Method and system for predicting mutation position of lung cancer gene

Through the intelligent processing of multi-dimensional clinical detection data, the target positioning network and optimization network are used to solve the problem of insufficient sensitivity and specificity of gene screening in the existing technology, the accurate detection and personalized treatment of lung cancer gene mutations are achieved, and the The effect of targeted therapy.

CN118969081AInactive Publication Date: 2024-11-15TIANJIN FUXUN TECH DEV CO LTD
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202411029295.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-30
Publication Date
2024-11-15
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

Existing gene screening technologies lack sensitivity and specificity in detecting complex gene mutations and low-frequency mutations, resulting in missed detection or misdiagnosis, affecting the effect of targeted treatment and resource utilization.

Method used

Multi-dimensional clinical detection data is used to intelligently process the target positioning network and optimization network to generate mutation prediction analysis information, and improve the accuracy and reliability of gene mutation detection.

Benefits of technology

It has achieved personalized localization and accurate detection of gene mutations related to lung cancer, improved the effectiveness of targeted treatment, reduced trial and error during the treatment process, and improved the treatment effect and survival rate of patients.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118969081A_ABST
    Figure CN118969081A_ABST
Patent Text Reader

Abstract

The invention relates to the field of data processing, in particular to a lung cancer gene mutation position prediction method and system. The method comprises the following steps: acquiring multi-dimensional clinical detection data of a target object; inputting the multi-dimensional clinical detection data into a target positioning network, and performing mutation positioning prediction processing on the target object through the target positioning network to obtain mutation positioning features of the target object; inputting the abrupt change positioning features and basic data of the target object into an optimization network, and performing self-positioning optimization on the abrupt change positioning features to obtain target abrupt change positioning features of the target object; mutation prediction analysis information of the target object is generated based on the target mutation positioning features, and is used for assisting lung cancer mutation detection optimization of the target object. According to the method, intelligent mutation position prediction of the lung cancer gene can be realized, and related personnel can conveniently refer to mutation prediction analysis to realize gene mutation detection of a target object, so that the accuracy and reliability of gene screening are improved in an auxiliary manner, and the targeted treatment effect is improved in an auxiliary manner.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of data processing, and in particular to a method and system for predicting the mutation position of a lung cancer gene. Background Art

[0002] Lung cancer is one of the most common cancers in the world with the highest morbidity and mortality. For example, the incidence of non-small cell lung cancer (NSCLC) is 85%-90%. With the development of precision medicine and genetic testing technology, more and more genes have been found to be associated with NSCLC.

[0003] Since there are many gene mutations in NSCLC that are sensitive to targeted therapy, genetic screening can help doctors determine whether patients have targets for these targeted therapies, thereby selecting the most appropriate treatment plan to improve treatment efficacy and patient survival. In addition, some patients may receive treatments that are insensitive or ineffective to them, which not only wastes medical resources but also delays the timing of effective treatment. With the help of genetic screening technology, the adaptability of such patients to related treatments can be improved to avoid this situation and further improve treatment efficiency.

[0004] Existing gene screening technologies may have limitations in the sensitivity and specificity of detecting specific gene variants, such as detecting complex gene variants (such as compound mutations or low-frequency variants). For example, in the detection of low-frequency variants, the sensitivity is limited by the specific primer design and reaction conditions, which leads to missed detection in PCR, resulting in missed detection or misdiagnosis of some potential targeted therapeutic targets. Therefore, it is urgent to design a new technical solution to improve the accuracy of gene screening. Summary of the invention

[0005] In view of the technical problems existing in the prior art, the present invention provides a method and system for predicting the mutation position of lung cancer genes, so as to realize intelligent prediction of the mutation position of lung cancer genes, so as to facilitate relevant personnel to refer to mutation prediction analysis to realize gene mutation detection of target objects, thereby assisting in improving the accuracy and reliability of gene screening and assisting in improving the effect of targeted treatment.

[0006] In a first aspect, an embodiment of the present application provides a method for predicting the mutation position of a lung cancer gene, the method comprising:

[0007] Acquire multidimensional clinical test data of the target object; the multidimensional clinical test data at least includes: gene sequence data and clinical test index data of the target object;

[0008] Inputting the multidimensional clinical test data into a target localization network, performing mutation localization prediction processing on a target object through the target localization network, and obtaining mutation localization features of the target object;

[0009] The mutation location feature and basic data of the target object are input into the optimization network, and the mutation location feature is self-optimized to obtain the target mutation location feature of the target object; the basic data at least includes: case data and basic physical attribute information of the target object;

[0010] Based on the target mutation location features, mutation prediction analysis information of the target object is generated to assist in optimizing lung cancer mutation detection of the target object.

[0011] In a second aspect, an embodiment of the present application provides a system for predicting the mutation position of a lung cancer gene, the system comprising the following units, wherein:

[0012] An acquisition unit is configured to acquire multi-dimensional clinical test data of a target object; the multi-dimensional clinical test data at least includes: gene sequence data and clinical test index data of the target object;

[0013] The detection unit is configured to input the multidimensional clinical detection data into a target positioning network, perform mutation positioning prediction processing on a target object through the target positioning network, and obtain a mutation positioning feature of the target object; input the mutation positioning feature and basic data of the target object into an optimization network, perform self-positioning optimization on the mutation positioning feature, and obtain a target mutation positioning feature of the target object; the basic data at least includes: case data and basic physical attribute information of the target object;

[0014] The analysis unit is configured to generate a defect analysis report of the target object based on the target defect feature to assist in optimizing the detection of the target object.

[0015] In a third aspect, an embodiment of the present application provides an electronic device, the electronic device comprising:

[0016] at least one processor, memory, and input-output unit;

[0017] The memory is used to store a computer program, and the processor is used to call the computer program stored in the memory to execute the mutation position prediction method of the lung cancer gene of the first aspect.

[0018] In a fourth aspect, a computer-readable storage medium is provided, which includes instructions. When the instructions are executed on a computer, the computer executes the method for predicting the mutation position of a lung cancer gene according to the first aspect.

