Molecular detection method and device, computer equipment, storage medium and program product

Abstract

The invention discloses a molecule detection method and device, computer equipment, a storage medium and a program product. The method comprises the following steps: acquiring an omics data set of a plurality of omics molecules; for each omics data set, performing correlation analysis on omics data in the omics data set to obtain intra-group correlation characteristics of the omics data set; performing correlation analysis on the plurality of intra-group correlation features to obtain inter-group correlation features of the plurality of intra-group correlation features; determining an influence value of each feature in the inter-group associated features on a preset omics model, and determining the molecule corresponding to the feature of which the influence value meets a preset condition as a target molecule; wherein the preset omics model is obtained based on inter-group association feature training of the plurality of omics data sets. By adopting the method, the molecular detection accuracy of the to-be-detected object can be improved.

Description

Technical Field

[0001] The present application relates to the field of bioinformatics, and in particular to a molecular detection method, apparatus, computer equipment, storage medium and program product. Background Art

[0002] With the development of personalized medical technology, technology based on model detection of user biological information has emerged. The model can detect important omics molecules based on the user's omics data and provide users with disease prediction, health assessment and other services.

[0003] At present, deep learning and other methods are often used to extract features and train models for multi-omics data. However, the accuracy of the models obtained based on deep learning is low. Summary of the invention

[0004] In view of the above-mentioned defects or deficiencies in the prior art, it is desirable to provide a molecular detection method, apparatus, computer equipment, storage medium and program product to improve the accuracy of molecular detection of an object to be detected.

[0005] In a first aspect, the present application provides a molecular detection method. The method comprises: Obtain omics data sets of multiple omics molecules; For each omics data set, correlation analysis is performed on the omics data in the omics data set to obtain the intra-group correlation features of the omics data set; correlation analysis is performed on multiple intra-group correlation features to obtain the inter-group correlation features of the multiple intra-group correlation features; Determine the influence value of each feature in the intergroup association feature on the preset omics model, and determine the molecule corresponding to the feature whose influence value meets the preset conditions as the target molecule; wherein the preset omics model is obtained by training based on the intergroup association features of multiple omics data sets.

[0006] In combination with the first aspect, in a possible implementation method, correlation analysis is performed on multiple intra-group correlation features to obtain inter-group correlation features of the multiple intra-group correlation features, including: fusing multiple intra-group correlation features to obtain fused features; and performing correlation analysis on feature data in the fused features to obtain inter-group correlation features.

[0007] In combination with the first aspect, in a possible implementation method, multiple intra-group correlation features are fused to obtain fused features, including: performing dimensionality reduction processing on each intra-group correlation feature to obtain reduced dimensionality features corresponding to each intra-group correlation feature; and fusing each reduced dimensionality feature to obtain a fused feature.

[0008] In combination with the first aspect, in a possible implementation method, the molecular detection method also includes: determining the total loss value of multiple omics data sets based on the dimensionality reduction features and the omics data sets corresponding to the dimensionality reduction features; adjusting the dimensionality reduction parameters according to the total loss value, and performing dimensionality reduction processing on the associated features within each group according to the adjusted dimensionality reduction parameters until the total loss value meets the first preset loss condition.

[0009] In combination with the first aspect, in a possible implementation method, the training process of the preset omics model includes: performing correlation analysis on multiple omics data sets to obtain inter-group correlation features, and clustering the inter-group correlation features to obtain predicted labels of molecules corresponding to each feature; determining the loss value of the preset omics model based on the predicted labels and the labeled labels; adjusting the model parameters of the preset omics model based on the loss value until the loss value meets the second preset loss condition.

[0010] In combination with the first aspect, in a possible implementation method, the influence value of each feature in the inter-group association feature on the preset omics model is determined, including: adjusting each feature data in the inter-group association feature to a preset value in turn to obtain an updated inter-group association feature; training the preset omics model according to the updated inter-group association feature to obtain an updated omics model; determining the influence value of each feature data according to the performance indicator difference value between the preset omics model and the updated omics model.

