A method and system for mining the relationship between lncRNA and diseases

Abstract

This invention relates to a method and system for mining the relationship between lncRNAs and diseases. It involves constructing a lncRNA hypergraph and a disease hypergraph, connecting them based on known associations between diseases and lncRNAs, and constructing a heterogeneous hypergraph containing hyperedges of the same and different types. A starting node is randomly selected from the heterogeneous hypergraph. During a walk from the current node, a bias factor is determined based on the current and total walk steps, and a set of nodes belonging to the same hyperedge as the current node is identified. The probability of the current node moving to each node in the set is determined based on the type of the hyperedge connecting the current node and the nodes in the set, as well as the bias factor. The next node to be moved to by the current node is determined according to the probability, and the walk order is recorded. The next node to be moved to is taken as the current node, and this process continues until the total number of walk steps is reached. After multiple walks, the visit frequency of lncRNA nodes and disease nodes is counted, and the mining results are obtained based on the visit frequency.

Description

Technical Field

[0001] This invention relates to the field of data mining, specifically to a method and system for mining the relationship between lncRNA and disease. Background Technology

[0002] Long non-coding RNAs (lncRNAs) are a class of non-coding RNA molecules exceeding 200 nucleotides in length. These RNAs do not encode proteins but regulate gene expression through various molecular mechanisms. lncRNAs regulate gene expression and cellular function through multiple mechanisms, such as enhancer action, promoter interference, transcription factor regulation, and chromatin remodeling. Unlike traditional small non-coding RNAs (such as miRNAs), lncRNAs have more diverse and complex functions. lncRNAs can regulate gene expression at the transcriptional level, participating in processes such as chromatin modification, gene transcription regulation, RNA splicing, and mRNA stability. Furthermore, lncRNAs can form complex regulatory networks through interactions with proteins or RNA. lncRNAs play a crucial role in various diseases, particularly cancer, neurodegenerative diseases, and cardiovascular diseases. Traditional methods can only study a few lncRNAs at a time, making large-scale screening and validation difficult. While high-throughput sequencing technology provides a wealth of omics data, effectively integrating and analyzing this data to extract valuable information is crucial for exploring the relationship between lncRNAs and diseases. Summary of the Invention

[0003] The purpose of this application is to provide a method and system for mining the relationship between lncRNAs and diseases, which can mine hidden or concealed relationships between lncRNAs and diseases based on existing relationships between lncRNAs and diseases.

[0004] To solve the above-mentioned technical problems, the embodiments of this application are implemented as follows:

[0005] The first aspect provides a method for mining the relationship between lncRNAs and diseases, the method comprising:

[0006] Based on lncRNA similarity and disease similarity, lncRNA hypergraphs and disease hypergraphs are constructed respectively. Based on the known association between diseases and lncRNAs, the lncRNA hypergraphs and disease hypergraphs are connected, and a heterogeneous hypergraph containing hyperedges of the same type and hyperedges of different types is constructed.

[0007] A starting node is randomly selected from the heterogeneous hypergraph. When the current node is walking, the bias factor is determined based on the current number of steps and the total number of steps. The set of nodes that belong to the same hyperedge as the current node is determined. The probability of the current node moving to each node in the node set is determined based on the type of the hyperedge connected to the current node and the nodes in the set and the bias factor. The next node to be moved to by the current node is determined according to the probability. The walking order is recorded. The next node to be moved to is taken as the current node. This process continues until the total number of steps is reached.

[0008] After multiple walks, the access frequency of lncRNA nodes and disease nodes is counted, and the mining results are obtained based on the access frequency.

[0009] Optionally, determining the bias factor based on the current number of steps and the total number of steps specifically involves:

[0010] The bias factor is calculated using a Gaussian function, where the center value of the Gaussian function curve is half of the total number of walks.

