism: use historical data to verify the performance of the updated model.

[0085] In a specific embodiment, the classification accuracy of the present invention is significantly improved. Figure 5 As shown in the figure, the accuracy of the proposed method is about 5% higher than that of the existing commonly used method Random Forest, and is higher than that of other classification methods Xgboost and MLP. It can reliably discover new tumor subtypes, and the AUC value for new subtype discrimination reaches 0.930, as shown in the figure. Figure 6 In addition, Figure 7 The data that needs to be labeled is screened using TADM2. By comparing the training on the complete data set and the training on the data set screened by TADM2, the model prediction accuracy remains unchanged, but the overall amount of data required is greatly reduced, reducing the manual labeling workload by about 70%. The model can be continuously optimized and has strong adaptability.

[0086] In one embodiment, the method further includes visualizing methylation data, inputting the methylation data into a first neural network model to obtain feature data, wherein the feature data includes a horizontal coordinate and a vertical coordinate of the feature data, and performing visualization processing based on the horizontal coordinate and the vertical coordinate to obtain two-dimensional visualization data.

[0087] In one embodiment, the abscissa and ordinate represent the distribution position of the classification type to which the methylation sample belongs, and the abscissa and ordinate are obtained based on the coordinates of similar samples.

[0088] In one embodiment, the two-dimensional visualization data includes distribution areas of various types.

[0089] In one embodiment, during the training process, the first neural network performs self-supervised comparative learning on the acquired methylation data set to obtain the horizontal coordinates and vertical coordinates of each type of methylation data, and repeats the training until the loss function remains unchanged to obtain a trained first neural network.

[0090] In one embodiment, when newly added methylation sample data is added, the newly added methylation sample data obtains two-dimensional abscissas and ordinates through a first neural network, wherein the two-dimensional abscissas and ordinates of the newly added samples are obtained based on coordinates of similar samples.

[0091] In one embodiment, the methylation data is input into the trained first neural network to obtain the two-dimensional coordinates of the methylation data, the two-dimensional coordinates are obtained based on the coordinates of similar samples, and the two-dimensional coordinates are visualized, wherein the visualized data is the distribution area of ​​each type, such as Figure 8 shown.

[0092] When the type of the newly added sample is a known type, the two-dimensional coordinates of the distribution area of ​​the type to which it belongs are obtained according to the first neural network; when the type of the newly added sample is a new subtype, when the model is updated, the classification type and coordinate mapping relationship in the first neural network is learned and updated.

[0093] In a specific embodiment, the methylation data is reduced in dimension by a dimensionality reduction network (a first neural network) to obtain a horizontal coordinate and a vertical coordinate of the methylation data, and the distribution position of the methylation data is obtained based on the horizontal coordinate and the vertical coordinate, and is visualized by a two-dimensional coordinate position, such as Figure 8 As shown, the distribution positions of various types of methylation data can be intuitively obtained, and when new samples are added, the categories to which the new samples belong can be obtained by comparing before and after. In addition, compared with the existing cluster type visualization method (although the category of the new samples can be displayed after the new samples are added, the distribution area of ​​each type will change), the two-dimensional coordinates obtained by the dimensionality reduction network of the present invention are visualized, and the distribution area of ​​each type remains unchanged after being determined. After the new samples are added, they are displayed in the fixed area of ​​the type. Based on this method, doctors can intuitively and easily observe and compare the changes before and after the new samples are added. In addition, the TADM2 of the present invention is compared with UMAP and tSNE, as shown Fig. 9 As shown, TADM2 is far superior to UMAP and tSNE.

[0094] The disclosed embodiments of the present invention further provide a computer program product or system, including a computer program, which, when executed by a processor, implements the above-mentioned tumor methylation typing and new subtype discovery method steps based on multi-task learning.