[0019] The beneficial effect of the present invention is that a method and system for predicting the mutation position of lung cancer genes are provided. In this technical solution, first, multidimensional clinical test data of the target object is obtained. The multidimensional clinical test data at least includes: gene sequence data and clinical test index data of the target object. Then, the multidimensional clinical test data is input into the target positioning network, and the target object is subjected to mutation positioning prediction processing through the target positioning network to obtain the mutation positioning feature of the target object. Then, the mutation positioning feature and the basic data of the target object are input into the optimization network, and the mutation positioning feature is self-positioned and optimized to obtain the target mutation positioning feature of the target object. Among them, the basic data at least includes: case data and basic physical attribute information of the target object. Finally, mutation prediction analysis information of the target object is generated based on the target mutation positioning feature to assist in optimizing the lung cancer mutation detection of the target object. The technical solution of this application realizes the intelligent mutation position prediction of lung cancer genes, which is convenient for relevant personnel to refer to the mutation prediction analysis to realize the gene mutation detection of the target object, thereby assisting in improving the accuracy and reliability of gene screening and assisting in improving the effect of targeted therapy. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a schematic diagram of a process of predicting the mutation position of a lung cancer gene according to an embodiment of the present application;

[0021] Figure 2 It is a schematic diagram of the structure of a lung cancer gene mutation position prediction system according to an embodiment of the present application;

[0022] Figure 3 is a schematic diagram of the structure of an electronic device according to an embodiment of the present application;

[0023] Figure 4 It is a structural schematic diagram of a medium device according to an embodiment of the present application. DETAILED DESCRIPTION

[0024] 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 those skilled in the art without creative work are within the scope of protection of this application.

[0025] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, "plurality" means two or more, unless otherwise clearly and specifically defined.

[0026] In the description of the present application, the term "for example" is used to mean "used as an example, illustration or description". Any embodiment described as "for example" in the present application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any technician in the field to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes will not be elaborated in detail to avoid unnecessary details to obscure the description of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed in the present application.

[0027] Lung cancer is one of the most common cancers in the world in terms of morbidity and mortality. For example, the incidence of non-small cell lung cancer (NSCLC) is 85%-90%, mainly including adenocarcinoma (40%-50%) and squamous cell carcinoma (20%-30%). About 30% of NSCLC patients have genetic changes that are sensitive to targeted therapy. For example, epidermal growth factor receptor (EGFR) mutations, anaplastic lymphoma kinase (ALK) rearrangements, and c-ros sarcoma oncogenic factor-receptor tyrosine kinase fusions are as high as 45% in lung adenocarcinoma.

[0028] With the development of precision medicine and genetic testing technology, more and more genes have been found to be associated with NSCLC, among which the v-rafmurine sarcoma viral oncogene homolog B1 (BRAF) gene may be another important driver gene in NSCLC. Since there are many gene mutations in NSCLC that are sensitive to targeted therapy, such as EGFR mutations, ALK rearrangements, and ROS1 rearrangements, genetic screening can help doctors determine whether patients have these targeted therapy targets, so as to select the most appropriate treatment plan and improve the treatment effect and patient survival rate. In addition, some patients may receive treatments that are insensitive or ineffective to them, which not only wastes medical resources, but also delays the opportunity for effective treatment. With the help of genetic screening technology, the adaptability of such patients to related treatments can be improved to avoid this situation and further improve treatment efficiency.

[0029] However, existing gene screening technologies may have limitations in the sensitivity and specificity of detecting specific gene variants. For example, in detecting complex gene variants (such as compound mutations or low-frequency variants), especially in detecting low-frequency variants, the sensitivity is limited by specific primer design and reaction conditions, which may lead to missed detections in PCR, resulting in missed detections or misdiagnosis of some potential targeted therapeutic targets. In addition, gene screening technology requires complex laboratory equipment and professional technical support, which makes it costly and difficult to promote and apply.

[0030] The embodiment of the present application provides a method and system for predicting the mutation position of lung cancer genes. In this technical solution, first, multidimensional clinical test data of the target object is obtained. The multidimensional clinical test data includes at least: gene sequence data and clinical test index data of the target object. Exemplarily, the gene sequence data is, for example, a DNA sequence obtained by high-throughput sequencing technology, and contains genomic variation data (such as SNPs, InDels, etc.). Clinical test index data, for example, includes biochemical indicators (such as blood test results, tumor markers), imaging data (such as CT, MRI reports), clinical symptoms, etc. Then, the multidimensional clinical test data is input into the target positioning network, and the target object is subjected to mutation positioning prediction processing through the target positioning network to obtain the mutation positioning feature of the target object. Then, the mutation positioning feature and the basic data of the target object are input into the optimization network, and the mutation positioning feature is self-positioned and optimized to obtain the target mutation positioning feature of the target object. Among them, the basic data includes at least: case data and basic physical attribute information of the target object. Finally, the mutation prediction analysis information of the target object is generated based on the target mutation positioning feature to assist in optimizing the lung cancer mutation detection of the target object.

[0031] The technical solution of this application can predict the mutation position of the target object in a personalized way through multi-dimensional clinical test data, including gene sequences and clinical test indicators, combined with the target positioning network. This personalized prediction can more accurately locate gene mutations related to lung cancer, avoiding generalization and errors in traditional methods. Secondly, the self-positioning optimization of mutation positioning features using the optimization network helps to further improve the accuracy and reliability of the mutation position. This means that the unique gene change characteristics of each patient can be more accurately grasped in diagnosis and treatment decisions. Thirdly, through intelligent mutation position prediction, technology can help medical professionals perform gene mutation detection more quickly, thereby speeding up the early screening and diagnosis process of lung cancer. This helps to improve the accuracy and reliability of screening, while improving the effectiveness of targeted therapy, so that patients can get the most appropriate treatment plan earlier. In addition, the generated mutation prediction analysis information provides a valuable reference for the medical team and can support personalized treatment decisions. This personalized treatment method can minimize trial and error during treatment and improve the treatment effect and survival rate of patients.

[0032] The mutation position prediction scheme of lung cancer gene provided by the embodiment of the present application can also be performed by an electronic device, which can be a server, a server cluster, or a cloud server. The electronic device can also be a terminal device such as a mobile phone, a computer, a tablet computer, a wearable device, or a special device (such as a special terminal device with a mutation position prediction system of lung cancer gene). The chip introduced in the above embodiment can also be carried in these electronic devices. Alternatively, these electronic devices can also be installed with a service program for executing the mutation position prediction scheme of lung cancer gene.

[0033] Figure 1 A schematic diagram of a method for predicting the mutation position of a lung cancer gene provided in an embodiment of the present application is shown in FIG. Figure 1 As shown, the method comprises the following steps:

[0034] 101, obtaining multi-dimensional clinical test data of a target object;

[0035] 102, inputting the multidimensional clinical test data into a target positioning network, performing mutation positioning prediction processing on a target object through the target positioning network, and obtaining mutation positioning features of the target object;

[0036] 103, inputting the mutation location feature and basic data of the target object into an optimization network, performing self-location optimization on the mutation location feature, and obtaining a target mutation location feature of the target object;

[0037] 104. Generate mutation prediction analysis information of the target object based on the target mutation location feature to assist in optimizing lung cancer mutation detection of the target object.

[0038] In the embodiment of the present application, the multidimensional clinical test data at least includes: gene sequence data of the target object and clinical test index data. Specifically, multidimensional clinical test data refers to multifaceted data used to comprehensively evaluate the health status of the target object. In the context of lung cancer gene mutation prediction, these data include at least the following two categories: First, gene sequence data, including but not limited to DNA sequence data and genome variation information. Second, clinical test index data, including but not limited to blood biochemical indicators, imaging examination results, and pathological examination data.