[0011] Optionally, determining the probability of the current node transitioning to each node in the node set based on the type of the hyperedges connecting the current node and the nodes in the set, as well as the bias factor, specifically involves:

[0012] Calculate the adjustment parameters of the same type of hyperedge and the adjustment parameters of the different types of hyperedge based on the bias factor;

[0013] For any node in the node set, calculate the set of hyperedges shared by the node and the current node; traverse each hyperedge in the hyperedge set, if the hyperedge is of the same type, calculate the internal weight of the hyperedge of the same type, and obtain the hyperedge weight based on the internal weight of the hyperedge and the adjustment parameter of the hyperedge of the same type; if the hyperedge is a different type of hyperedge, calculate the external weight of the different type of hyperedge, and obtain the hyperedge weight based on the external weight of the hyperedge and the adjustment parameter of the different type of hyperedge; take the average of the sum of the hyperedge weights corresponding to each hyperedge in the hyperedge set as the intermediate probability of the current node being transferred to a node in the node set;

[0014] The intermediate probabilities of all nodes in the node set are normalized, and the normalized intermediate probabilities are used as the probability of the current node being transferred to the node set.

[0015] Optionally, the internal weights of the same type of hyperedges are calculated as follows:

[0016] Calculate the average similarity of nodes inside hyperedges of the same type, and use the average value as the weight of the hyperedges of the same type.

[0017] Optionally, the external weight of the heterogeneity hyperedge is specifically:

[0018] Obtain the strongest association strength between all disease nodes and all lncRNA nodes in the heterogeneity hyperedge, and use the strongest association strength as the external weight of the heterogeneity hyperedge.

[0019] In another aspect, a system for mining the relationship between lncRNAs and diseases is provided, the system comprising:

[0020] The hypergraph construction module is used to construct lncRNA hypergraphs and disease hypergraphs based on lncRNA similarity and disease similarity, respectively. It connects the lncRNA hypergraph and disease hypergraph based on the known association between diseases and lncRNAs, and constructs a heterogeneous hypergraph containing hyperedges of the same type and hyperedges of different types.

[0021] The walking module is used to randomly select a starting node from the heterogeneous hypergraph. When walking at the current node, it determines the bias factor based on the current walking step count and the total walking step count, and determines the set of nodes that belong to the same hyperedge as the current node. Based on the type of the hyperedge connected to the current node and the nodes in the set, as well as the bias factor, it determines the probability that the current node will move to each node in the node set. According to the probability, it determines the next node to be moved to by the current node, records the walking order, and takes the next node to be moved to as the current node, until the total walking step count is reached.

[0022] The mining module is used to count the access frequency of lncRNA nodes and disease nodes after multiple walks, and obtain mining results based on the access frequency.

[0023] Optionally, determining the bias factor based on the current number of steps and the total number of steps specifically involves:

[0024] The bias factor is calculated using a Gaussian function, where the center value of the Gaussian function curve is half of the total number of walks.

[0025] Optionally, determining the probability of the current node transitioning to each node in the node set based on the type of the hyperedges connecting the current node and the nodes in the set, as well as the bias factor, specifically involves:

[0026] Calculate the adjustment parameters of the same type of hyperedge and the adjustment parameters of the different types of hyperedge based on the bias factor;

[0027] For any node in the node set, calculate the set of hyperedges shared by the node and the current node; traverse each hyperedge in the hyperedge set, if the hyperedge is of the same type, calculate the internal weight of the hyperedge of the same type, and obtain the hyperedge weight based on the internal weight of the hyperedge and the adjustment parameter of the hyperedge of the same type; if the hyperedge is a different type of hyperedge, calculate the external weight of the different type of hyperedge, and obtain the hyperedge weight based on the external weight of the hyperedge and the adjustment parameter of the different type of hyperedge; take the average of the sum of the hyperedge weights corresponding to each hyperedge in the hyperedge set as the intermediate probability of the current node being transferred to a node in the node set;

[0028] The intermediate probabilities of all nodes in the node set are normalized, and the normalized intermediate probabilities are used as the probability of the current node being transferred to the node set.

[0029] Optionally, the internal weights of the same type of hyperedges are calculated as follows:

[0030] Calculate the average similarity of nodes inside hyperedges of the same type, and use the average value as the weight of the hyperedges of the same type.

[0031] Optionally, the external weight of the heterogeneity hyperedge is specifically:

[0032] Obtain the strongest association strength between all disease nodes and all lncRNA nodes in the heterogeneity hyperedge, and use the strongest association strength as the external weight of the heterogeneity hyperedge.

