Shared unit-based circular RNA and disease association prediction model construction method, prediction system and prediction method

Abstract

The invention discloses a construction method, a prediction system and a prediction method of a circular RNA (Ribonucleic Acid) and disease association prediction model based on a shared unit, and association prediction is realized through the constructed circular RNA and disease association prediction model. Comprising the following steps: constructing a circular RNA meta-path network, a circular RNA similarity network, a disease meta-path network and a disease similarity network based on a known associated data set, and respectively carrying out feature extraction on the networks to obtain circular RNA similarity features, circular RNA meta-path features, disease similarity features and disease meta-path features; the similarity features and the meta-path features are simultaneously input to a sharing unit; similarity features output by the sharing unit are input into a multi-layer perceptron to carry out circular RNA and disease association prediction and update model parameters. A sharing unit is designed for the model, and the prediction performance of the model is improved by constructing a similarity network and a meta-path network of circular RNA and diseases and capturing potential cross-view information in a multi-view feature fusion process.

Description

Technical Field

[0001] The present invention belongs to the field of computer bioinformatics, and in particular relates to a construction method, a prediction system and a prediction method of a circular RNA and disease association prediction model based on a shared unit. Background Art

[0002] As a special type of non-coding RNA molecule, circular RNA plays a key role in a variety of biological processes in cells. They participate in regulating gene expression and signal transduction by interacting with molecules such as miRNA and proteins, thereby affecting the occurrence and development of diseases. Predicting the association between circular RNA and specific diseases can help us uncover the molecular veil of diseases and understand their complex biological networks. This understanding is the basis for developing new therapies, early diagnostic tools and personalized treatment plans. By identifying circular RNA associated with diseases, new biomarkers can be developed, which can be used for early identification of diseases, disease course monitoring and treatment effect evaluation, thereby improving the accuracy of treatment. In addition, this association information may also reveal new therapeutic targets and provide direction for the development of targeted drugs, especially in disease areas where traditional treatments have limited effects. At the same time, this predictive ability can also help us predict patients' responses to specific drugs and optimize clinical medication plans. More importantly, understanding the role of circular RNA in diseases can provide new strategies for disease prevention and reduce disease risks by regulating the expression of specific circular RNA. In summary, the prediction of circular RNA-disease association is not only of great significance to basic scientific research, but also has a profound impact on clinical practice, public health policy formulation and economic benefits. Through this prediction, the transformation from basic research to clinical application can be achieved, promoting the progress of medical science and ultimately improving human health.

[0003] At present, the methods for predicting the association between circular RNA (circRNA) and diseases mainly include the following categories: Network-based methods predict associations by building heterogeneous networks that make full use of different types of biological data. However, how to effectively integrate multiple types of data is a problem that requires in-depth research, and there is currently no good solution.

[0004] Based on the traditional machine learning method, this method predicts circRNA-disease associations by integrating different features combined with traditional machine learning methods, such as gradient decision trees and matrix decomposition, but this method relies on different similarity strategies when extracting features, and not all strategies are effective for the model.

[0005] Based on deep learning, this method can learn and predict potential disease-related circular RNAs, automatically extract high-level features contained in descriptors, and more accurately predict new circular RNA disease associations. However, existing deep learning-based methods fail to fully utilize the potential information between views and do not consider the importance differences between views, resulting in limited prediction performance. Summary of the invention

[0006] The purpose of the present invention is to propose a construction method, a prediction system and a prediction method for a circular RNA and disease association prediction model based on a shared unit in response to the problems existing in the prior art.

[0007] In order to achieve the above object, the present invention adopts the following technical solutions: A method for constructing a circular RNA and disease association prediction model based on shared units, comprising: Construct circular RNA metapathway network, circular RNA similarity network, disease metapathway network and disease similarity network based on known association datasets; The circular RNA similarity network, circular RNA metapath network, disease similarity network and disease metapath network were respectively extracted to obtain circular RNA similarity features, circular RNA metapath features, disease similarity features and disease metapath features; CircRNA similarity features and circRNA metapath features are simultaneously input into the shared unit for multi-view interaction and fusion; Disease similarity features and disease meta-path features are simultaneously input into the shared unit for multi-view interaction and fusion; The circular RNA similarity features and disease similarity features output by the shared unit are input into the multi-layer perceptron MLP to predict the association between circular RNA and disease; Based on the model loss function The model parameters are updated according to the association prediction results to enable the model to have association prediction capabilities.