[0095] Figure 2 A schematic diagram of a tumor methylation typing and new subtype discovery system based on multi-task learning provided in an embodiment of the present invention specifically includes: an acquisition unit: acquiring tumor methylation data;

[0096] Classification unit: inputting the tumor methylation data into a first neural network and a second neural network that are trained in parallel, wherein the first neural network performs feature processing to obtain feature data, and the second neural network performs classification to obtain classification types;

[0097] Judgment unit: input the feature data and classification type into a trained third neural network to obtain a classification prediction value, compare the classification prediction value with a preset threshold, and when the classification prediction value is less than the preset threshold interval, determine it as a new subtype, otherwise determine it as a known type.

[0098] Figure 3 A schematic diagram of a device for tumor methylation typing and new subtype discovery based on multi-task learning provided by an embodiment of the present invention specifically includes:

[0099] A memory and a processor; the memory is used to store program instructions; the processor is used to call program instructions, and when the program instructions are executed, any one of the above-mentioned tumor methylation typing and new subtype discovery methods based on multi-task learning is performed.

[0100] The disclosed embodiments of the present invention further provide a computer-readable storage medium storing a computer program, which, when executed by a processor, is any one of the above-mentioned methods for tumor methylation classification and new subtype discovery based on multi-task learning.

[0101] The verification results of this verification embodiment show that assigning inherent weights to indications can improve the performance of the method relative to the default setting. Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments, and will not be repeated here. In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the coupling or direct coupling or communication connection between each other shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms. The unit described as a separate component may or may not be physically separated, and the component displayed as a unit may or may not be a physical unit, that is, it may be located in one place, or it may be distributed on multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units. A person of ordinary skill in the art may understand that all or part of the steps in the various methods of the above-mentioned embodiments may be completed by instructing the relevant hardware through a program, and the program may be stored in a computer-readable storage medium, and the storage medium may include: a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, etc.

[0102] A person skilled in the art can understand that all or part of the steps in the above-mentioned embodiment method can be implemented by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium. The above-mentioned medium storage can be a read-only memory, a disk or an optical disk, etc.

[0103] The above is a detailed introduction to a computer device provided by the present invention. For a person skilled in the art, according to the concept of the embodiments of the present invention, there may be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A method for tumor methylation typing and new subtype discovery based on multi-task learning, characterized in that: include: S1. Obtain tumor methylation data; S2. Input the tumor methylation data into a first neural network and a second neural network that are trained in parallel, the first neural network performs feature processing to obtain feature data, and the second neural network performs classification to obtain a classification type; S3. Input the feature data and the classification type into a third neural network that has been trained to obtain a classification prediction value, and compare the classification prediction value with a preset threshold value. When the classification prediction value is less than the preset threshold value interval, it is determined to be a new subtype, otherwise it is determined to be a known type.

2. The method for tumor methylation typing and new subtype discovery based on multi-task learning according to claim 1, characterized in that: When a new subtype is determined based on the classification prediction value, it also includes new subtype confirmation, inputting the feature data determined as the new subtype into the fourth neural network for classification to obtain a final classification result, and when the final classification result includes a new subtype, annotating the new subtype; Optionally, the fourth neural network is a clustering network.