[0039] In practical applications, DNA sequence data refers to the genomic DNA sequence information of the target object obtained through sequencing technology. These data can reveal the variations in the genome, including single nucleotide polymorphisms (SNPs) and structural variations. Genomic variation information includes known gene variations related to lung cancer, such as TP53 and EGFR. These variations can directly affect the occurrence and development of tumors.

[0040] Blood biochemical indicators, such as the levels of tumor markers (such as CEA, CA125, etc.), can provide information about the status of the tumor. Imaging test results, such as CT scans, MRI and other test results, can provide information such as the size, location and morphology of the tumor. Pathological examination data, such as histopathological examination results, can provide detailed information such as histological type and tumor grade.

[0041] The comprehensive analysis of these multi-dimensional clinical test data helps the medical team to fully understand the patient's lung cancer status and develop a personalized treatment plan. Processing these data through the target positioning network and optimization network can achieve more accurate mutation location prediction and provide more effective treatment support for lung cancer patients.

[0042] In 101, the multidimensional clinical test data of the target object is obtained. Specifically, the multidimensional clinical test data is obtained by sequencing the DNA sample of the target object through a professional gene sequencing institution or laboratory. This usually involves collecting blood, saliva or other biological samples from patients and extracting DNA for high-throughput sequencing. The sequencing results will provide the complete gene sequence data of the target object, including the nucleotide sequence of all genes. Further data analysis can identify mutation sites in genes, which may be related to lung cancer risk or drug response. For the collection of clinical test indicator data, the clinical test indicator data of the target object is obtained from the medical records of the hospital or clinic. These data may include the results of blood routine, biochemical tests, tumor marker tests, etc. For lung cancer patients, special attention is paid to indicators related to lung cancer, such as the levels of tumor markers such as CEA (carcinoembryonic antigen). These data can be retrieved and sorted through the hospital's electronic medical record system or laboratory information system. Then, the gene sequence data and clinical test indicator data are integrated to ensure that they can be effectively used for subsequent analysis and prediction. It may be necessary to standardize the data to eliminate systematic errors between different laboratories or detection methods so that the data are comparable. Further, optionally, the collected multidimensional clinical test data is quality controlled and verified to ensure the accuracy and reliability of the data. This may include checks on data integrity, consistency and rationality, as well as necessary data cleaning and preprocessing steps. Through the above steps, the multidimensional clinical test data of the target object can be systematically obtained, providing a solid foundation for the subsequent prediction of lung cancer mutation locations and the formulation of personalized treatment plans.

[0043] In the embodiment of the present application, the basic data at least includes: case data and basic physical attribute information of the target object. Specifically, in the embodiment of the present application, the basic data covers the case data and basic physical attribute information of the target object. For example, in the case data, the diagnostic history includes the target object's past diagnostic records, such as whether there has been a diagnosis of lung cancer or other related diseases, and any previous treatment records. Family medical history, recording whether there is a history of lung cancer or other cancers in the target object's family, is important for assessing genetic risks and predicting disease progression. Medication records, including the target object's past medication and reactions, especially drugs related to lung cancer treatment. Surgery and radiotherapy records, if the target object has undergone surgery or radiotherapy before, these records will also be included in the case data. For another example, in the basic physical attribute information, age and gender are important factors to consider when assessing lung cancer risks and treatment options. Height and weight are used to calculate body mass index (BMI), which is related to the patient's overall health and drug dosage calculation. Lifestyle habits, including smoking history, drinking history, eating habits, etc., all of which may affect the risk and treatment effect of lung cancer. Co-morbidities: record whether the target subject has other chronic diseases, such as heart disease, diabetes, etc., which may affect the choice of treatment options.

[0044] In summary, case data and basic physical attribute information together constitute the basic data of the target object, which is crucial for assessing lung cancer risk, formulating personalized treatment plans, and monitoring treatment response. By comprehensively considering these data, the medical team can have a more comprehensive understanding of the patient's health status and provide more accurate medical services.

[0045] As an optional embodiment, it is assumed that the target localization network includes at least the following structures: a preprocessing layer, a feature extraction layer, and a prediction layer. Based on the above structure, in 102, the multidimensional clinical test data is input into the target localization network, and the target object is subjected to mutation localization prediction processing by the target localization network to obtain the mutation localization features of the target object, including the following steps:

[0046] 1021, performing multi-branch preprocessing on the multi-dimensional clinical test data through a preprocessing layer to obtain multiple groups of test preprocessing data;

[0047] 1022, performing multi-dimensional feature extraction processing on the multiple groups of detection pre-processing data through a feature extraction layer to obtain a local feature coding sequence and a global feature coding sequence of the target object;

[0048] 1023, through the prediction layer, according to the standard gene structure parameters, the local feature coding sequence and the global feature coding sequence of the target object are predicted for mutation location, and the mutation location feature of the target object is obtained.

[0049] In the target localization network, the preprocessing layer, feature extraction layer and prediction layer are the key components of the network, and these structures are responsible for different stages of data processing.

[0050] It is understandable that the purpose of the preprocessing layer is to standardize and filter the multidimensional clinical test data so that it can be better processed by the subsequent layers. The dynamically configured preprocessing branch processes each type of data separately. For example, image data may be converted to a different format, resized, etc., while text data may need to be segmented, cleaned, etc. The preprocessing mechanism may include data cleaning, normalization, format conversion, etc. to ensure the consistency of quality and structure of different types of data, which is convenient for subsequent feature extraction.

[0051] Exemplarily, in 1021, the dynamically configured preprocessing branch refers to performing specific preprocessing operations for different types of data to ensure that each data type can be effectively utilized in the subsequent processing stage. The following are several examples to illustrate the application of dynamically configured preprocessing branches: For medical imaging data, the preprocessing branch may include image format conversion (such as conversion from DICOM to JPEG or PNG) and image size adjustment (scaling or cropping) to meet the requirements of model input. It may also include enhancement processing, such as contrast adjustment, noise removal, or magnification of local areas to better observe potential mutations or abnormalities. For genomic data, the preprocessing branch may involve sequence normalization processing to ensure that all base sequences are represented in a unified format. It may also include annotation of gene mutation sites, or screening and filtering of gene mutations based on known databases. For clinical text data, the preprocessing branch may include natural language processing techniques, such as word segmentation, removal of stop words, stem extraction, or part-of-speech tagging, to prepare text data for feature extraction. In addition, entity recognition can also be performed to identify key information such as diseases, drugs, or treatment methods in the text. For blood or biochemical data, the preprocessing branch may involve the processing of outliers, such as detecting and correcting possible measurement errors or laboratory errors. It may also include unit unification or data normalization of different biomarkers to ensure data comparability and consistency. In summary, in 1021, each preprocessing branch designs a specific processing flow and algorithm according to the data type it processes and the characteristics of the data source to ensure that the data has been fully prepared and optimized before entering the feature extraction layer. This dynamic configuration can be flexibly adjusted according to the characteristics and requirements of different data types to adapt to complex medical data analysis tasks.