[0033] This invention explores the potential relationship between lncRNAs and diseases by constructing a heterogeneous hypergraph. First, a heterogeneous hypergraph containing both homogeneous and heterogeneous hyperedges is constructed. Then, a bias factor is used to regulate the migration bias. In the early stages of migration, the bias factor is small, and migration tends to occur within local neighborhoods. In the middle stages, the bias factor is large, and migration tends to move between lncRNAs and diseases. In the later stages, the bias factor decreases again, and migration tends to occur within local neighborhoods. This approach is more effective in uncovering the potential relationship between lncRNAs and diseases. Attached Figure Description

[0034] Figure 1 This is a flowchart illustrating Example 1;

[0035] Figure 2 A diagram showing the relationship between known diseases and lncRNAs is presented;

[0036] Figure 3 This is a schematic diagram of the superedge;

[0037] Figure 4 This is a schematic diagram of the walking sequence. Detailed Implementation

[0038] To enable those skilled in the art to better understand the technical solutions in this application, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this application.

[0039] Figure 1This is a flowchart of the method for mining the relationship between lncRNA and disease in the embodiments of this application. Figure 1 The method is executed by a computing device. In specific applications, the computing device has the ability to store code and execute programs, etc., and the embodiments of this application are not limited thereto. Figure 1 The methods include:

[0040] S101, construct lncRNA hypergraph and disease hypergraph based on lncRNA similarity and disease similarity respectively, connect lncRNA hypergraph and disease hypergraph based on known association between disease and lncRNA, and construct heterogeneous hypergraph containing hyperedges of the same type and hyperedges of different types;

[0041] There are relatively few existing lncRNAs with clearly defined disease relationships. Based on the principle that similar lncRNAs have similar functions and are associated with similar diseases, a lncRNA hypergraph is constructed based on lncRNA similarity. LncRNA similarity is calculated based on sequence similarity, functional similarity, and expression profile similarity. The more similar two lncRNAs are in sequence, function, or expression pattern, the higher their similarity. In one embodiment, the weights of sequence similarity and expression profile similarity are determined based on existing lncRNA similarity, and then the lncRNA similarity is calculated based on these weights. After obtaining the lncRNA similarity, similarity scores greater than a preset value are retained, and a lncRNA hypergraph is constructed. The hypergraph consists of nodes and hyperedges; compared to a regular graph structure, the hyperedges of a hypergraph can connect multiple nodes. Similar to lncRNA similarity, disease similarity is calculated based on the clinical manifestations, pathological features, and gene associations of diseases. The more similar the symptoms, causes, or related genes of two diseases, the higher their similarity. Similarly, similarity scores greater than a certain value are retained, and a disease hypergraph is constructed. In the lncRNA hypergraph, each node represents a lncRNA, and hyperedges connect similar lncRNAs; in the disease hypergraph, each node represents a disease, and hyperedges connect similar diseases.

[0042] The association between lncRNAs and disease has been confirmed through experiments or other methods. Figure 2The diagram illustrates the known relationships between diseases and lncRNAs, connecting two hypergraphs through these known associations. For example, if lncRNA 1 and lncRNA 2 are known to be associated with disease X, a hyperedge is created connecting nodes lncRNA 1 and lncRNA 2 in the lncRNA hypergraph to node X in the disease hypergraph. Besides constructing heterogeneous hyperedges based on diseases, heterogeneous hyperedges can also be based on lncRNAs, identifying diseases associated with the lncRNAs and connecting these lncRNAs to known associated diseases as a single hyperedge. The resulting hypergraph includes two types of hyperedges: those connecting nodes of the same type (lncRNA or disease) and those connecting nodes of different types. For example, a hyperedge connecting only lncRNAs or diseases is a homogeneous hyperedge; a hyperedge connecting both lncRNAs and diseases is a heterogeneous hyperedge. Figure 3 The superedges in the array are all of the same type.