[0011] In a second aspect, the present application also provides a molecular detection device. The device comprises: An acquisition module, used to acquire omics data sets of multiple omics molecules; An analysis module is used to perform correlation analysis on the omics data in the omics data set for each omics data set to obtain the intra-group correlation features of the omics data set; perform correlation analysis on multiple intra-group correlation features to obtain the inter-group correlation features of the multiple intra-group correlation features; A determination module is used to determine the influence value of each feature in the inter-group association feature on the preset omics model, and determine the molecule corresponding to the feature whose influence value meets the preset conditions as the target molecule; wherein the preset omics model is obtained by training the inter-group association features of multiple omics data sets.

[0012] In a third aspect, the present application further provides a computer device, which includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the method described in the first aspect is implemented.

[0013] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the method described in the first aspect is implemented.

[0014] In a fifth aspect, the present application further provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, the method described in the first aspect is implemented.

[0015] The embodiments of the present application provide molecular detection methods, devices, computer equipment, storage media and program products, first collect multiple omics data sets of the object to be detected, and perform correlation analysis on the omics data in each omics data set to obtain multiple intra-group correlation features, and then perform correlation analysis on multiple intra-group correlation features to obtain inter-group correlation features. The embodiments of the present application sequentially realize the learning of the interaction relationship between molecules within a substance and between molecules of different substances through two correlation analyses, effectively improving the richness and accuracy of the feature data of the inter-group correlation features, and improving the accuracy of the preset omics model obtained by training based on the component correlation features. Furthermore, the target molecules determined based on the influence on the preset omics model are more accurate, that is, the accuracy of the molecular detection of the object to be detected is effectively improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Other features, objects and advantages of the present application will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings: Figure 1-6 is a schematic diagram of a molecular detection method in one embodiment; Figure 7 Another schematic diagram of a molecular detection method in one embodiment; Figure 8 It is a comparison chart of the test results of various molecular detection methods; Fig. 9 is a structural block diagram of a molecular detection device in one embodiment; Fig.10 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0017] The present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the relevant invention, rather than to limit the invention. It is also necessary to explain that, for ease of description, only the parts related to the invention are shown in the accompanying drawings.

[0018] It should be noted that, in the absence of conflict, the embodiments in the present application and the features in the embodiments can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments. In addition, the term "and / or" herein is merely a description of the association relationship of associated objects, indicating that three relationships may exist, for example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. The terms "first" and "second" in the description and claims of the embodiments of the present application are used to distinguish different objects, rather than to describe a specific order of objects.

[0019] With the development of personalized medical technology, technology based on model detection of user biological information has emerged. The model can detect important omics molecules based on the user's omics data and provide users with disease prediction, health assessment and other services.

[0020] At present, deep learning and other methods are often used to extract features and train models for multi-omics data. However, there are complex interactions between omics molecules of the same substance and between omics molecules of different substances. Therefore, the accuracy of the model obtained based on deep learning is relatively low.

[0021] In one embodiment, a molecular detection method is provided. This embodiment uses the method applied to a terminal as an example for illustration. It is understandable that the method can also be applied to a server, and can also be applied to a system including a terminal and a server, and is implemented through the interaction between the terminal and the server.

[0022] like Figure 1 As shown, the method comprises the following steps: Step 101: Obtain omics data sets of multiple omics molecules.

[0023] In an embodiment of the present application, body fluid samples such as blood and urine of the object to be detected can be first obtained, and routine blood tests, multi-omics sequencing, etc. can be performed on the obtained body fluid samples to obtain multi-omics data of the object to be detected, that is, a set of omics data corresponding to the omics molecules of multiple substances.

[0024] Among them, omics molecules can include genome, transcriptome, proteome, phenotype group, metabolome, etc. Each omics molecule corresponds to an omics data set.

[0025] In a possible implementation, after obtaining the multi-omics data, the multi-omics data can be preprocessed first. For example, the difference analysis of each omics data can be performed, that is, the P value of each omics data can be calculated by variance analysis, chi-square distribution, t-test, or Z-test, etc., and the omics data with a P value less than or equal to a preset threshold are determined, indicating that the substance (gene, protein, etc.) corresponding to the group of omics data has a significant difference relative to the normal state of the substance, and there may be pathological molecules, which require further molecular testing. Therefore, the group of omics data is retained, and the omics data with a P value greater than the preset threshold is discarded; wherein the preset threshold can be 0.05. Furthermore, the missing values ​​and abnormal values ​​in the retained omics data can also be filled with zeros.