[0008] In the above-mentioned method for constructing a circular RNA and disease association prediction model based on a shared unit, at least two of the circular RNA similarity networks are included, and the circular RNA similarity features of at least one circular RNA similarity network and the circular RNA metapath features are simultaneously input into the shared unit; The circular RNA similarity features after the shared unit and the other circular RNA similarity features are input into the multi-channel attention mechanism, and then input into the multi-layer perceptron after being processed by the multi-channel attention mechanism; At least two disease similarity networks are included, and the disease similarity features of at least one disease similarity network and the disease meta-path features are simultaneously input into the sharing unit; The disease similarity view features passing through the shared unit and the other disease similarity features are input into the multi-channel attention mechanism, and then input into the multi-layer perceptron after being processed by the multi-channel attention mechanism.

[0009] In the above-mentioned method for constructing a circular RNA and disease association prediction model based on shared units, the method also includes a comparative learning process: Construct positive and negative sample pairs in contrastive learning, select the features of the same circular RNA / the same disease in different views as positive sample pairs, and select the features of different entities in different views as negative sample pairs. In the process of updating model parameters, the contrastive learning loss function of circular RNA is used to calculate the positive and negative sample pairs. Comparative learning loss function with disease Increase the similarity between positive sample pairs and reduce the similarity between negative sample pairs; Based on the loss function during model training Update model parameters.

[0010] In the above-mentioned method for constructing a circular RNA and disease association prediction model based on shared units, the output of the multi-channel attention mechanism on the similarity features of multiple diseases and the disease meta-pathway features processed by the shared units participate in the comparative learning process of the disease; The output of the multi-channel attention mechanism on multiple circular RNA similarity features and the circular RNA meta-path features processed by shared units participate in the comparative learning process of circular RNA.

[0011] In the above-mentioned method for constructing a circular RNA and disease association prediction model based on shared units, two shared units are included, corresponding to the disease and the circular RNA respectively; CircRNA similarity features and circRNA metapath features were simultaneously input into a shared unit; Disease similarity features and disease metapath features are simultaneously input into another shared unit; Includes two multi-channel attention mechanisms, corresponding to diseases and circular RNAs respectively; The circular RNA similarity features after the shared unit and the other circular RNA similarity features are input into a multi-channel attention mechanism for processing; The disease similarity view features and other disease similarity features that pass through the shared unit are input into another multi-channel attention mechanism for processing.

[0012] In the above-mentioned method for constructing a circular RNA and disease association prediction model based on shared units, the circular RNA similarity network includes a circular RNA functional similarity network and a circular RNA GIP similarity network; The circular RNA GIP similarity features and the circular RNA metapath features extracted from the circular RNA GIP similarity network are simultaneously input into the shared unit; The disease similarity network includes a disease semantic similarity network and a disease GIP similarity network; The disease GIP similarity features and disease meta-path features extracted from the disease GIP similarity network are simultaneously input into the shared unit.

[0013] In the above-mentioned method for constructing a circular RNA and disease association prediction model based on shared units, the disease semantic similarity network is constructed in the following manner: in: and Represents diseases and diseases ancestral disease collection; Indicates ancestral disease right The semantic contribution of Indicates ancestral disease right The semantic contribution of The disease GIP similarity network is constructed as follows: in: and Disease and Column vector in the circRNA-disease association matrix; is the normalized bandwidth parameter; The circular RNA functional similarity network is constructed in the following manner: in: and Circular RNA and A collection of associated diseases; Indicates disease With Collection Similarity scores for all diseases in ; Indicates disease With Collection Similarity scores for all diseases in ; The circular RNA GIP similarity network was constructed as follows: and Circular RNA and row vector in the incidence matrix; is the normalized bandwidth parameter.

[0014] In the above-mentioned method for constructing a circular RNA and disease association prediction model based on shared units, constructing the circular RNA meta-pathway network based on a known association data set includes: The disease meta-pathway network is constructed based on known association data sets, including: in: is the correlation matrix of the known correlation data set; The association matrix between circular RNA and diseases The transpose of .

[0015] The above association data set is in the form of an association matrix, or the association data set is converted into an association matrix.

[0016] A circular RNA and disease association prediction system based on a shared unit includes a processor and a storage medium, wherein the storage medium stores a prediction model constructed by the above method, and the processor is used to execute the prediction model, that is, to execute the algorithm and calculation in the prediction model.