[0052] In an embodiment of the present application, the preprocessing layer is dynamically configured with preprocessing branches corresponding to different data types in the multidimensional clinical test data, and each preprocessing branch is provided with a corresponding type of preprocessing mechanism. In an embodiment of the present application, the dynamic configuration of the preprocessing layer is to design specific preprocessing branches according to the characteristics and requirements of different data types in the multidimensional clinical test data, and each branch is provided with a specific type of preprocessing mechanism to ensure that the data can be effectively utilized in the subsequent processing stage. First, in the image data preprocessing branch, for medical imaging data, such as CT scans or MRI images, the preprocessing branch may include the following operations:

[0053] Image format conversion: Convert medical images in DICOM format to standard image formats (such as JPEG, PNG) for easy model processing.

[0054] Image resizing: Adjust the resolution or size of the image based on the model input requirements to ensure consistency of the input model.

[0055] Image enhancement: This may include contrast enhancement, noise removal, or magnification of local areas to help the model better detect and analyze potential anomalies or mutations.

[0056] Second, in the gene sequence data preprocessing branch, for genomic data, the preprocessing branch may involve the following processing:

[0057] Sequence normalization: Ensure that all gene sequences are represented in a uniform format, such as standardizing base sequences to uppercase letters or a specific encoding format.

[0058] Variant site annotation: Based on known databases or genome references, possible gene variation sites are marked and screened for subsequent feature extraction and analysis.

[0059] Third, in the clinical text data preprocessing branch, for clinical text data extracted from medical records, the preprocessing branch may include the following steps:

[0060] Natural Language Processing (NLP) techniques: such as tokenization, stop word removal, stemming, or part-of-speech tagging, to convert text into structured data that can be used in models.

[0061] Entity Recognition: Identify and tag key entities in text, such as disease names, drug names, or treatments, to help the model understand and analyze the content of the text.

[0062] Fourth, in the biochemical data preprocessing branch, for blood or biochemical data, the preprocessing branch may perform the following operations:

[0063] Outlier processing: Detect and correct possible measurement errors or laboratory errors to ensure data accuracy and reliability.

[0064] Data normalization: The measurement results of different biomarkers are unified or normalized to eliminate the differences between different laboratories or equipment.

[0065] Each preprocessing branch uses a specific algorithm and processing flow according to the specific data type it processes, with the goal of converting the original multidimensional clinical test data into a standardized, structured data format suitable for model input. This dynamic configuration and personalized preprocessing mechanism can effectively optimize the data and provide high-quality input for the subsequent feature extraction and prediction stages, thereby improving the overall analysis and prediction results.

[0066] The function of the feature extraction layer is to extract coding sequences that can represent the local and global features of the target object from the preprocessed data. Local feature coding sequences may include gene expression levels and protein levels in a specific region, while global feature coding sequences may involve variation information of the entire genome, the overall health status of the patient, etc. For example, feature extraction usually uses deep learning models, which can automatically learn complex features in the data.

[0067] The prediction layer receives the local and global feature coding sequences output by the feature extraction layer, and combines standard gene structure parameters (such as known gene mutation sites, gene expression patterns, etc.) to predict mutation location. Standard gene structure parameters can be used as prior knowledge input to guide and correct the prediction results of the deep learning model. The prediction layer may use a fully connected layer or other neural network structure to combine the input features and parameters to output the mutation location features of the target object.

[0068] When predicting mutation location, the target location network uses the above structure to perform multi-branch preprocessing, multi-dimensional feature extraction and mutation location prediction on the input multi-dimensional clinical test data, and finally obtains the mutation location features of the target object. These features can be used for further analysis, such as the formulation of personalized treatment plans, drug response prediction, etc.

[0069] It is worth noting that the specific design and steps of the above structure may vary depending on the specifics of the actual application requirements and available data, and experiments and adjustments are needed to obtain optimal performance.

[0070] As an optional embodiment, it is assumed that the prediction layer includes at least the following structures: a reference prediction layer, a target prediction layer, and an output layer. Based on the above structure, in 1023, the prediction layer performs mutation location prediction on the local feature coding sequence and the global feature coding sequence of the target object according to the standard gene structure parameters to obtain the mutation location feature of the target object, including the following steps:

[0071] 231, through the reference prediction layer, according to the standard gene structure parameters, the local feature coding sequence and the global feature coding sequence of the target object are reference predicted to obtain a reference gene structure sample of the target object;

[0072] 232, through the target prediction layer, target prediction is performed on the local feature coding sequence and the global feature coding sequence of the target object to obtain a target gene structure sample of the target object;

[0073] 233, through the output layer, the reference gene structure sample is compared and predicted with the target gene structure sample to obtain the mutation location feature of the target object.

[0074] In the embodiment of the present application, the target prediction and the reference prediction correspond to different structural sample parameters respectively. Further optionally, the structural sample parameters of the target prediction are dynamically configured by the preset indicators of the target object in the multidimensional clinical test data. That is, the structural sample parameters refer to the standard gene structure parameters used in the prediction layer, which are usually prior knowledge, including known gene mutation sites, characteristics of normal gene structure samples, etc. These parameters are used in the reference prediction and target prediction stages to help predict the mutation location characteristics of the target object.

[0075] 231, in the reference prediction layer, the local feature coding sequence and the global feature coding sequence of the target object are predicted based on the standard gene structure parameters. The purpose of this step is to generate a reference gene structure sample of the target object. The reference prediction stage may use known normal samples or common gene structures to predict the gene structure characteristics of the target object, such as normal genotype or unmutated state.

[0076] 232. In the target prediction layer, the local feature coding sequence and the global feature coding sequence of the target object are predicted based on the standard gene structure parameters. The purpose of this step is to generate a target gene structure sample of the target object. The target prediction stage is similar to the reference prediction, but it is aimed at the possible mutations or abnormal gene structures of the target object, and the comparison with the known gene structure.

[0077] In 233, the output layer receives the reference gene structure sample and the target gene structure sample for comparison and analysis. Through comparative prediction, the output layer can determine the mutation location characteristics of the target object, that is, identify the mutation or variation position in the target object that is different from the reference gene structure. The result of the output layer can be the specific location and type of the mutation, or the potential impact of the mutation on gene function and health.