[0026] Step 102: for each omics data set, performing correlation analysis on the omics data in the omics data set to obtain intra-group correlation features of the omics data set; performing correlation analysis on multiple intra-group correlation features to obtain inter-group correlation features of the multiple intra-group correlation features.

[0027] In the embodiments of the present application, considering the interaction relationship between the omics molecules of the same substance and between the omics molecules of different substances, correlation analysis can be performed on each omics data set first, and the correlation relationship between each omics data in each omics data set can be learned, so as to obtain the intra-group correlation feature of each omics data set, and each feature data in the intra-group correlation feature can represent the interaction relationship between each omics molecules of the corresponding substance.

[0028] Then, correlation analysis is performed on multiple intra-group correlation features, and the correlation relationship between the feature data in multiple different intra-group correlation features is learned, so as to obtain the inter-group correlation features of multiple intra-group correlation features. Each feature data in the inter-group correlation feature can simultaneously represent the interaction relationship between each group of molecules of the corresponding substance and the interaction relationship between the group of molecules of different substances.

[0029] In one possible implementation, the self-attention mechanism can be used to first perform correlation analysis on each omics data set. That is, for each omics data in the omics data set, the influence of the remaining omics data on the omics data is determined, and the similarity matrix is ​​calculated, as shown in the following equations (1)-(3): (1) (2) (3) in,

[0030] in, represents the query matrix corresponding to the i-th group of omics data; A linear transformation matrix representing the query value; represents the bond matrix corresponding to the i-th group of omics data; The linear transformation matrix representing the key value; represents the value matrix corresponding to the i-th group of omics data; represents the linear transformation matrix of the value; i represents the i-th omics molecule, that is, the i-th omics data set; It represents the dimensional characteristics of omics data; n represents the number of samples, that is, the amount of omics data; v represents the number of omics molecular species; d represents the dimension.

[0031] Then, the intra-group correlation features are calculated based on multiple similarity matrices, as shown in the following formula (4): (4) in, Represents the intra-group association characteristics of the i-th group of omics data sets.

[0032] Then, the self-attention mechanism is also used to perform correlation analysis on multiple intra-group related features. For each feature data in the multiple intra-group related features, the influence of the remaining feature data on the feature data is determined, or the influence of the remaining feature data on the feature data except the feature data of the intra-group related feature where the feature data is located is determined, so as to calculate the similarity matrix, as shown in the following equations (5)-(7): (5) (6) (7) in, Represents the query matrix corresponding to multiple groups of intra-group correlation features; A linear transformation matrix representing the query value; Represents the key matrix corresponding to multiple groups of intra-group correlation features; The linear transformation matrix representing the key value; Represents the value matrix corresponding to multiple groups of intra-group correlation features; A linear transformation matrix representing a value; It is an integrated representation of multiple groups of intra-group correlation features.

[0033] Furthermore, the inter-group correlation features of multiple groups of intra-group correlation features are calculated based on multiple similarity matrices, as shown in the following formula (8): (8) Step 103: determine the influence value of each feature in the inter-group association feature on the preset omics model, and determine the molecule corresponding to the feature whose influence value meets the preset conditions as the target molecule.

[0034] The preset omics model is obtained by training based on the inter-group association features of multiple omics data sets. The preset condition can be that the influence value is greater than a preset influence threshold, or it can be the top M positions after the influence values ​​are arranged in descending order.

[0035] In the embodiment of the present application, a preset omics model can be first trained with the inter-group association features as sample data. The preset omics model is trained based on the multi-omics data of the object to be detected, so the omics molecules contained in any multi-omics data of the object to be detected can be accurately classified.

[0036] Similarly, the influence of each feature data in the inter-group association feature used to train the preset omics model on the preset omics model can represent the importance of the feature data, and further, can represent the abnormality of the omics molecule corresponding to the feature data. Therefore, influence analysis can be performed on each feature data in the inter-group association feature to determine the influence value of each feature data on the preset omics model, and the omics molecule corresponding to the feature data with a larger influence value is determined as the target molecule. For example, the omics molecule corresponding to the influence value greater than the preset influence threshold is determined as the target molecule, or the omics molecule corresponding to the top M influence values ​​after the influence values ​​are arranged in descending order is determined as the target molecule.