[0017] A method for predicting the association between circular RNA and diseases based on shared units, using the model constructed by the above method to perform association prediction, the input of the model includes the target circular RNA and target disease whose association relationship is to be predicted, the known association between the target circular RNA and the disease, and the known association between the target disease and the circular RNA, the model constructs a meta-path network and a similarity network of the target circular RNA, as well as a meta-path network and a similarity network of the target disease based on the input and performs feature extraction, and finally a multi-layer perceptron performs association prediction based on the extracted features and outputs a prediction result.

[0018] The advantages of the present invention are: 1) This scheme designs a shared unit for the model. By constructing a similarity network and a meta-path network between circular RNA and diseases, the model realizes information interaction and fusion between similarity views and meta-path views. By capturing potential cross-view information in the process of multi-view feature fusion, the model's performance in predicting circular RNA-disease associations is improved. 2) This scheme adopts a multi-channel attention mechanism while designing shared units, dynamically assigning weights according to the importance of features in different similarity networks, thereby effectively integrating the features of multiple similarity networks and improving the model's ability to model complex associations; 3) In addition, this scheme introduces a contrastive learning strategy to make full use of the complementary information between multiple views. By maximizing the similarity between positive samples and minimizing the similarity between negative samples, it further optimizes the feature representation and improves the prediction accuracy and generalization ability of the model. 4) This solution solves the problem that the model fails to fully utilize the potential information between different views, and proposes an optimization algorithm that combines multi-view interaction and multi-channel attention mechanism. Compared with existing methods, MSMCDA shows significant advantages in the accuracy and generalization ability of circular RNA-disease association prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 Shown is a flow chart of a method for predicting the association between circular RNA and diseases based on shared units provided by an embodiment of the present invention; Figure 2 The figure shows a schematic diagram of a method for predicting the association between circular RNA and diseases based on shared units provided in an embodiment of the present invention; Figure 3 Shown are the comparison results of AUC, AUPR, ACC and F1 of the present invention on 5 independent test data sets; Figure 4 The AUC comparison results of the present invention and the existing method on 5 training data sets are shown. Figure 5 It is a schematic diagram showing the comparison of the results of the ablation experiment performed on the CircR2disease dataset by the present invention; Figure 6 It is a schematic diagram showing the comparison of the results of the ablation experiment performed on the Circ2disease dataset by the present invention; Figure 7 It is a schematic diagram showing the comparison of the results of the ablation experiment performed on the CircRNAdisease dataset in the present invention; Figure 8 It is a schematic diagram showing the comparison of the results of the ablation experiment performed on the CircRDS dataset by the present invention; Fig. 9 Shown is a schematic diagram of the comparison of the results of the ablation experiment performed by the present invention on the dataset CircR2diseasev2.0. DETAILED DESCRIPTION

[0020] This scheme provides a circular RNA and disease association prediction method based on shared units, which can achieve higher performance association prediction between circular RNA and diseases. First, a method for constructing a circular RNA and disease association prediction model based on shared units for implementing this method is proposed. Secondly, a circular RNA and disease association prediction system based on shared units is proposed, which includes a storage medium and a processor for storing and executing the aforementioned constructed model.

[0021] For the convenience of description, the constructed model is called MSMCDA here. This method constructs multi-view features of circular RNA and diseases, combines shared units and attention mechanisms, and uses a contrastive learning strategy to optimize feature representation, thereby more accurately predicting the potential association between circular RNA and diseases. Figure 1 As shown, the method provided by this scheme includes: Collect and organize known association data sets between circular RNA and diseases in multiple public databases; Based on the obtained association datasets, similarity networks and meta-pathway networks of circular RNA, as well as similarity networks and meta-pathway networks of diseases were constructed.

[0022] The constructed similarity network and meta-path network are input into the shared unit, and the shared unit is used to promote information interaction between multi-view features, thereby capturing the potential information between different views and enhancing the information expression capability across views.

[0023] The multi-channel attention mechanism is used to adaptively adjust the importance weights of multi-view features of circular RNA and diseases to optimize the expression quality of each feature; A contrastive learning strategy is introduced to maximize the feature similarity of positive sample pairs from the same view, while minimizing the similarity between negative sample pairs across views, further enhancing the feature representation capability. Through the above processes, the characteristics of circular RNA and disease are extracted, and the multi-layer perceptron MLP performs association prediction based on these characteristics. Finally, based on the loss function The model parameters are updated according to the association prediction results to enable the model to have association prediction capabilities.