[0078] In the entire prediction process, structural sample parameters play an important role, providing prior knowledge about normal and abnormal gene structures, helping the prediction model to accurately locate and analyze possible mutations in the target object. This hierarchical prediction method can effectively improve the model's analytical ability and prediction accuracy for complex multidimensional clinical test data.

[0079] Another optional embodiment is that in 232, target prediction is performed on the local feature coding sequence and the global feature coding sequence of the target object to obtain a target gene structure sample of the target object, including the following steps:

[0080] 2321, dividing the genome sequence of the target object into a plurality of nodes; wherein each node identifies a specific region in the genome, and the specific region includes at least one of a chromosome, a chromatin, a promoter factor, and an enhancer factor;

[0081] 2322, splitting the local feature coding sequence of the target object according to the corresponding node positions to obtain multiple local feature nodes of the target object;

[0082] 2323, constructing a connection relationship between multiple local feature nodes of the target object according to the dependency relationship between each local feature coding element in the local feature coding sequence; wherein the connection relationship represents the gene space structure relationship between the corresponding specific regions; the gene space structure relationship at least includes: chromosome distance, and three-dimensional spatial structure characteristics of chromatin;

[0083] 2324, based on the overall structural characteristics in the global feature coding sequence, global feature aggregation is performed on multiple local feature nodes and corresponding connection relationships to obtain a target gene structure sample of the target object.

[0084] In the above steps, first, the genome sequence of the target object is divided into multiple nodes according to certain rules. Each node represents a specific region in the genome, which includes at least one of chromosomes, chromatin, promoter factors, and enhancer factors. Such segmentation can decompose the complex structure of the genome into smaller manageable units. For example, assuming that the structure of a specific gene is to be analyzed, the genome sequence can be divided into multiple nodes in the following way:

[0085] Chromosome Region: Chromosomes are one of the largest structural units in the genome and contain many genes. The segmentation of chromosomes can help distinguish the genomic structural features on different chromosomes, such as chromosome 21 or chromosome X in the human genome.

[0086] Chromatin regions: Chromatin is a complex of DNA and proteins on chromosomes, and its structure and organization may vary from region to region. Segmenting chromatin regions can help analyze gene regulatory regions or epigenetic features.

[0087] Promoters and enhancers: These are regulatory elements in the genome that can influence the expression of genes. Dissecting these regions helps understand the regulatory networks of genes and their role in biological processes.

[0088] For example, for a chromosome region containing multiple genes, it can be divided into different nodes, each of which represents a gene or a group of related genes on the chromosome. Each node can contain the coding region of the gene, regulatory sequences, and its relationship with neighboring genes. This segmentation and regional definition makes the complexity of the genome easier to manage and analyze, providing an important foundation for a deep understanding of genome structure and function.

[0089] Then, for each segmented node, the local feature encoding sequence of the target object is split. This step aims to extract local features from each node, which reflect the gene sequence, structure or other important characteristics in a specific region of the genome.

[0090] After the target object's genome sequence is node-segmented, the local feature coding sequence is split for each node. These local features reflect the important characteristics in a specific genomic region. The following is an example to illustrate this process. Suppose a specific chromosome region in the human genome is being analyzed, such as chromosome 13. This chromosome region contains multiple genes, which are divided into several nodes, each of which represents a gene on the chromosome or a genomic region.

[0091] Node 1: Coding region of gene A. The coding region of gene A is selected as the first node on the chromosome. The local characteristic coding sequence may include the specific base sequence of gene A, such as ATCG sequence, and epigenetic markers associated therewith.

[0092] Node 2: Regulatory regions of gene B. The second node may involve regulatory regions of gene B, which are located close to gene B on the chromosome. Local feature coding sequences may include specific sequence elements of these regulatory regions, such as structural features of promoters or enhancers.

[0093] Node 3: Region between neighboring genes C and D. The third node may involve the region between neighboring genes C and D. The local characteristic coding sequence will reflect the genomic structural features between these two genes, which may include the distance between genes C and D and the interaction of their regulatory sequences.

[0094] In the process of splitting the local feature coding sequence of each node, the key is to identify and extract information that reflects the important characteristics of the specific region of the genome. This information includes not only the composition of the base sequence, but also epigenetic modifications, DNA structural characteristics, and regulatory sequences that may affect gene expression. By analyzing and comparing these local features, we can have a deeper understanding of the structure and function of the genomic region, thus providing a basis for predicting changes in gene structure and its potential impact.

[0095] Next, based on the dependency between each local feature coding element in the local feature coding sequence, the connection relationship between multiple local feature nodes of the target object is constructed. These connection relationships can reflect the association and distance in the gene spatial structure, such as the distance between chromosomes or the three-dimensional spatial structural characteristics of chromatin.

[0096] Finally, based on the overall structural characteristics in the global feature coding sequence, multiple local feature nodes and the connection relationships between them are aggregated globally. This process can integrate the local information of a single node to form a target gene structure sample of the target object. Global feature aggregation takes into account the comprehensive properties of the entire genome or important genomic regions, such as the chromosome organization state of the whole genome or specific structural features on chromatin.

[0097] In the process of constructing connection relationships between local feature coding sequences and aggregating global features, the following is an example to illustrate this process. It is assumed that the analysis is of a chromosomal region in the human genome, which contains multiple genes and related regulatory elements.

[0098] Node 1: Coding region of gene A. Local characteristic coding sequence, ATCG sequence, methylation modification site information. This node may be associated with neighboring nodes 2 and 3, because the expression of gene A may be affected by these neighboring regions.

[0099] Node 2: Regulatory region of gene B. Local characteristic coding sequence, promoter structure, enhancer location and regulatory factor binding site. This node may be connected to nodes 1 and 3, because the regulatory region of gene B may interact with the coding region of gene A and the coding region of the neighboring gene C.

[0100] Node 3: The region between neighboring genes C and D. Local feature coding sequences, the distance between genes C and D, and the distribution of common regulatory elements. This node is also connected to nodes 1 and 2 because it involves the genomic region adjacent to genes A and B.

[0101] When constructing the connectivity relationships, the dependencies and interactions between these nodes are taken into account to reflect the associations and distances in the spatial structure of genes, such as the interactions between different regions on chromosomes and their possible influences.

[0102] After the connection relationship is constructed, the local information of each node is integrated through global feature aggregation to form a target gene structure sample of the target object. The global feature coding sequence provides the state and structural characteristics of the overall genomic region. By considering the connection relationship between nodes and their respective local feature coding sequences, the comprehensive properties of the overall gene structure can be calculated. For example, the DNA sequences, epigenetic features and organizational structures of different nodes in three-dimensional space are integrated to form a comprehensive understanding of the chromosome region. This process not only helps to predict the gene structure of the target object, but also provides important data support for understanding the complex regulatory networks and functional regions in the genome. Through the fine analysis of local features and the comprehensive processing of global features, the complexity of the structure and function of the genome can be deeply explored.