[0043] S102, randomly select a starting node from the heterogeneous hypergraph. When the current node is walking, determine the bias factor based on the current walking step and the total walking step, and determine the set of nodes that belong to the same hyperedge as the current node. Determine the probability of the current node moving to each node in the node set based on the type of the hyperedge connected to the current node and the nodes in the set and the bias factor. Determine the node to which the current node will move next according to the probability, and record the walking order. Take the node to which the current node will move next as the current node, until the total walking step is reached.

[0044] In a random walk of a hypergraph, multiple walks are involved. Each random walk begins by selecting a starting node, typically randomly chosen from the hypergraph. During each walk, after determining the starting node, the current walk number and the total number of steps are determined. For example, if a walk has 5 steps and this is the 3rd step, the current step count is 3, and the total number of steps is 5. The bias factor is determined based on the current and total steps. A Gaussian function or activation function is used to calculate the bias factor, ensuring a smaller bias factor at the beginning and end of the walk, and a larger bias factor in the middle. In a preferred embodiment, a Gaussian function is used to calculate the bias factor, where the center value of the Gaussian function curve is half the total number of steps. To avoid a bias factor of 1, in another embodiment, the value obtained from the Gaussian function is scaled to the range [0.1, 0.9].

[0045] In a hypergraph, a node can belong to multiple hyperedges. All other nodes on the same hyperedge as the current node constitute the set of nodes to be traversed next. For example, if the current node is L1 and there are 3 hyperedges, then the nodes connected by these 3 hyperedges together form the node set. That is, the current node may move to any node in the node set. The nodes in the node set are called candidate nodes.

[0046] Once the node set is determined, the probability of the current node transferring to each candidate node in the node set is calculated. In one embodiment, weights are assigned to hyperedges of the same type and hyperedges of different types, optionally with the weight of hyperedges of the same type being greater than that of hyperedges of different types. The type of hyperedge shared by the current node and each candidate node is calculated. If the hyperedge type is the same type, the intermediate probability of the same type hyperedge is obtained by multiplying the weight of the same type hyperedge by (1 - bias factor). If the hyperedge type is a different type hyperedge, the intermediate probability of the different type hyperedge is obtained by multiplying the weight of the different type hyperedge by the bias factor. The average or maximum value of all intermediate probabilities of the hyperedges shared by the current node and candidate nodes is taken as the probability of the current node transferring to the chosen node. After calculating the transfer probabilities of all nodes in the node set, the transfer probabilities are normalized to obtain the final transfer probabilities. Then, the node to which the current node transfers to the node set is determined based on the final transfer probabilities. In this embodiment, the bias factor is maximized when the number of steps reached half of the total number of steps, the intermediate probability of the heterogeneous hyperedge is maximized, and it contributes the most to the final transition probability. That is, when the number of steps reached half of the total number of steps, if the current node and the candidate node include more heterogeneous hyperedges, the probability of transitioning to that candidate node increases. Thus, in the early stage of the walk, it is more inclined to move to nodes within the same type of hyperedge; when the walk reaches the middle stage, it is more inclined to move to nodes within the heterogeneous hyperedge; and when the walk reaches the end stage, it is more inclined to move to nodes within the same type of hyperedge. This promotes the walk between lncRNA-disease-lncRNA or disease-lncRNA-disease, and better discovers the potential relationship between lncRNA and disease.

[0047] In another optional embodiment, determining the probability of the current node moving to each node in the node set based on the type of the hyperedge connecting the current node and the nodes in the set, as well as the bias factor, specifically involves:

[0048] Calculate the adjustment parameters of the same type of hyperedge and the adjustment parameters of the different types of hyperedge based on the bias factor;

[0049] For any node in the node set, calculate the set of hyperedges shared by the node and the current node; traverse each hyperedge in the hyperedge set, if the hyperedge is of the same type, calculate the internal weight of the hyperedge of the same type, and obtain the hyperedge weight based on the internal weight of the hyperedge and the adjustment parameter of the hyperedge of the same type; if the hyperedge is a different type of hyperedge, calculate the external weight of the different type of hyperedge, and obtain the hyperedge weight based on the external weight of the hyperedge and the adjustment parameter of the different type of hyperedge; take the average of the sum of the hyperedge weights corresponding to each hyperedge in the hyperedge set as the intermediate probability of the current node being transferred to a node in the node set;

[0050] The intermediate probabilities of all nodes in the node set are normalized, and the normalized intermediate probabilities are used as the probability of the current node being transferred to the node set.

