Lung cancer driver gene prediction method, device, equipment, medium and product
By constructing the feature matrix and combining gene expression patterns and mutation information, using deep learning models to predict lung cancer driver genes, the problem of insufficient prediction accuracy in the existing technology is solved, and higher accuracy and performance indicators are achieved.
Patent Information
- Application Number
- CN202410929468.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-11
- Publication Date
- 2024-11-12
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
When predicting lung cancer driver genes, the prior art simply merges the characteristics extracted from different data sources, resulting in the failure of the effectiveness of lung cancer driver genes to be improved.
By calculating the similarity between each gene in multiple target lung cancer tumor samples, constructing a feature matrix, combining the similarity of gene expression patterns and mutation information, predictions were made using deep learning models.
It significantly improves the accuracy of lung cancer driver gene prediction, improves AUC, accuracy and recall, and is superior to traditional methods such as random forests and support vector machines.
Smart Images

Figure CN118942550A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of machine learning technology, and in particular to a method, device, equipment, medium and product for predicting lung cancer driver genes. Background Art
[0002] Identifying lung cancer driver genes from millions of somatic mutations is an extremely challenging task. In order to accurately predict the driver genes of lung cancer, researchers have proposed many computational methods, most of which use machine learning techniques such as random forests and support vector machines. However, the effectiveness of identifying lung cancer driver genes by simply merging features extracted from different data sources and using the selected features to train machine learning models still needs to be improved. Summary of the invention
[0003] The purpose of this application is to provide a method, device, equipment, medium and product for predicting lung cancer driver genes, which can improve the accuracy of lung cancer driver gene prediction.
[0004] To achieve the above objectives, this application provides the following solutions:
[0005] In a first aspect, the present application provides a method for predicting lung cancer driver genes, comprising:
[0006] Calculate the similarity between genes in multiple target lung cancer tumor samples to obtain k similar genes for each gene; k is an integer greater than 1;
[0007] The ratio of different mutation types of each gene is obtained from the lung cancer dataset, and the ratio of different mutation types of each gene is used as the feature vector of each gene;
[0008] A feature submatrix is constructed according to k similar genes of each gene and the feature vector of each gene, and each constructed feature submatrix constitutes a feature matrix;
[0009] The feature matrix is input into a lung cancer driver gene prediction model, and the lung cancer driver gene is output; the lung cancer driver gene prediction model is obtained by training a deep learning model.
[0010] Optionally, the characteristic matrix of the ith gene includes 2k vectors, the odd-numbered columns in the 2k vectors are all characteristic vectors of the ith gene, and the even-numbered columns in the 2k vectors are characteristic vectors of k similar genes in sequence.
[0011] Optionally, the similarity between genes in multiple target lung cancer tumor samples is calculated to obtain k similar genes for each gene, specifically including:
[0012] Based on multiple target lung cancer tumor samples, the Pearson correlation coefficient was used to calculate the similarity between genes and obtain k similar genes for each gene.
[0013] Optionally, the ratio of different mutation types of each gene is obtained from the lung cancer dataset, and the ratio of different mutation types of each gene is used as the feature vector of each gene, specifically including:
[0014] The ratio of different mutation types of each gene whose mutation times are greater than the set times is obtained from the lung cancer dataset, and the obtained ratio of different mutation types of each gene is used as the feature vector of each gene.
[0015] Optionally, the mutation types of the gene include silent mutations, nonsense mutations, splice site mutations and missense mutations; and the lung cancer dataset includes a LUAD dataset and a LUSC dataset.
[0016] Optionally, the deep learning model is a bidirectional long short-term memory network.
[0017] In a second aspect, the present application provides a lung cancer driver gene prediction device, comprising:
[0018] A gene similarity determination module is used to calculate the similarity between genes in multiple target lung cancer tumor samples to obtain k similar genes for each gene; k is an integer greater than 1;
[0019] A gene feature vector determination module is used to obtain the ratio of different mutation types of each gene from the lung cancer data set, and use the ratio of different mutation types of each gene as the feature vector of each gene;
[0020] A feature matrix construction module is used to construct a feature sub-matrix according to k similar genes of each gene and a feature vector of each gene, and each constructed feature sub-matrix constitutes a feature matrix;
[0021] A prediction module is used to input the feature matrix into a lung cancer driver gene prediction model and output the lung cancer driver gene; the lung cancer driver gene prediction model is obtained by training a deep learning model.
[0022] In a third aspect, the present application provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement any of the above-described lung cancer driver gene prediction methods.
[0023] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described lung cancer driver gene prediction methods.
[0024] In a fifth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements any of the above-mentioned lung cancer driver gene prediction methods.