[0024] The trained model can be used for circular RNA-disease association prediction. Finally, this scheme verifies the prediction results of the model with actual biological data and compares them with other methods. The results show that this scheme is superior to other methods in all indicators and has the highest overall performance.

[0025] Specifically, Figure 2 As shown in the figure, the construction method of similarity network and meta-path network includes the following steps: (1) Similarity network construction In this embodiment, the disease similarity network includes a disease semantic similarity network and a disease Gaussian interaction kernel (GIP) similarity network.

[0026] Construct a disease semantic similarity network, use DOID (Disease Ontology Identifier, disease ontology identifier) ​​to retrieve the identifier (DOID) of each disease from the disease ontology, and calculate the disease and The semantic similarity between : (1) in, and Represents diseases and ancestral disease collection; Indicates ancestral disease right The semantic contribution of is calculated as: (2) In the disease semantic similarity network, the weight of the edge is the semantic similarity The value of .

[0027] Construct a disease GIP similarity network and use Gaussian interaction kernel (GIP) similarity to calculate the similarity between diseases: (3) in: and Disease and Column vector in the circRNA-disease association matrix; To normalize the bandwidth parameter, the calculation formula is: (4) in, is the number of columns of the incidence matrix, is the norm of each column vector.

[0028] In the disease GIP similarity network, the edge weight is the GIP similarity The value of .

[0029] In this embodiment, the similarity network of circular RNA includes a circular RNA function similarity network and a circular RNA GIP similarity network.

[0030] CircRNAs were calculated by and Functional similarity: (5) in: and Circular RNA and A collection of associated diseases; Indicates disease With Collection Similarity scores for all diseases in .

[0031] In the circular RNA functional similarity network, the weight of the edge is the functional similarity The value of .

[0032] The GIP similarity of circular RNA was calculated as follows: (6) and Circular RNA and row vector in the incidence matrix; To normalize the bandwidth parameter, the calculation formula is: (7) in, is the number of rows in the incidence matrix, is the norm of each row vector.

[0033] In the circular RNA GIP similarity network, the weight of the edge is the GIP similarity The value of .

[0034] (2) Meta-path construction: Constructing a metapathway network of circular RNA using association matrices : (8) in: is the correlation matrix; is the transpose of the association matrix between circular RNA and diseases.

[0035] Meta-path network The weight of the edge is the functional similarity The value of .

[0036] The meta-pathway network of the disease is also constructed using the association matrix : (9) Meta-path network The weight of the edge is the functional similarity The value of .

[0037] Furthermore, the multi-view feature interaction and fusion method of the shared unit includes the following process: A layer of convolution operation is performed on the circular RNA functional similarity network and the GIP similarity network respectively. The update feature calculation process of the two networks is as follows: (10) in, , is the adjacency matrix of the circular RNA functional similarity network or the GIP similarity network, is the added identity matrix; is the corresponding degree matrix; Represents learnable parameters is the activation function; Represents the circular RNA node features in the lth layer.

[0038] After a layer of GCN convolution for the disease semantic similarity network and the disease GIP similarity network, the updated features are calculated as follows: (11) in, , is the adjacency matrix of the semantic similarity network of the disease or the GIP similarity network, and I is the added identity matrix; is the corresponding degree matrix; represents a learnable parameter; is the activation function; Represents the disease node features in the lth layer.

[0039] Similarly, the two meta-path networks are also convolved, and the updated features of the meta-path network are calculated as follows: (12) in, , is the adjacency matrix of the meta-path network, and I is the added identity matrix; is the corresponding degree matrix; represents a learnable parameter; is the activation function; Indicates The disease node features or circular RNA node features in the layer.

[0040] For circular RNA and disease, the obtained GIP similarity features and their meta-path features are input into the shared unit together, where four independent linear modules are used to perform element-by-element operations on the multi-view features, adjust the weights of the features, and output the updated features. The feature updating process is as follows: (13) (14) in, and Represents the updated similarity view features and metapath view features of circular RNA or disease respectively; operator represents element-wise multiplication; , , and is a trainable parameter.