[0051] The adjustment parameters for similar and dissimilar hyperedges are calculated based on the bias factor. These parameters dynamically adjust the importance of different types of hyperedges. Preferably, the adjustment parameter for similar hyperedges = 1 - bias factor; the adjustment parameter for dissimilar hyperedges = bias factor. However, the calculation methods for these parameters are not limited to the one described above. It is sufficient that the adjustment parameter for similar and dissimilar hyperedges is greater than the adjustment parameter for dissimilar hyperedges in the early and late stages of the walk, and less in the middle stages. If the bias factor is large, such as in the middle stages of the walk, the focus is more on exploring relationships between different types of nodes, thus the adjustment parameter for dissimilar hyperedges will be relatively high. Conversely, if the bias factor is small, such as in the early or late stages of the walk, the focus is more on exploring within nodes of the same type, thus the adjustment parameter for similar hyperedges will be relatively high.

[0052] Find the set of hyperedges that the current node and every node in the node set belong to. For example, if the hyperedges that the third node in the node set and the current node belong to are hyperedges 1 and 4, then the hyperedge set of the third node is {hyperedge 1, hyperedge 4}. If a hyperedge connects nodes of the same type, calculate the internal weight of the hyperedge. The internal weight is calculated based on the similarity between the nodes within the hyperedge. Preferably, calculate the average similarity of the nodes within the same type of hyperedge and use this average as the internal weight of the same type of hyperedge. Then, multiply this internal weight by the previously calculated adjustment parameter for the same type of hyperedge to obtain the final weight of the hyperedge. If a hyperedge connects nodes of different types, calculate the external weight of the hyperedge. This external weight is calculated based on the known lncRNA-disease association strength. Preferably, obtain the strongest association strength between all disease nodes and all lncRNA nodes in the heterogeneous hyperedge and use this strongest association strength as the external weight of the heterogeneous hyperedge. Then, multiply this external weight by the previously calculated adjustment parameter for the heterogeneous hyperedge to obtain the final weight of the hyperedge.

[0053] For a candidate node in the node set, add up the weights of all hyperedges connecting the current node and this candidate node, and then divide by the number of these hyperedges to obtain an average value, which is used as the intermediate probability. Normalize the intermediate probabilities of all candidate nodes in the node set so that their sum is 1, and then obtain the probability of the current node transferring to each node in the node set. Make a random selection based on these probabilities to determine the node for the next walk. The greater the probability, the higher the likelihood of being selected. Record the node sequence of each walk. For example, if the current node is the initial node L1 and the next selected node is L2, then the walk sequence is L1 - L2. Repeat the above process until the preset total number of walk steps is reached. For example, if the total number of walk steps is 5, then a possible walk sequence is L1 - L2 - D3 - D8 - L1, as Figure 4 shown.

[0054] S103. After completing multiple walks, count the access frequencies of lncRNA nodes and disease nodes, and obtain the mining results based on the access frequencies.

[0055] Each walk will select an initial node and obtain a walk sequence according to the total number of walk steps. When this process is repeated multiple times, multiple walk sequences will be obtained. Count the access frequencies of each lncRNA node and each disease node, as well as the access frequencies from each disease node to each lncRNA node. The access frequency represents the degree of association between lncRNA and disease. If the access frequency from a certain lncRNA node to a certain disease node is very high, it means that during the random walk process, the walk paths starting from this lncRNA are more likely to reach this disease, indicating that there may be a strong association between them.

[0056] In one embodiment, among all the walk sequences, count the total number of times a certain lncRNA and a certain disease are adjacent before and after, and use the ratio of the total number of times to the number of walk sequences as the access frequency. Sort <lncRNA, disease> according to the access frequency and perform visual display according to the sorting result. In a preferred embodiment, the known <lncRNA, disease> is marked in the visual display. For example, a possible sorting result is <lncRNA1, disease2>, <lncRNA5, disease1>, <lncRNA4, disease1>. If <lncRNA5, diseasex1> is known, then <lncRNA5, disease1> is displayed in bold.