[0025] According to the specific embodiments provided in this application, this application discloses the following technical effects:
[0026] The present application provides a method, apparatus, device, medium and product for predicting lung cancer driver genes. A feature submatrix is constructed based on k similar genes of each gene and the feature vector of each gene. Each constructed feature submatrix constitutes a feature matrix. Lung cancer driver genes are predicted based on the feature matrix of each gene, effectively combining the two types of information, namely the similarity of gene expression patterns and gene mutations, to improve the prediction accuracy. BRIEF DESCRIPTION OF THE DRAWINGS
[0027] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0028] Figure 1 A schematic diagram of a process for predicting lung cancer driver genes provided in one embodiment of the present application;
[0029] Figure 2 A schematic diagram of a feature matrix provided in an embodiment of the present application;
[0030] Figure 3 A schematic diagram of a bidirectional long short-term memory network structure provided in an embodiment of the present application;
[0031] Figure 4 A schematic diagram of a ROC curve based on a LUAD data set provided in an embodiment of the present application;
[0032] Figure 5 A schematic diagram of a ROC curve based on a LUSC data set provided in an embodiment of the present application;
[0033] Figure 6 A schematic diagram of the structure of a lung cancer driver gene prediction device provided in one embodiment of the present application;
[0034] Figure 7 A schematic diagram of the structure of a computer device provided in one embodiment of the present application. DETAILED DESCRIPTION
[0035] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0036] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.
[0037] The present application embodiment provides a method for predicting lung cancer driver genes, such as Figure 1 As shown, a lung cancer driver gene prediction method includes:
[0038] Step 101: Calculate the similarity between genes in multiple target lung cancer tumor samples to obtain k similar genes for each gene; k is an integer greater than 1.
[0039] Step 102: Obtain the ratio of different mutation types of each gene from the lung cancer data set, and use the ratio of different mutation types of each gene as the feature vector of each gene.
[0040] Step 103: construct a feature sub-matrix according to the k similar genes of each gene and the feature vector of each gene, and each constructed feature sub-matrix constitutes a feature matrix.
[0041] Step 104: input the feature matrix into a lung cancer driver gene prediction model, and output the lung cancer driver gene; the lung cancer driver gene prediction model is obtained by training a deep learning model.
[0042] The present application constructs a feature submatrix based on k similar genes of each gene and the feature vector of each gene. Each constructed feature submatrix constitutes a feature matrix. Lung cancer driver genes are predicted based on the feature matrix of each gene, effectively combining the two types of information, the similarity of gene expression patterns and gene mutations, to improve the prediction accuracy.
[0043] In order to characterize the relationship between genes in disease states, the Pearson Correlation Coefficient (PCC) is used to calculate the similarity. Step 101 specifically includes:
[0044] Based on multiple (m) target lung cancer tumor samples, the Pearson correlation coefficient is used to calculate the similarity between genes, and k similar genes of each gene are obtained. More specifically, the k genes closest to the current gene are arranged according to the value of the Pearson correlation coefficient. Each gene is connected to its own k genes with the largest Pearson correlation coefficient to construct a gene similarity network.i =(e i1 , e i2 , ..., e im ) represents the i-th gene g in m target lung cancer tumor samples i The expression value, e iq represents the expression value of the i-th gene in the q-th target lung cancer tumor sample, and the value range of q is 1≤q≤m.
[0045]
[0046] Among them, PCC (g i ,g j ) is the similarity between the ith gene and the jth gene, for e i The average expression value of the ith gene in m target tumor samples, for e j The average expression value of the jth gene in m target tumor samples.
[0047] The lung cancer dataset includes a LUAD dataset and a LUSC dataset. The mutation annotation format (MAF) file is a common file format for describing genomic variation information. For each gene in the lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC) datasets, significant mutation genes are found from the MAF file for driver gene prediction research, and the mutation rates of different mutation types of these genes are counted.
[0048] The types of gene mutations include silent mutations, nonsense mutations, splice site mutations and missense mutations.
[0049] In order to improve the effectiveness of mutation effects on gene drive prediction, the names of all mutated genes are counted in the MAF file, and then genes with mutation times greater than 5 are considered to be significantly mutated genes, which are considered to be statistically significant and can be used for research on driver gene prediction. This application obtains the ratio of different mutation types of each gene from a lung cancer dataset, and uses the ratio of different mutation types of each gene as a feature vector of each gene, specifically including: obtaining the ratio of different mutation types of each gene with mutation times greater than a set number from a lung cancer dataset, and using the obtained ratio of different mutation types of each gene as a feature vector of each gene.