[0103] Through the above steps 2321 to 2324, the target prediction stage can effectively analyze and reconstruct the gene structure of the target object, from local features to global structure modeling and prediction, which helps to identify possible gene mutations or other structural changes. This method not only provides a deep understanding of the genome, but also provides important basic data for subsequent mutation location and functional analysis.

[0104] As an optional embodiment, it is assumed that the optimization network includes the following structure: a recognition layer, a self-positioning layer, and an optimization layer. Based on this structure, in 103, the mutation positioning feature and the basic data of the target object are input into the optimization network, and the mutation positioning feature is self-positioned optimized to obtain the target mutation positioning feature of the target object, including the following steps:

[0105] 301, identifying, through the recognition layer, the candidate mutation type to which the target object belongs and the applicable application scenario based on the basic data;

[0106] 302, performing self-positioning reconstruction processing on the mutation positioning feature through the self-positioning layer to obtain candidate mutation positioning features of the target object; wherein the construction mechanism of the dynamic configuration in the self-positioning reconstruction processing is determined based on the candidate mutation type and the application scenario;

[0107] 303 , optimizing the candidate mutation location features of the target object through the optimization layer to obtain the target mutation location features of the target object.

[0108] In this optional embodiment, the optimization network consists of three main layers: recognition layer, self-positioning layer and optimization layer. These layers work together to accurately optimize the mutation location features. First, the recognition layer receives basic data of the target object as input. These data may include genome sequence, clinical manifestations, family history, etc. Based on these data, the recognition layer uses machine learning and pattern recognition techniques to identify the mutation type that the target object may belong to, such as point mutation, insertion or deletion. At the same time, this layer also determines in which specific application scenarios these mutation types are of great significance, such as disease diagnosis, drug response prediction, etc. Next, the self-positioning layer will perform self-positioning reconstruction processing on the mutation location features based on the candidate mutation types and application scenario information provided by the recognition layer. The purpose of this step is to fine-tune the original features to improve the accuracy of subsequent analysis. In the self-positioning reconstruction process, the dynamically configured construction mechanism plays a key role. This mechanism adjusts the reconstruction algorithm and parameters according to different mutation types and application scenarios to ensure the pertinence and effectiveness of feature extraction. For example, when facing a certain type of mutation, the construction mechanism may emphasize certain specific genomic regions or sequence patterns to more accurately locate the mutation position. After the self-positioning reconstruction process, the candidate mutation location features of the target object are obtained. These features have been optimized according to the specific mutation type and application scenario, laying a solid foundation for subsequent optimization processing. Finally, the optimization layer will further optimize the candidate mutation location features. The purpose of this step is to eliminate possible noise and redundant information, highlight key features, and thus obtain more accurate and reliable target mutation location features. In fact, the optimization process may include a variety of techniques, such as feature selection, dimensionality reduction, clustering, etc. These techniques help to extract the most representative features while reducing the complexity and computational complexity of the data. After the optimization process, the target mutation location features of the target object are obtained, which will be used for subsequent analysis and decision support.

[0109] In general, this optimization network can achieve accurate optimization of mutation location features through the collaborative work of the recognition layer, self-localization layer and optimization layer. This not only improves the accuracy of gene mutation analysis, but also provides strong support for related biomedical research and clinical applications.

[0110] Further optionally, the construction mechanism includes at least a gene positioning microstructure parameter and a mutant gene position offset correction parameter.

[0111] Among them, the dynamic configuration factor T of the gene positioning microstructure parameter or the mutation gene position offset correction parameter i It is calculated by the following function; wherein the function is expressed as the following formula:

[0112]

[0113] Among them, T i represents the dynamic configuration factor of the ith mutation gene type, k represents the total number of mutation structures of the ith mutation gene type, and y i1 represents the factor to be configured for the i-th mutant gene type, y i2 represents the actual dynamic configuration factor of the ith mutant gene type, represents the variable characteristic distance of the dynamic configuration factor, β and α are the adjustable empirical thresholds corresponding to the i-th mutant gene type, and p i The gene structure preset parameters corresponding to the i-th mutant gene type are determined by the stage of the gene mutation, the gene mutation family factors and / or the gene mutation expression performance.

[0114] The purpose of this formula is to adaptively update the dynamic configuration factor according to the characteristics of each mutant gene type and environmental conditions to optimize the accuracy and adaptability of the mutation location characteristics. Through such a configuration mechanism, more accurate optimization results can be provided under different mutation conditions, thereby enhancing the prediction and analysis capabilities of the target object's mutation location characteristics.

[0115] As an optional embodiment, in 104, generating mutation prediction analysis information of the target object based on the target mutation location feature includes the following steps:

[0116] 401, constructing a mutation localization spatial model according to the target mutation localization feature;

[0117] 402, input the mutation localization space model into the mutation analysis model, analyze the mutation distribution law between each mutation gene position point of the target object in the mutation localization space model, and determine whether the mutation distribution law meets the pre-configured gene mutation detection condition to obtain the mutation analysis result;

[0118] 403 , predicting candidate genes that may mutate in the target object in the current state based on the mutation analysis result, and outputting the candidate genes as the mutation prediction analysis information.

[0119] These steps describe the process of generating mutation prediction analysis information of the target object based on the target mutation location features. First, based on the obtained target mutation location features, we construct a mutation location space model. This model is based on the spatial representation of mutation locations and related features, usually including the location coordinates of the mutant gene, environmental factors on the genome, and possible influencing factors such as chromosome structure or epigenetic modifications.

[0120] Secondly, the constructed mutation localization space model is input into the mutation analysis model, which is used to analyze the mutation distribution pattern between the various mutation gene position points of the target object in the mutation localization space model. The goal of this step is to evaluate the distribution of mutations in the genome, as well as the mutation types, density and possible aggregation. In this process, the pre-configured gene mutation detection conditions are also used to determine whether the mutation distribution pattern meets expectations, such as whether it meets the distribution pattern of certain disease-related mutations or specific gene family mutation patterns.

[0121] Finally, based on the results of mutation analysis, predict candidate genes that may mutate in the target object in the current state. These candidate genes are usually based on the output of the mutation analysis model, combined with relevant bioinformatics and medical databases, to infer the genes that may change and their mutation types. This prediction information can include important biomedical information such as potential disease-related gene mutations and drug response-related gene mutations, providing support and guidance for subsequent clinical diagnosis, personalized treatment or research.

[0122] In summary, this process explores the mutation characteristics of the target object and its possible biological significance at the genomic level through multi-level data integration and analysis, providing important data support and decision-making basis for precision medicine and genomics research.