[0037] The method provided in the embodiment of the present application first collects a multi-omics data set of the object to be detected, and performs a correlation analysis on the omics data in each omics data set to obtain multiple intra-group correlation features, and then performs a correlation analysis on multiple intra-group correlation features to obtain inter-group correlation features. The embodiment of the present application realizes the learning of the interaction relationship between molecules within a substance and between molecules of different substances in turn through two correlation analyses, effectively improving the richness and accuracy of the feature data of the inter-group correlation features, and improving the accuracy of the preset omics model obtained by training based on the component correlation features. Furthermore, the target molecules determined based on the influence on the preset omics model are more accurate, that is, the accuracy of the molecular detection of the object to be detected is effectively improved.

[0038] The above-mentioned embodiments introduce that multiple intra-group correlation features can be first fused into a feature data set, and then correlation analysis can be performed on the fused feature data set. For example, the above-mentioned "performing correlation analysis on multiple intra-group correlation features to obtain inter-group correlation features of multiple intra-group correlation features" includes the following: Figure 2 Steps shown: Step 201: fuse multiple intra-group related features to obtain fused features.

[0039] Step 202: Perform correlation analysis on the feature data in the fused features to obtain inter-group correlation features.

[0040] In the embodiment of the present application, in order to facilitate the correlation analysis of multiple intra-group correlation features, multiple intra-group correlation features can be fused first to merge multiple groups of data into one group of data. For example, multiple intra-group correlation features can be spliced ​​to obtain fused features, and the splicing method is shown in the following formula (9): (9) Then, the fusion features By performing correlation analysis, you can obtain component correlation characteristics.

[0041] The method provided in the embodiment of the present application can first merge multiple intra-group correlation features into one feature data set, and perform correlation analysis on the merged feature data set, thereby improving the convenience of performing correlation analysis on multiple intra-group correlation features.

[0042] In one embodiment, in order to improve the efficiency of correlation analysis and shorten the time of acquiring fusion features, multiple intra-group correlation features can be subjected to dimensionality reduction processing before fusion processing. This embodiment includes the following steps: Figure 3 Steps shown: Step 301: Perform dimensionality reduction processing on the associated features within each group to obtain the reduced dimensionality features corresponding to the associated features within each group.

[0043] Step 302: perform fusion processing on each dimension reduction feature to obtain a fusion feature.

[0044] In the embodiment of the present application, a fully connected layer may be set to perform dimensionality reduction processing on the intra-group correlation features to reduce the data volume of the intra-group correlation features. Furthermore, to ensure that the data volume of the intra-group correlation features after dimensionality reduction processing is small enough or within a preset data volume range, each intra-group correlation feature may be processed through multiple fully connected layers, for example, each intra-group correlation feature is processed through two layers of fully connected layers to obtain the dimensionality reduction features corresponding to each intra-group correlation feature. Then, each dimensionality reduction feature is fused, for example, spliced, to obtain a fused feature.

[0045] The method provided in the embodiment of the present application first performs dimensionality reduction processing on the intra-group correlation features before fusing multiple intra-group correlation features, which effectively reduces the data volume of the intra-group correlation features, that is, reduces the data volume of the fused features, and effectively improves the correlation analysis efficiency of the fused features.

[0046] In one embodiment, in order to ensure the validity of the feature data in the dimension reduction feature, the dimension reduction feature and the data before dimension reduction can be analyzed for differences. If the difference is large, the dimension reduction process is performed again on the intra-group associated features until the difference meets expectations. This embodiment includes the following: Figure 4 Steps shown: Step 401: determine the total loss value of multiple omics data sets according to the dimensionality reduction features and the omics data sets corresponding to the dimensionality reduction features.

[0047] Step 402: adjust the dimension reduction parameters according to the total loss value, and perform dimension reduction processing on the associated features in each group according to the adjusted dimension reduction parameters until the total loss value meets the first preset loss condition.