[0050] The characteristic matrix of the i-th gene includes 2k vectors. The odd columns in the 2k vectors are the characteristic vectors of the i-th gene, and the even columns in the 2k vectors are the characteristic vectors of k similar genes in turn. The characteristic matrix is as follows: Figure 2 As shown, represents the feature matrix, x fis the feature vector of each gene in the feature matrix.
[0051] Among them, step 103 matches the gene names obtained in step 101 and step 102 to obtain each gene g i The eigenvector x fi , and g s1 , g s2 , ..., g sk Yes i The k nearest neighbor genes of PCC(g i , g s1 )>PCC(g i , g s2 )>…>PCC(g i , g sk ). By arranging the 2k vectors into 2k×x f Matrix to construct the feature matrix Feature Matrix like Figure 2 As shown in the construction,
[0052] The present application specifically trains the deep learning model including: using a training set to train the deep learning model to obtain a lung cancer driver gene prediction model; the training set includes input data and label data, the input data includes feature matrix samples, and the label data is the lung cancer driver gene corresponding to the feature matrix sample.
[0053] The deep learning model is a bidirectional long short-term memory network (Bidirectional Long Short-Term Memory, Bi-LSTM), a bidirectional long short-term memory network such as Figure 3 As shown, x1, x2, x3 and x4 are all inputs of Bi-LSTM, h1, h2, h3, h4, h1^, h2^, h3^ and h4^ are the outputs of each LSTM unit. This application inputs the constructed feature matrix into Bi-LSTM to perform the binary classification task of genes. Bi-LSTM is a deep learning architecture for sequence modeling. Bi-LSTM is composed of multiple LSTM units. These LSTM units can simultaneously consider the past and future information of the sequence to better understand the temporal relationship in the sequence. When processing one-dimensional sequence data, Bi-LSTM captures temporal dependencies by gradually traversing the sequence and updating the internal state. One of its core advantages is the ability to simultaneously learn the past and future information of the sequence through forward and backward propagation, thereby better understanding the entire sequence. As Figure 3As shown in the figure, a single-layer Bi-LSTM is composed of two groups of LSTMs, one of which processes the input sequence forward and the other processes the sequence backward. After the processing is completed, the outputs of the two LSTMs are concatenated. Only after all time steps are calculated can the final Bi-LSTM output result be obtained.
[0054] In the present application, m (an integer greater than 1) target lung cancer tumor samples are lung cancer tumor samples from m patients, and the purpose of the present application is to predict the lung cancer driver genes of these m patients.
[0055] This application predicts driver genes by integrating similarity networks and features that characterize mutation frequencies, and effectively combines the two types of information by using the similarity of gene expression patterns and the functional impact of mutations, thereby improving the accuracy of predictions.
[0056] This application uses the area under the curve (AUC) score, recall rate (Recall) and precision (Precision) as evaluation indicators of the model. Figure 4 and Figure 5 The ROC curves of BiLSTM-Driver, Transformer, Random Forest (Rf) and Support Vector Machine (SVM) on LUSC and LUAD datasets are given. BiLSTM-Driver refers to the lung cancer driver gene prediction model of this application.
[0057] Table 3 Comparison based on LUAD dataset
[0058] AUC Precision Recall BiLSTM-Driver 0.994 0.991 0.987 Transformer 0.981 0.986 0.807 RandomForest 0.965 0.985 0.554 SVM 0.819 0.805 0.521
[0059] Table 4 Comparison based on LUSC dataset
[0060] AUC Precision Recall BiLSTM-Driver 0.987 0.993 0.991 Transformer 0.983 0.991 0.725 RandomForest 0.965 0.978 0.636 SVM 0.696 0.803 0.374
[0061] The experiments on the LUAD and LUSC datasets are recorded in Tables 1 and 2. BiLSTM-Driver achieved significant improvements in the three key performance indicators of AUC, precision, and recall. For the LUAD dataset, BiLSTM-Driver improved AUC by 2.9% and 17.5% compared to Random Forest and SVM, respectively, and improved precision by almost 18.6% (compared to SVM), while recall improved by 18% compared to the Transformer model. This trend was further confirmed in the evaluation of the LUSC dataset, where BiLSTM-Driver improved AUC by 29.1% compared to SVM, achieved a 19.0% improvement in precision, and improved recall by 26.6% compared to Transformer. These results not only highlight the effectiveness of BiLSTM-Driver in predicting lung cancer driver genes, but also highlight its significant advantage over existing methods in capturing true positive samples, indicating the application potential of BiLSTM-Driver in the field of bioinformatics.
[0062] Based on the same inventive concept, the embodiment of the present application also provides a lung cancer driver gene prediction device for implementing the lung cancer driver gene prediction method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more lung cancer driver gene prediction device embodiments provided below can refer to the limitations of the lung cancer driver gene prediction method above, and will not be repeated here.