[0123] In an embodiment of the present application, a method for predicting the mutation position of a lung cancer gene is provided. First, multidimensional clinical test data of a target object is obtained. The multidimensional clinical test data at least includes: gene sequence data and clinical test index data of the target object. Then, the multidimensional clinical test data is input into a target positioning network, and the target object is subjected to mutation positioning prediction processing through the target positioning network to obtain the mutation positioning feature of the target object. Then, the mutation positioning feature and the basic data of the target object are input into an optimization network, and the mutation positioning feature is self-positioned and optimized to obtain the target mutation positioning feature of the target object. Among them, the basic data at least includes: case data and basic physical attribute information of the target object. Finally, mutation prediction analysis information of the target object is generated based on the target mutation positioning feature to assist in optimizing the lung cancer mutation detection of the target object. In an embodiment of the present application, intelligent mutation position prediction of lung cancer genes is realized, which is convenient for relevant personnel to refer to mutation prediction analysis to realize gene mutation detection of the target object, thereby assisting in improving the accuracy and reliability of gene screening and assisting in improving the effect of targeted therapy.

[0124] In another embodiment of the present application, a system for predicting the mutation position of lung cancer genes is also provided. Figure 2 The lung cancer gene mutation position prediction system comprises the following units:

[0125] An acquisition unit is configured to acquire multi-dimensional clinical test data of a target object; the multi-dimensional clinical test data at least includes: gene sequence data and clinical test index data of the target object;

[0126] The detection unit is configured to input the multidimensional clinical detection data into a target positioning network, perform mutation positioning prediction processing on a target object through the target positioning network, and obtain a mutation positioning feature of the target object; input the mutation positioning feature and basic data of the target object into an optimization network, perform self-positioning optimization on the mutation positioning feature, and obtain a target mutation positioning feature of the target object; the basic data at least includes: case data and basic physical attribute information of the target object;

[0127] The analysis unit is configured to generate a defect analysis report of the target object based on the target defect feature to assist in optimizing the detection of the target object.

[0128] Further optionally, the target localization network comprises at least the following structures: a preprocessing layer, a feature extraction layer, and a prediction layer; the detection unit, when inputting the multidimensional clinical detection data into the target localization network, performing mutation localization prediction processing on the target object through the target localization network, and obtaining the mutation localization features of the target object, is configured as follows:

[0129] Through the preprocessing layer, the multi-dimensional clinical test data is subjected to multi-branch preprocessing to obtain multiple groups of test preprocessing data; wherein the preprocessing layer is dynamically configured with preprocessing branches corresponding to different data types in the multi-dimensional clinical test data, and each preprocessing branch is provided with a corresponding type of preprocessing mechanism;

[0130] Through the feature extraction layer, multi-dimensional feature extraction processing is performed on the multiple groups of detection pre-processed data to obtain local feature coding sequences and global feature coding sequences of the target object;

[0131] Through the prediction layer, according to the standard gene structure parameters, the mutation location prediction is performed on the local feature coding sequence and the global feature coding sequence of the target object to obtain the mutation location characteristics of the target object.

[0132] Further optionally, the prediction layer at least includes the following structures: a reference prediction layer, a target prediction layer, and an output layer; the detection unit, through the prediction layer, performs mutation location prediction on the local feature coding sequence and the global feature coding sequence of the target object according to the standard gene structure parameters to obtain the mutation location feature of the target object, is configured as follows:

[0133] Through the reference prediction layer, according to the standard gene structure parameters, the local feature coding sequence and the global feature coding sequence of the target object are reference predicted to obtain the reference gene structure sample of the target object;

[0134] Through the target prediction layer, target prediction is performed on the local feature coding sequence and the global feature coding sequence of the target object to obtain the target gene structure sample of the target object; wherein the target prediction and the reference prediction correspond to different structure sample parameters respectively; the structure sample parameters of the target prediction are dynamically configured by the preset indicators of the target object in the multidimensional clinical test data;

[0135] Through the output layer, the reference gene structure sample is compared and predicted with the target gene structure sample to obtain the mutation location characteristics of the target object.

[0136] Further optionally, when the detection unit performs target prediction on the local feature coding sequence and the global feature coding sequence of the target object to obtain the target gene structure sample of the target object, it is configured as follows:

[0137] Divide the genome sequence of the target object into multiple nodes; wherein each node identifies a specific region in the genome, and the specific region includes at least one of a chromosome, a chromatin, a promoter factor, and an enhancer factor;

[0138] According to the corresponding node positions, the local feature coding sequence of the target object is split to obtain multiple local feature nodes of the target object;

[0139] According to the dependency relationship between each local feature coding element in the local feature coding sequence, a connection relationship between multiple local feature nodes of the target object is constructed; wherein the connection relationship represents the gene space structure relationship between the corresponding specific regions; the gene space structure relationship at least includes: chromosome distance, three-dimensional spatial structure characteristics of chromatin;

[0140] Based on the overall structural characteristics in the global feature coding sequence, global feature aggregation is performed on multiple local feature nodes and the corresponding connection relationships to obtain the target gene structure sample of the target object.

[0141] Further optionally, the optimization network includes the following structures: a recognition layer, a self-positioning layer, and an optimization layer; the detection unit, when inputting the mutation positioning feature and basic data of the target object into the optimization network, performs self-positioning optimization on the mutation positioning feature, and obtains the target mutation positioning feature of the target object, is configured as follows:

[0142] Through the recognition layer, based on the basic data, the candidate mutation type to which the target object belongs and the applicable application scenario are identified;

[0143] Through the self-positioning layer, the mutation positioning feature is self-positioned and reconstructed to obtain the candidate mutation positioning feature of the target object; wherein the construction mechanism of the dynamic configuration in the self-positioning and reconstruction process is determined based on the candidate mutation type and the application scenario;

[0144] Through the optimization layer, the candidate mutation location features of the target object are optimized to obtain the target mutation location features of the target object.

[0145] Further optionally, the construction mechanism includes at least a gene positioning microstructure parameter and a mutation gene position offset correction parameter; wherein the dynamic configuration factor T of the gene positioning microstructure parameter or the mutation gene position offset correction parameter is i It is calculated by the following function; wherein the function is expressed as the following formula:

[0146]

[0147] Among them, T i represents the dynamic configuration factor of the ith mutation gene type, k represents the total number of mutation structures of the ith mutation gene type, and y i1 represents the factor to be configured for the i-th mutant gene type, y i2 represents the actual dynamic configuration factor of the ith mutant gene type, represents the variable characteristic distance of the dynamic configuration factor, β and α are the adjustable empirical thresholds corresponding to the i-th mutant gene type, and p i The gene structure preset parameters corresponding to the i-th mutant gene type are determined by the stage of the gene mutation, the gene mutation family factors and / or the gene mutation expression performance.