[0048] Among them, the first preset loss condition may be that the total loss value is less than or equal to the first preset loss threshold.

[0049] In the embodiment of the present application, the dimensionality reduction features and the data before dimensionality reduction (for example, the dimensionality reduction features may correspond to the intra-group association features or the omics data set) may be differentially analyzed to determine the total loss values ​​of multiple groups of intra-group association features, as shown in the following formula (10): (10) in, Represents the data before dimensionality reduction; represents the dimension reduction feature; F represents the norm; Represents the total loss value, which can represent the difference between the dimensionality reduction features and the data before dimensionality reduction.

[0050] If the total loss value is greater than the first preset loss threshold, the parameters in the fully connected layer, that is, the dimensionality reduction parameters, are adjusted, and then the intra-group correlation features are re-reduced in dimension according to the fully connected layer after the parameter adjustment, and a difference analysis is performed between the reduced dimensionality features obtained after the re-reduction and the data before dimensionality reduction until the difference reaches the expected value, that is, the total loss value is less than or equal to the first preset loss threshold.

[0051] The method provided in the embodiment of the present application can perform a difference analysis on the reduced dimensionality features and the data before dimensionality reduction after the dimensionality reduction processing of the intra-group associated features, adjust the dimensionality reduction parameters when the difference is large, and re-perform dimensionality reduction processing on the intra-group associated features until the difference reaches the expected level. On the basis of reducing the amount of intra-group associated feature data, the validity of each feature data in the reduced dimensionality features is guaranteed, and the data validity and accuracy of the fused features obtained based on the fusion of the reduced dimensionality features are improved.

[0052] In one embodiment, a training process of a preset omics model is provided, which includes: Figure 5 Steps shown: Step 501: perform correlation analysis on multiple omics data sets to obtain inter-group correlation features, and perform clustering processing on the inter-group correlation features to obtain predicted labels of molecules corresponding to each feature.

[0053] Step 502: Determine the loss value of the preset omics model according to the predicted label and the labeled label.

[0054] Step 503: Adjust the model parameters of the preset omics model according to the loss value until the loss value meets the second preset loss condition.

[0055] Among them, the second preset loss condition may be that the loss value is less than or equal to a second preset loss threshold.

[0056] In the embodiment of the present application, a correlation analysis is performed on multiple omics data sets (i.e., multi-omics data) of the collected object to be detected to obtain inter-group correlation features. The correlation analysis process refers to the above embodiment, and this embodiment will not be repeated here.

[0057] After obtaining the inter-group association features, each feature data in the inter-group association features can be used as a training sample, and each feature data can be labeled to obtain a labeled label. The inter-group association features can be clustered to obtain a predicted label for each feature data in the inter-group association features. For example, the K-means algorithm can be used to cluster the inter-group association features. A preset omics model can be trained based on the training samples, the labeled labels, and the predicted labels.

[0058] Then, the difference analysis of the predicted labels and the annotated labels is performed to obtain the loss value of the preset omics model, as shown in the following equations (11)-(12): (11) (12) in, represents the number of samples in the training set, p represents the pth sample, represents the true label of the p-th sample, represents the predicted label of the p-th sample, represents the cross entropy loss, represents the mean square error loss, and a is a hyperparameter.

[0059] If the loss value L is greater than the second preset loss threshold, the inter-group association features are clustered again to obtain updated prediction labels, and the updated preset omics model is trained based on the training samples, the annotation labels and the updated prediction labels (this step completes the adjustment of the model parameters of the preset omics model). The loss value corresponding to the updated preset omics model is calculated until the loss value is less than or equal to the second preset loss threshold, and the training of the preset omics model is completed.

[0060] The above-mentioned embodiment introduces a scheme for determining the influence of each feature data in the inter-group association feature on the preset omics model. In another embodiment, the influence of each feature data on the preset omics model can be determined by the performance change of the preset omics model. For example, the above-mentioned “determining the influence value of each feature in the inter-group association feature on the preset omics model” includes the following: Figure 6Steps shown: Step 601: sequentially adjust each feature data in the inter-group association feature to a preset value to obtain an updated inter-group association feature.