[0041] Furthermore, the specific method of optimizing feature weights by the multi-channel attention mechanism includes the following process: The various similarity view features of circular RNA are input into the multi-channel attention mechanism. In the multi-channel attention mechanism, the input similarity view features are first globally averaged and pooled, and then the weight coefficient is calculated. The calculation formula is as follows: (15) (16) in: Indicates functional similarity features; Indicates GIP similarity features; Represents a global average pooling operation.

[0042] and is the weight matrix of the fully connected layer; ReLU activation function. Represents the Sigmoid activation function.

[0043] Using the calculated weight coefficient , weighted processing of the similarity features of circular RNA, integrating view features: (17) in: Represents a two-dimensional convolutional neural network used to fuse multi-view features.

[0044] The multiple similarity view features of the disease are input into the multi-channel attention mechanism. In the multi-channel attention mechanism, the input similarity view features are first globally averaged and pooled, and then the weight coefficient of the disease similarity view is calculated: (18) (19) in: Represents semantic similarity features; Indicates GIP similarity features; Represents a global average pooling operation.

[0045] and is the weight matrix of the fully connected layer; ReLU activation function. Represents the Sigmoid activation function.

[0046] Using the calculated weight coefficient , weight the similarity features of the disease and integrate the features: (20) in: Represents a two-dimensional convolutional neural network used to fuse multi-view features.

[0047] The processing of the multi-layer perceptron is as follows: (twenty one) in: and Circular RNA and diseases The embedded features of , i.e., the corresponding circular RNA features and corresponding disease features processed by shared units, multi-channel attention mechanism and contrastive learning; represents element-by-element multiplication; FNN is a fully connected layer, which is used to extract features from the combined features; sigmoid is an activation function, which is used to normalize the prediction results to the range of [0, 1], indicating circular RNA and disease The probability score of the association.

[0048] Furthermore, the specific method of contrastive learning to optimize feature representation includes the following process: Construct positive and negative sample pairs in contrastive learning, select the features of the same entity (circRNA or disease) in different views as positive sample pairs, and select the features of different entities in different views as negative sample pairs. During the model update process, increase the similarity between positive sample pairs and reduce the similarity between negative sample pairs.

[0049] The contrastive learning loss function for circular RNA is calculated by the following formula: (twenty two) in: and Circular RNA features in similarity networks and metapath networks; represents the set of negative samples i≠j, sim(a,b) is the cosine similarity function, and the calculation formula is: (twenty three) The contrastive learning loss function for diseases has the same form as that for circular RNAs and is defined as: (twenty four) in: and Disease features in similarity networks and metapath networks; represents the set of negative samples i≠j, and sim(a,b) is the cosine similarity function.

[0050] Specifically, this model adopts the binary cross entropy loss function To optimize: (25) Where: Y + and Y - Represent the positive samples and negative samples in the training data respectively; when , the actual label ;when , the actual label .

[0051] The final loss function Combining binary cross entropy loss and contrastive learning loss: Furthermore, this scheme uses 5-fold cross validation to optimize the classification performance of the model. First, the training data is divided into five subsets, one subset is selected as the test set, and the remaining four subsets are selected as training sets. The process is repeated 5 times, and the average index is calculated to evaluate the model performance.

[0052] After training the circular RNA and disease association prediction model, the process of using the model for prediction is as follows: The model is input with the target circular RNA and target disease whose association is to be predicted, as well as the known association between the target circular RNA and the disease, and the known association between the target disease and the circular RNA. For example, taking circular RNA1 and disease 1 as an example, assuming that the association between circular RNA1 and disease 1 is to be predicted, and disease 1 is known to be associated with circular RNA2 and circular RNA3, and circular RNA1 is known to be associated with disease 2 and disease 3, then the following association matrix can be input: Disease 1 Disease 2 Disease 3 CircRNA1 To be tested 1 1 CircRNA2 1 0 1 CircRNA3 1 0 1 In the table, 1 indicates that the corresponding circular RNA is associated with the corresponding disease, and 0 indicates an unknown association.

[0053] The model constructs the meta-path network and similarity network of the target circular RNA based on the input, as well as the meta-path network and similarity network of the target disease and performs feature extraction. Finally, the multi-layer perceptron is used to Figure 2 The features extracted in step B are used for association prediction and the prediction results are output.

[0054] The proposed method MSMCDA is different from the traditional circular RNA and disease association prediction network. It introduces shared units to promote feature interaction between different views, achieve complementary integration in the fusion process, and capture potential cross-view information. In addition, the MSMCDA of this scheme uses the attention mechanism to assign different importance weights according to the features of different similar networks, thereby enhancing the model's adaptability to multi-view information. Finally, MSMCDA introduces a contrastive learning method to make full use of complementary information across views and further improve feature representation capabilities.