3. The method for tumor methylation typing and new subtype discovery based on multi-task learning according to claim 1, characterized in that: The first neural network includes N layers of fully connected modules, and the second neural network includes N layers of fully connected modules. Every two fully connected modules are connected by a feature exchange layer. The methylation data passes through the fully connected modules and the feature exchange layer alternately, and feature data is obtained in the first neural network, and classification types are obtained in the second neural network. Optionally, the feature exchange layer of the first neural network also includes the fusion of features of the second neural network. After the first fully connected module of the first neural network outputs features, the output features of the first fully connected module of the second neural network are obtained for fusion to obtain fusion features, and the fusion features are input to the feature exchange layer of the first neural network for processing to obtain first exchange features, and the first exchange features are input to the second fully connected module of the first neural network. The feature exchange layer repeats the above process until The Nth fully connected module; optionally, the feature exchange layer of the second neural network also includes the fusion of the features of the first neural network; after the first fully connected module of the second neural network outputs the features, the output features of the first fully connected module of the first neural network are obtained for fusion to obtain fusion features, the fusion features are input to the feature exchange layer of the second neural network for processing to obtain second exchange features, the second exchange features are input to the second fully connected module of the second neural network, and the feature exchange layer repeats the above process until the Nth fully connected module; optionally, the fully connected module includes a batch normalization layer and an activation layer; optionally, the loss function of the neural network adopts one or more of the following: cross entropy loss function, hinge loss function, Huber loss function; optionally, the loss function of the first neural network The calculation method is as follows: Among them, x is the target sample, x i is any one of the m noise samples, q θ is the encoder network, p represents the probability distribution, θ represents the parameters of the neural network, and x m represents the mth noise sample, ξ represents the target sample; optionally, the first neural network model or the second neural network adopts one or more of the following: convolutional neural network, residual network, Transformer.

4. The method for tumor methylation typing and new subtype discovery based on multi-task learning according to claim 1, characterized in that: The third neural network adopts one or more of the following: GMM, DBSCAN, K-means clustering, hierarchical clustering, Bayesian non-parametric method, infinite Gaussian mixture model.

5. The method for tumor methylation typing and new subtype discovery based on multi-task learning according to claim 1, characterized in that: The method also includes model updating. When a new subtype type appears, the first neural network, the second neural network, and the third neural network are updated after accumulating sample data of the new subtype type to obtain updated first neural network, second neural network, and third neural network, characteristic data are obtained based on the updated first neural network, classification type data are obtained based on the updated second neural network, and whether there is a new subtype is determined based on the updated third neural network; optionally, the updating method updates the neural network through incremental learning; optionally, the model updating also includes model verification, verifying the updated model performance through historical data, and when the model performance reaches or exceeds the historical model performance, the first neural network, the second neural network, and the third neural network are updated to obtain updated first neural network, second neural network, and third neural network.

6. The method for tumor methylation typing and new subtype discovery based on multi-task learning according to claim 1, characterized in that: The method also includes methylation data visualization, inputting the methylation data into a trained first neural network model to obtain feature data, wherein the feature data includes a horizontal coordinate and a vertical coordinate of the feature data, and performing visualization processing based on the horizontal coordinate and the vertical coordinate to obtain two-dimensional visualization data.

7. The method for tumor methylation typing and new subtype discovery based on multi-task learning according to claim 6, characterized in that: The abscissa and ordinate represent the distribution position of the classification type to which the methylation sample belongs; the abscissa and ordinate are obtained based on the coordinates of similar samples; optionally, the two-dimensional visualization data includes the distribution area of ​​each type; Optionally, during the training process, the first neural network performs self-supervised comparative learning on the acquired methylation data set to obtain the horizontal coordinates and vertical coordinates of the methylation data in each type, and repeats the training until the loss function remains unchanged, thereby obtaining a trained first neural network; Optionally, when newly added methylation sample data is added, the newly added methylation sample data obtains two-dimensional abscissas and ordinates through the first neural network, wherein the two-dimensional abscissas and ordinates of the newly added samples are obtained based on coordinates of similar samples.

8. A computer program product comprising a computer program or instructions, characterized in that: The computer program or instructions are executed by a processor to implement the tumor methylation typing and new subtype discovery method based on multi-task learning as described in any one of claims 1 to 7.

9. A computer device comprising a memory, a processor and a computer program or instruction stored in the memory, characterized in that: The computer program or instructions are executed by a processor to implement the tumor methylation typing and new subtype discovery method based on multi-task learning as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program or instruction stored thereon, characterized in that: The computer program or instructions are executed by a processor to implement the tumor methylation typing and new subtype discovery method based on multi-task learning as described in any one of claims 1 to 7.

Images