[0061] Step 602: Train the preset omics model according to the updated inter-group association features to obtain an updated omics model.

[0062] Step 603: Determine the influence value of each characteristic data according to the performance index difference value between the preset omics model and the updated omics model.

[0063] Among them, the preset value can be 0.

[0064] In the embodiment of the present application, each feature data of the inter-group association feature is set to 0 in turn, and the other feature data remain unchanged, so as to obtain the updated inter-group association feature corresponding to each feature data. Then, based on the updated inter-group association feature, the updated preset omics model is retrained to obtain the updated preset omics model corresponding to each feature data.

[0065] Then, the performance parameters of each updated preset omics model are determined, and the performance indicator difference value is calculated with the performance parameters of the preset omics model to obtain the performance indicator difference value corresponding to each characteristic data. The performance indicator difference value can represent the influence of the characteristic data on the preset omics model. Therefore, the performance indicator difference value can be used as the influence value of the characteristic data.

[0066] The method provided in the embodiment of the present application can determine the influence value of each feature data of the inter-group association feature on the omics data model through performance changes, and then determine the important target molecules in the omics molecules corresponding to the multi-omics data based on the influence value, thereby improving the accuracy of molecular detection of the object to be detected.

[0067] In summary, the molecular detection method (i.e., Omicsformer model) provided in the embodiments of the present application can be used as follows Figure 7 As shown. Further, the classification effect of the preset omics model is tested, as shown in Figure 8 As shown, a comparison of experimental results based on the method described in the embodiments of the present application, (7) k-nearest neighbor (KNN) classifier, support vector machine SVM classifier, fully connected neural network classifier, classification based on Lasso regularization, Multi-Omics Graph Convolutional Networks (MOGONET), Multi-Omics Variational Autoencoders (MOVE) and DeepIMV is provided respectively.

[0068] Among them, the KNN method connects multi-omics data and classifies them using the KNN algorithm. The SVM method uses the SVM algorithm to classify multi-omics data. The fully connected neural network classifier method uses a fully connected deep network architecture and is trained with cross-entropy loss. The classification method based on Lasso regularization uses linear regression with Lasso regularization for classification, and assigns the highest probability prediction value to each sample as the final class prediction label. The MOGONET method uses graph convolutional networks for specific and cross-omics learning in multi-omics data classification. The MOVE method uses variational autoencoders to explore associations between omics layers and evaluate drug treatment effects. The DeepIMV method implements an information bottleneck framework that focuses on the edge and joint representation of omics to achieve optimal performance.

[0069] In order to evaluate the performance of the above models, three key indicators, Accuracy (ACC), F1score (F1), and Purity, were compared. Each method was executed multiple times (for example, 10 times), and the mean and variance of the results showed that the new model based on Omicsformer outperformed other state-of-the-art methods in the classification of multi-omics supervised tasks. This application has significant improvements in the above three performance indicators, with an average increase of 8.3%, 6.0% and 8.3% respectively, indicating the advantage of the Omicsformer-based model provided in the embodiments of this application in capturing complex relationships within and between groups in omics research.

[0070] Among them, ACC indicates the proportion of correctly classified samples to the total number of samples. The higher the value, the better the model performance. F1score (F1) is a weighted average of precision and recall. Precision indicates the proportion of samples predicted to be positive that are actually positive. Recall indicates the proportion of samples predicted to be positive that are actually positive. F1 score takes into account both precision and recall. Its value is generally between 0 and 1. The larger the value, the better the model performance. The closer Purity is to 1, the better the clustering result.

[0071] It should be noted that although the operations of the method of the present invention are described in a particular order in the accompanying drawings, this does not require or imply that the operations must be performed in this particular order, or that all the operations shown must be performed to achieve the desired results. On the contrary, the steps depicted in the flow chart can be performed in a different order. Additionally or alternatively, some steps can be omitted, multiple steps can be combined into one step, and / or one step can be decomposed into multiple steps.

[0072] Further references Fig. 9 , which shows an exemplary structural block diagram of an apparatus for molecular detection according to an embodiment of the present application.

[0073] In one embodiment, the molecular detection device includes: an acquisition module 701, an analysis module 702 and a determination module 703, wherein: The acquisition module 701 is used to acquire omics data sets of multiple omics molecules.