[0057] This application also provides a system, software, or program for mining the relationship between lncRNA and disease, which includes:

[0058] The hypergraph construction module is used to construct lncRNA hypergraphs and disease hypergraphs based on lncRNA similarity and disease similarity, respectively. It connects the lncRNA hypergraph and disease hypergraph based on the known association between diseases and lncRNAs, and constructs a heterogeneous hypergraph containing hyperedges of the same type and hyperedges of different types.

[0059] The walking module is used to randomly select a starting node from the heterogeneous hypergraph. When walking at the current node, it determines the bias factor based on the current walking step count and the total walking step count, and determines the set of nodes that belong to the same hyperedge as the current node. Based on the type of the hyperedge connected to the current node and the nodes in the set, as well as the bias factor, it determines the probability that the current node will move to each node in the node set. According to the probability, it determines the next node to be moved to by the current node, records the walking order, and takes the next node to be moved to as the current node, until the total walking step count is reached.

[0060] The mining module is used to count the access frequency of lncRNA nodes and disease nodes after multiple walks, and obtain mining results based on the access frequency.

[0061] Optionally, determining the bias factor based on the current number of steps and the total number of steps specifically involves:

[0062] The bias factor is calculated using a Gaussian function, where the center value of the Gaussian function curve is half of the total number of walks.

[0063] Optionally, determining the probability of the current node transitioning to each node in the node set based on the type of the hyperedges connecting the current node and the nodes in the set, as well as the bias factor, specifically involves:

[0064] Calculate the adjustment parameters of the same type of hyperedge and the adjustment parameters of the different types of hyperedge based on the bias factor;

[0065] For any node in the node set, calculate the set of hyperedges shared by the node and the current node; traverse each hyperedge in the hyperedge set, if the hyperedge is of the same type, calculate the internal weight of the hyperedge of the same type, and obtain the hyperedge weight based on the internal weight of the hyperedge and the adjustment parameter of the hyperedge of the same type; if the hyperedge is a different type of hyperedge, calculate the external weight of the different type of hyperedge, and obtain the hyperedge weight based on the external weight of the hyperedge and the adjustment parameter of the different type of hyperedge; take the average of the sum of the hyperedge weights corresponding to each hyperedge in the hyperedge set as the intermediate probability of the current node being transferred to a node in the node set;

[0066] The intermediate probabilities of all nodes in the node set are normalized, and the normalized intermediate probabilities are used as the probability of the current node being transferred to the node set.

[0067] Optionally, the internal weights of the same type of hyperedges are calculated as follows:

[0068] Calculate the average similarity of nodes inside hyperedges of the same type, and use the average value as the weight of the hyperedges of the same type.

[0069] Optionally, the external weight of the heterogeneity hyperedge is specifically:

[0070] Obtain the strongest association strength between all disease nodes and all lncRNA nodes in the heterogeneity hyperedge, and use the strongest association strength as the external weight of the heterogeneity hyperedge.

[0071] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0072] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0073] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0074] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