[0055] In order to verify the effectiveness and prediction accuracy of the proposed method, this scheme is used for verification on different test sets, and the performance is compared with six currently advanced methods such as AMHMDA, MDGF-MCEC, Bi-SGTAR, GMNN2CD, GraphCDA, and DMFCDA on the same data set. At the same time, ablation experiments are also carried out: In each figure, ACC represents the prediction accuracy, that is, the ratio of the number of samples correctly classified by the classifier to the total number of samples. F1-score represents the comprehensive evaluation of the accuracy and recall of the classifier. AUC is the area under the ROC curve, which can objectively express the classification ability of the model. AUPR is the area under the PR curve. In the prediction task of the association between circular RNA and disease, the associated circular RNA and disease pairs are positive samples, and the unassociated circular RNA and disease pairs are negative samples. Precision represents the proportion of all samples predicted to be positive that are actually positive, which measures how many of the samples predicted to be positive by the model are correct. Recall represents the proportion of all samples that are actually positive that are correctly predicted by the model.

[0056] Figure 3 This is the result of verifying this method on 5 independent test set data of different sizes. From the results, MSMCDA has high indicators on all 5 data sets, proving that the method proposed in this scheme has high generalization ability for data sets of different sizes.

[0057] Figure 4 The performance is compared with six advanced methods such as AMHMDA, MDGF-MCEC, Bi-SGTAR, GMNN2CD, GraphCDA, and DMFCDA on the same data set. As can be seen from the figure, MSMCDA has the highest AUC value on all data sets, which is better than other comparison methods, showing the effectiveness of the method of the present invention.

[0058] Figure 5-Figure 9 The following are the results of ablation experiments on the datasets CircR2disease, Circ2disease, CircRNAdisease, CircRDS, and CircR2diseasev2.0. The models include: MSMCDA, a circular RNA-disease association prediction model using multi-view shared units and contrastive learning strategy, also known as the method of this scheme; MSMCDA-noatten: MSMCDA model without the attention mechanism, where all views are considered equally important; MSMCDA-nocl: MSMCDA removes the contrastive learning model and directly uses the aggregated similarity features for prediction; MSMCDA-noshare: MSMCDA removes the shared unit model and directly applies two layers of GCN to the original input features without any feature sharing processing.

[0059] Depend on Figure 5-Figure 9 It can be seen that the various indicators of MSMCDA on the above five data sets are better than other methods. It can be seen that this scheme still has outstanding advantages and higher prediction performance compared with the current mainstream and advanced models.

[0060] The specific embodiments described herein are merely examples of the spirit of the present invention. Those skilled in the art may make various modifications or additions to the specific embodiments described or replace them in similar ways, but they will not deviate from the spirit of the present invention or exceed the scope defined by the appended claims.

Claims

1. A method for constructing a circular RNA and disease association prediction model based on shared units, characterized in that: include: Construct circular RNA metapathway network, circular RNA similarity network, disease metapathway network and disease similarity network based on known association datasets; The circular RNA similarity network, circular RNA metapath network, disease similarity network and disease metapath network were respectively extracted to obtain circular RNA similarity features, circular RNA metapath features, disease similarity features and disease metapath features; CircRNA similarity features and circRNA metapath features are simultaneously input into the shared unit for multi-view interaction and fusion; Disease similarity features and disease meta-path features are simultaneously input into the shared unit for multi-view interaction and fusion; The circular RNA similarity features and disease similarity features output by the shared unit are input into the multi-layer perceptron to predict the association between circular RNA and disease; Based on the model loss function The model parameters are updated according to the association prediction results to enable the model to have association prediction capabilities.

2. The method for constructing a circular RNA and disease association prediction model based on shared units according to claim 1, characterized in that: At least two of the circular RNA similarity networks are included, and the circular RNA similarity features of at least one circular RNA similarity network and the circular RNA metapath features are simultaneously input into the sharing unit; The circular RNA similarity features after the shared unit and the other circular RNA similarity features are input into the multi-channel attention mechanism, and then input into the multi-layer perceptron after being processed by the multi-channel attention mechanism; At least two disease similarity networks are included, and the disease similarity features of at least one disease similarity network and the disease meta-path features are simultaneously input into the sharing unit; The disease similarity view features passing through the shared unit and the other disease similarity features are input into the multi-channel attention mechanism, and then input into the multi-layer perceptron after being processed by the multi-channel attention mechanism.