[0063] In an exemplary embodiment, Figure 6 As shown, a lung cancer driver gene prediction device is provided, comprising:
[0064] The gene similarity determination module 201 is used to calculate the similarity between genes in multiple target lung cancer tumor samples to obtain k similar genes for each gene; k is an integer greater than 1.
[0065] The gene feature vector determination module 202 is used to obtain the ratio of different mutation types of each gene from the lung cancer data set, and use the ratio of different mutation types of each gene as the feature vector of each gene.
[0066] The feature matrix construction module 203 is used to construct a feature sub-matrix according to the k similar genes of each gene and the feature vector of each gene, and each constructed feature sub-matrix constitutes a feature matrix.
[0067] The prediction module 204 is used to input the feature matrix into a lung cancer driver gene prediction model and output the lung cancer driver gene; the lung cancer driver gene prediction model is obtained by training a deep learning model.
[0068] In an exemplary embodiment, a computer device is provided. The computer device may be a server or a terminal. The internal structure diagram thereof may be as follows: Figure 7 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store lung cancer driver gene prediction data. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a lung cancer driver gene prediction method is implemented.
[0069] Those skilled in the art will understand that Figure 7 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.
[0070] In an exemplary embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.
[0071] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.
[0072] In an exemplary embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.
[0073] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.
[0074] Those of ordinary skill 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 the present 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), magnetic 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. By way of illustration and not limitation, RAM may be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).
[0075] The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. The non-relational database may include a distributed database based on blockchain, etc., but is not limited thereto. 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, etc., but is not limited thereto.
[0076] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0077] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.
Claims
1. A method for predicting lung cancer driver genes, characterized in that: The lung cancer driver gene prediction method comprises: Calculate the similarity between genes in multiple target lung cancer tumor samples to obtain k similar genes for each gene; k is an integer greater than 1; The ratio of different mutation types of each gene is obtained from the lung cancer dataset, and the ratio of different mutation types of each gene is used as the feature vector of each gene; A feature submatrix is constructed according to k similar genes of each gene and the feature vector of each gene, and each constructed feature submatrix constitutes a feature matrix; The feature matrix is input into a lung cancer driver gene prediction model, and the lung cancer driver gene is output; the lung cancer driver gene prediction model is obtained by training a deep learning model.
2. The lung cancer driver gene prediction method according to claim 1, characterized in that: The characteristic matrix of the i-th gene includes 2k vectors, the odd columns in the 2k vectors are the characteristic vectors of the i-th gene, and the even columns in the 2k vectors are the characteristic vectors of k similar genes in turn.
3. The method for predicting lung cancer driver genes according to claim 1, characterized in that: Calculate the similarity between genes in multiple target lung cancer tumor samples and obtain k similar genes for each gene, including: Based on multiple target lung cancer tumor samples, the Pearson correlation coefficient was used to calculate the similarity between genes and obtain k similar genes for each gene.
4. The method for predicting lung cancer driver genes according to claim 1, characterized in that: The ratio of different mutation types of each gene is obtained from the lung cancer dataset, and the ratio of different mutation types of each gene is used as the feature vector of each gene, including: The ratio of different mutation types of each gene whose mutation times are greater than the set times is obtained from the lung cancer dataset, and the obtained ratio of different mutation types of each gene is used as the feature vector of each gene.
5. The method for predicting lung cancer driver genes according to claim 1, characterized in that: The mutation types of the gene include silent mutation, nonsense mutation, splice site mutation and missense mutation; the lung cancer dataset includes a LUAD dataset and a LUSC dataset.
6. The method for predicting lung cancer driver genes according to claim 1, characterized in that: The deep learning model is a bidirectional long short-term memory network.
7. A lung cancer driver gene prediction device, characterized in that: The lung cancer driver gene prediction device comprises: A gene similarity determination module is used to calculate the similarity between genes in multiple target lung cancer tumor samples to obtain k similar genes for each gene; k is an integer greater than 1; A gene feature vector determination module is used to obtain the ratio of different mutation types of each gene from the lung cancer data set, and use the ratio of different mutation types of each gene as the feature vector of each gene; A feature matrix construction module is used to construct a feature sub-matrix according to k similar genes of each gene and a feature vector of each gene, and each constructed feature sub-matrix constitutes a feature matrix; A prediction module is used to input the feature matrix into a lung cancer driver gene prediction model and output the lung cancer driver gene; the lung cancer driver gene prediction model is obtained by training a deep learning model.
8. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the lung cancer driver gene prediction method according to any one of claims 1 to 6.
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 method for predicting lung cancer driver genes according to any one of claims 1 to 6 is implemented.
10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for predicting lung cancer driver genes according to any one of claims 1 to 6 is implemented.