[0074] The analysis module 702 is used to perform correlation analysis on the omics data in the omics data set for each omics data set to obtain the intra-group correlation features of the omics data set; perform correlation analysis on multiple intra-group correlation features to obtain the inter-group correlation features of the multiple intra-group correlation features.

[0075] The determination module 703 is used to determine the influence value of each feature in the inter-group association feature on the preset omics model, and determine the molecule corresponding to the feature whose influence value meets the preset conditions as the target molecule; wherein the preset omics model is obtained by training the inter-group association features of multiple omics data sets.

[0076] In one embodiment, the analysis module 702 is specifically used to fuse multiple intra-group correlation features to obtain fused features; and perform correlation analysis on feature data in the fused features to obtain inter-group correlation features.

[0077] In one embodiment, the analysis module 702 is further used to perform dimensionality reduction processing on the associated features within each group to obtain the reduced dimensionality features corresponding to the associated features within each group; and perform fusion processing on the reduced dimensionality features to obtain fusion features.

[0078] In one embodiment, the molecular detection device also includes a feature optimization module 704, which is used to determine the total loss value of multiple omics data sets based on the dimensionality reduction features and the omics data sets corresponding to the dimensionality reduction features; adjust the dimensionality reduction parameters according to the total loss value, and perform dimensionality reduction processing on the associated features within each group according to the adjusted dimensionality reduction parameters until the total loss value meets the first preset loss condition.

[0079] In one embodiment, the training process of the preset omics model includes: performing correlation analysis on multiple omics data sets to obtain inter-group correlation features, and clustering the inter-group correlation features to obtain predicted labels of molecules corresponding to each feature; determining the loss value of the preset omics model based on the predicted labels and the labeled labels; and adjusting the model parameters of the preset omics model based on the loss value until the loss value meets the second preset loss condition.

[0080] In one embodiment, the determination module 703 is specifically used to adjust each feature data in the inter-group association feature to a preset value in turn to obtain an updated inter-group association feature; train the preset omics model according to the updated inter-group association feature to obtain an updated omics model; and determine the influence value of each feature data according to the performance indicator difference value between the preset omics model and the updated omics model.

[0081] It should be understood that the units or modules described in the molecular detection device are the same as those described in the reference Figure 1 The various steps in the described method correspond to each other. Thus, the operations and features described above for the method are also applicable to the molecular detection device and the units contained therein, and will not be repeated here. The molecular detection device can be pre-implemented in a browser or other security application of an electronic device, or can be loaded into a browser or its security application of an electronic device by downloading or the like. The corresponding units in the molecular detection device can cooperate with the units in the electronic device to implement the scheme of the embodiment of the present application.

[0082] Reference below Fig.10 , which shows a schematic diagram of the structure of a computer system 800 of a terminal device or server suitable for implementing an embodiment of the present application.

[0083] like Fig.10 As shown, the computer system 800 includes a central processing unit (CPU) 801, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 802 or a program loaded from a storage part 808 to a random access memory (RAM) 803. In the RAM 803, various programs and data required for the operation of the system 800 are also stored. The CPU 801, the ROM 802, and the RAM 803 are connected to each other through a bus 804. An input / output (I / O) interface 805 is also connected to the bus 804.

[0084] The following components are connected to the I / O interface 805: an input section 806 including a keyboard, a mouse, etc.; an output section 807 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 808 including a hard disk, etc.; and a communication section 809 including a network interface card such as a LAN card, a modem, etc. The communication section 809 performs communication processing via a network such as the Internet. A drive 810 is also connected to the I / O interface 805 as needed. A removable medium 811, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 810 as needed, so that a computer program read therefrom is installed into the storage section 808 as needed.

[0085] In particular, according to the embodiments of the present disclosure, the above reference Figure 1 The described processes may be implemented as computer software programs. For example, embodiments of the present disclosure include a computer program product comprising a computer program tangibly embodied on a machine-readable medium, the computer program comprising a computer program for executing Figure 1In such an embodiment, the computer program may be downloaded and installed from a network via the communication section 809 and / or installed from the removable medium 811.