3. The method for constructing a circular RNA and disease association prediction model based on shared units according to claim 2, characterized in that: The method also includes a contrastive learning process: Construct positive and negative sample pairs in contrastive learning, select the features of the same circular RNA / the same disease in different views as positive sample pairs, and select the features of different entities in different views as negative sample pairs. In the process of updating model parameters, the contrastive learning loss function of circular RNA is used to calculate the positive and negative sample pairs. Comparative learning loss function with disease Increase the similarity between positive sample pairs and reduce the similarity between negative sample pairs; Based on the loss function during model training Update model parameters.

4. The method for constructing a circular RNA and disease association prediction model based on shared units according to claim 3, characterized in that: The output of the multi-channel attention mechanism on the similarity features of multiple diseases and the disease meta-path features processed by the shared unit participate in the comparative learning process of the disease; The output of the multi-channel attention mechanism on the similarity features of multiple circular RNAs and the RNA meta-path features processed by shared units participate in the comparative learning process of circular RNA.

5. The method for constructing a circular RNA and disease association prediction model based on shared units according to any one of claims 2 to 4, characterized in that: It includes two shared units, corresponding to diseases and circular RNAs; CircRNA similarity features and circRNA metapath features were simultaneously input into a shared unit; Disease similarity features and disease metapath features are simultaneously input into another shared unit; Includes two multi-channel attention mechanisms, corresponding to diseases and circular RNAs respectively; The circular RNA similarity features after the shared unit and the other circular RNA similarity features are input into a multi-channel attention mechanism for processing; The disease similarity view features and other disease similarity features that pass through the shared unit are input into another multi-channel attention mechanism for processing.

6. The method for constructing a circular RNA and disease association prediction model based on shared units according to claim 2, characterized in that: The circular RNA similarity network includes a circular RNA functional similarity network and a circular RNA GIP similarity network; The circular RNA GIP similarity features and the circular RNA metapath features extracted from the circular RNA GIP similarity network are simultaneously input into the shared unit; The disease similarity network includes a disease semantic similarity network and a disease GIP similarity network; The disease GIP similarity features and disease meta-path features extracted from the disease GIP similarity network are simultaneously input into the shared unit.

7. The method for constructing a circular RNA and disease association prediction model based on shared units according to claim 6, characterized in that: The disease semantic similarity network is constructed as follows: in: and Represents diseases and diseases ancestral disease collection; Indicates ancestral disease right The semantic contribution of Indicates ancestral disease right The semantic contribution of The disease GIP similarity network is constructed as follows: in: and Disease and Column vector in the circRNA-disease association matrix; is the normalized bandwidth parameter; The circular RNA functional similarity network is constructed in the following manner: in: and Circular RNA and A collection of associated diseases; Indicates disease With Collection Similarity scores for all diseases in ; Indicates disease With Collection Similarity scores for all diseases in ; The circular RNA GIP similarity network was constructed as follows: and Circular RNA and row vector in the incidence matrix; is the normalized bandwidth parameter.

8. The method for constructing a circular RNA and disease association prediction model based on shared units according to claim 1, characterized in that: The circular RNA metapathway network is constructed based on a known association data set, including: The disease meta-pathway network is constructed based on known association data sets, including: in: is the correlation matrix of the known correlation data set; The association matrix between circular RNA and diseases The transpose of .

9. A circular RNA and disease association prediction system based on shared units, characterized in that: It comprises a processor and a storage medium, wherein the storage medium stores a prediction model constructed by the method described in any one of claims 1 to 8, and the processor is used to execute the prediction model.

10. A method for predicting the association between circular RNA and disease based on shared units, characterized in that: The model constructed by the method described in any one of claims 1 to 8 is used for association prediction, wherein the input of the model includes the target circular RNA and the target disease whose association relationship is to be predicted, as well as the known association between the target circular RNA and the disease, and the known association between the target disease and the circular RNA. Based on the input, the model constructs a meta-path network and a similarity network of the target circular RNA, as well as a meta-path network and a similarity network of the target disease, and performs feature extraction. Finally, a multi-layer perceptron performs association prediction based on the extracted features and outputs a prediction result.

Images