Claims

1. A method for mining the relationship between lncRNA and disease, characterized in that, The method includes: Based on lncRNA similarity and disease similarity, lncRNA hypergraphs and disease hypergraphs are constructed respectively. Based on the known association between diseases and lncRNAs, the lncRNA hypergraphs and disease hypergraphs are connected, and a heterogeneous hypergraph containing hyperedges of the same type and hyperedges of different types is constructed. A starting node is randomly selected from the heterogeneous hypergraph. When the current node is walking, the bias factor is determined based on the current number of steps and the total number of steps. The set of nodes that belong to the same hyperedge as the current node is determined. The probability of the current node moving to each node in the node set is determined based on the type of the hyperedge connected to the current node and the nodes in the set and the bias factor. The next node to be moved to by the current node is determined according to the probability. The walking order is recorded. The next node to be moved to is taken as the current node. This process continues until the total number of steps is reached. After completing multiple walks, the access frequency of lncRNA nodes and disease nodes is counted, and the mining results are obtained based on the access frequency. The process of determining the bias factor based on the current number of steps and the total number of steps is as follows: The bias factor is calculated using a Gaussian function, where the center value of the Gaussian function curve is half of the total number of steps. The process of determining the probability of the current node moving to each node in the node set based on the type of the hyperedges connecting the current node and the nodes in the set, as well as the bias factor, is as follows: Assign weights to hyperedges of the same type and hyperedges of different types, with the weight of hyperedges of the same type being greater than the weight of hyperedges of different types; Calculate the type of superedge shared by the current node and each candidate node. If the superedge type is the same type, obtain the intermediate probability of the same type superedge by (1 - bias factor) * weight of the same type superedge. If the superedge type is a different type superedge, obtain the intermediate probability of the different type superedge by bias factor * weight of the different type superedge. The average or maximum value of all intermediate probabilities of the current node and the candidate node sharing a common superedge is taken as the probability of the current node transitioning to the candidate node. After calculating the transition probabilities of all nodes in the node set, the transition probabilities are normalized to obtain the final transition probabilities.

2. The method as described in claim 1, characterized in that, The internal weights of the same type of hyperedges are calculated as follows: Calculate the average similarity of nodes inside hyperedges of the same type, and use the average value as the weight of the hyperedges of the same type.

3. The method as described in claim 1, characterized in that, The external weights of the heterogeneous hyperedges are specifically as follows: Obtain the strongest association strength between all disease nodes and all lncRNA nodes in the heterogeneity hyperedge, and use the strongest association strength as the external weight of the heterogeneity hyperedge.

4. A system for mining the relationship between lncRNA and disease, characterized in that, The system includes: The hypergraph construction module is used to construct lncRNA hypergraphs and disease hypergraphs based on lncRNA similarity and disease similarity, respectively. It connects the lncRNA hypergraph and disease hypergraph based on the known association between diseases and lncRNAs, and constructs a heterogeneous hypergraph containing hyperedges of the same type and hyperedges of different types. The walking module is used to randomly select a starting node from the heterogeneous hypergraph. When walking at the current node, it determines the bias factor based on the current walking step count and the total walking step count, and determines the set of nodes that belong to the same hyperedge as the current node. Based on the type of the hyperedge connected to the current node and the nodes in the set, as well as the bias factor, it determines the probability that the current node will move to each node in the node set. According to the probability, it determines the next node to be moved to by the current node, records the walking order, and takes the next node to be moved to as the current node, until the total walking step count is reached. The mining module is used to count the access frequency of lncRNA nodes and disease nodes after multiple walks, and obtain the mining results based on the access frequency. The process of determining the bias factor based on the current number of steps and the total number of steps is as follows: The bias factor is calculated using a Gaussian function, where the center value of the Gaussian function curve is half of the total number of steps. The process of determining the probability of the current node moving to each node in the node set based on the type of the hyperedges connecting the current node and the nodes in the set, as well as the bias factor, is as follows: Assign weights to hyperedges of the same type and hyperedges of different types, with the weight of hyperedges of the same type being greater than the weight of hyperedges of different types; Calculate the type of superedge shared by the current node and each candidate node. If the superedge type is the same type, obtain the intermediate probability of the same type superedge by (1 - bias factor) * weight of the same type superedge. If the superedge type is a different type superedge, obtain the intermediate probability of the different type superedge by bias factor * weight of the different type superedge. The average or maximum value of all intermediate probabilities of the current node and the candidate node sharing a common superedge is taken as the probability of the current node transitioning to the candidate node. After calculating the transition probabilities of all nodes in the node set, the transition probabilities are normalized to obtain the final transition probabilities.

5. The system as described in claim 4, characterized in that, The internal weights of the same type of hyperedges are calculated as follows: Calculate the average similarity of nodes inside hyperedges of the same type, and use the average value as the weight of the hyperedges of the same type.

6. The system as described in claim 4, characterized in that, The external weights of the heterogeneous hyperedges are specifically as follows: Obtain the strongest association strength between all disease nodes and all lncRNA nodes in the heterogeneity hyperedge, and use the strongest association strength as the external weight of the heterogeneity hyperedge.

Images