[0086] It should be noted that the computer-readable medium shown in the present application may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium such as a computer-readable storage medium that can send, propagate or transmit a program for use by or in conjunction with an instruction execution system, device or device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0087] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present invention. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, program segment or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a different order from the order marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0088] The units or modules involved in the embodiments described in the present application may be implemented by software or hardware. The units or modules described may also be arranged in a processor. The names of these units or modules do not, in some cases, constitute limitations on the units or modules themselves.

[0089] As another aspect, the present application further provides a computer-readable storage medium, which may be included in the computer device described in the above embodiment, or may exist independently without being installed in the computer device. The above computer-readable storage medium stores one or more programs, and when the above programs are used by one or more processors to execute the method described in the present application. For example, it can be executed Figure 1 The individual steps of the method are shown.

[0090] The present application embodiment provides a computer program product, which includes instructions. When the instructions are executed, the method described in the present application embodiment is executed. For example, Figure 1 The individual steps of the method are shown.

[0091] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0092] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the inventive concept. For example, the above features are replaced with (but not limited to) technical features with similar functions disclosed in the present application.

Claims

1. A molecular detection method, characterized in that: The method comprises Obtain omics data sets of multiple omics molecules; For each of the omics data sets, performing correlation analysis on the omics data in the omics data set to obtain intra-group correlation features of the omics data set; Performing correlation analysis on multiple intra-group correlation features to obtain inter-group correlation features of the multiple intra-group correlation features; Determine the influence value of each feature in the inter-group association feature on the preset omics model, and determine the molecule corresponding to the feature whose influence value meets the preset conditions as the target molecule; wherein the preset omics model is obtained by training based on the inter-group association features of the multiple omics data sets.

2. The method according to claim 1, characterized in that The performing correlation analysis on the multiple intra-group correlation features to obtain the inter-group correlation features of the multiple intra-group correlation features includes: Performing fusion processing on the multiple intra-group correlation features to obtain fusion features; Perform correlation analysis on the feature data in the fused features to obtain the inter-group correlation features.

3. The method according to claim 2, characterized in that The fusing the multiple intra-group related features to obtain the fused features includes: Performing dimensionality reduction processing on each of the intra-group correlation features to obtain dimensionality reduction features corresponding to each of the intra-group correlation features; The dimensionality reduction features are fused to obtain the fused features.

4. The method according to claim 3, characterized in that The method further comprises: Determine a total loss value of the plurality of omics data sets according to the dimensionality reduction feature and the omics data set corresponding to the dimensionality reduction feature; The dimension reduction parameters are adjusted according to the total loss value, and the dimension reduction processing is performed on each of the intra-group associated features according to the adjusted dimension reduction parameters until the total loss value meets the first preset loss condition.

5. The method according to claim 1, characterized in that The training process of the preset omics model includes: Performing correlation analysis on the plurality of omics data sets to obtain the inter-group correlation features, and performing clustering processing on the inter-group correlation features to obtain predicted labels of molecules corresponding to each of the features; Determining a loss value of the preset omics model according to the predicted label and the annotated label; The model parameters of the preset omics model are adjusted according to the loss value until the loss value meets a second preset loss condition.

6. The method according to claim 1, characterized in that The determining of the influence value of each feature in the inter-group association feature on the preset omics model includes: Sequentially adjusting each feature data in the inter-group association feature to a preset value to obtain an updated inter-group association feature; Training the preset omics model according to the updated inter-group association features to obtain an updated omics model; The influence value of each characteristic data is determined according to the performance index difference value between the preset omics model and the updated omics model.

7. A molecular detection device, characterized in that: The device comprises: An acquisition module, used to acquire omics data sets of multiple omics molecules; An analysis module, for performing correlation analysis on the omics data in the omics data set for each of the omics data sets to obtain intra-group correlation features of the omics data set; performing correlation analysis on multiple intra-group correlation features to obtain inter-group correlation features of the multiple intra-group correlation features; A determination module is used to determine the influence value of each feature in the inter-group association feature on a preset omics model, and determine the molecule corresponding to the feature whose influence value meets the preset conditions as the target molecule; wherein the preset omics model is obtained by training based on the inter-group association features of the multiple omics data sets.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images