[0148] Further optionally, when the analysis unit generates the mutation prediction analysis information of the target object based on the target mutation location feature, it is configured to:

[0149] Constructing according to the target mutation location characteristics to obtain a mutation location space model;

[0150] Inputting the mutation localization space model into the mutation analysis model, analyzing the mutation distribution law between the various mutation gene position points of the target object in the mutation localization space model, and determining whether the mutation distribution law meets the pre-configured gene mutation detection conditions to obtain the mutation analysis result;

[0151] Based on the mutation analysis results, candidate genes of the target object that may have mutations in the current state are predicted and output as the mutation prediction analysis information.

[0152] The system can implement various steps in the above method embodiment, which will not be expanded here.

[0153] In the embodiments of the present application, intelligent mutation position prediction of lung cancer genes can be achieved, which is convenient for relevant personnel to refer to mutation prediction analysis to realize gene mutation detection of target objects, thereby assisting in improving the accuracy and reliability of gene screening and assisting in improving the effect of targeted treatment.

[0154] See also Figure 3 , Figure 3 Schematic diagram of an electronic device provided by an embodiment of the present invention. Figure 3 As shown, an embodiment of the present invention provides an electronic device 500, including a memory 510, a processor 5102, and a computer program 511 stored in the memory 510 and executable on the processor 5102. When the processor 5102 executes the computer program 511, the following steps are implemented: obtaining multidimensional clinical test data of a target object; the multidimensional clinical test data at least includes: gene sequence data and clinical test index data of the target object; inputting the multidimensional clinical test data into a target positioning network, performing mutation positioning prediction processing on the target object through the target positioning network, and obtaining a mutation positioning feature of the target object; inputting the mutation positioning feature and the basic data of the target object into an optimization network, performing self-positioning optimization on the mutation positioning feature, and obtaining a target mutation positioning feature of the target object; the basic data at least includes: case data and basic physical attribute information of the target object; generating mutation prediction analysis information of the target object based on the target mutation positioning feature to assist in optimizing lung cancer mutation detection of the target object.

[0155] See also Figure 4 , Figure 4 A schematic diagram of an embodiment of a computer-readable storage medium provided in an embodiment of the present invention. Figure 4 As shown, this embodiment provides a computer-readable storage medium 600, on which a computer program 611 is stored. When the computer program 611 is executed by a processor, the following steps are implemented: obtaining multidimensional clinical test data of a target object; the multidimensional clinical test data at least includes: gene sequence data and clinical test index data of the target object; inputting the multidimensional clinical test data into a target positioning network, performing mutation positioning prediction processing on the target object through the target positioning network, and obtaining a mutation positioning feature of the target object; inputting the mutation positioning feature and basic data of the target object into an optimization network, performing self-positioning optimization on the mutation positioning feature, and obtaining a target mutation positioning feature of the target object; the basic data at least includes: case data and basic physical attribute information of the target object; generating mutation prediction analysis information of the target object based on the target mutation positioning feature to assist in optimizing lung cancer mutation detection of the target object.

[0156] It should be noted that in the above embodiments, the description of each embodiment has its own emphasis, and for parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0157] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented 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.

[0158] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0159] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0160] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0161] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0162] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for predicting the mutation position of lung cancer genes, It is characterized in that The method comprises: Acquire multidimensional clinical test data of the target object; the multidimensional clinical test data at least includes: gene sequence data and clinical test index data of the target object; Inputting the multidimensional clinical test data into a target localization network, performing mutation localization prediction processing on a target object through the target localization network, and obtaining mutation localization features of the target object; The mutation location feature and basic data of the target object are input into the optimization network, and the mutation location feature is self-optimized to obtain the target mutation location feature of the target object; the basic data at least includes: case data and basic physical attribute information of the target object; Based on the target mutation location features, mutation prediction analysis information of the target object is generated to assist in optimizing lung cancer mutation detection of the target object.

2. The method for predicting the mutation position of a lung cancer gene according to claim 1, It is characterized in that The target positioning network at least includes the following structures: a preprocessing layer, a feature extraction layer, and a prediction layer; The step of inputting the multidimensional clinical test data into a target positioning network, performing mutation positioning prediction processing on a target object through the target positioning network, and obtaining mutation positioning features of the target object comprises: Through the preprocessing layer, the multi-dimensional clinical test data is subjected to multi-branch preprocessing to obtain multiple groups of test preprocessing data; wherein the preprocessing layer is dynamically configured with preprocessing branches corresponding to different data types in the multi-dimensional clinical test data, and each preprocessing branch is provided with a corresponding type of preprocessing mechanism; Through the feature extraction layer, multi-dimensional feature extraction processing is performed on the multiple groups of detection pre-processed data to obtain local feature coding sequences and global feature coding sequences of the target object; Through the prediction layer, according to the standard gene structure parameters, the mutation location prediction is performed on the local feature coding sequence and the global feature coding sequence of the target object to obtain the mutation location characteristics of the target object.

3. The method for predicting the mutation position of a lung cancer gene according to claim 2, It is characterized in that The prediction layer at least includes the following structures: a reference prediction layer, a target prediction layer, and an output layer; The prediction layer performs mutation location prediction on the local feature coding sequence and the global feature coding sequence of the target object according to the standard gene structure parameters to obtain the mutation location features of the target object, including: Through the reference prediction layer, according to the standard gene structure parameters, the local feature coding sequence and the global feature coding sequence of the target object are reference predicted to obtain the reference gene structure sample of the target object; Through the target prediction layer, target prediction is performed on the local feature coding sequence and the global feature coding sequence of the target object to obtain the target gene structure sample of the target object; wherein the target prediction and the reference prediction correspond to different structure sample parameters respectively; the structure sample parameters of the target prediction are dynamically configured by the preset indicators of the target object in the multidimensional clinical test data; Through the output layer, the reference gene structure sample is compared and predicted with the target gene structure sample to obtain the mutation location characteristics of the target object.

4. The method for predicting the mutation position of a lung cancer gene according to claim 3, It is characterized in that Target prediction is performed on the local feature coding sequence and the global feature coding sequence of the target object to obtain the target gene structure sample of the target object, including: Divide the genome sequence of the target object into multiple nodes; wherein each node identifies a specific region in the genome, and the specific region includes at least one of a chromosome, a chromatin, a promoter factor, and an enhancer factor; According to the corresponding node positions, the local feature coding sequence of the target object is split to obtain multiple local feature nodes of the target object; According to the dependency relationship between each local feature coding element in the local feature coding sequence, a connection relationship between multiple local feature nodes of the target object is constructed; wherein the connection relationship represents the gene space structure relationship between the corresponding specific regions; the gene space structure relationship at least includes: chromosome distance, three-dimensional spatial structure characteristics of chromatin; Based on the overall structural characteristics in the global feature coding sequence, global feature aggregation is performed on multiple local feature nodes and the corresponding connection relationships to obtain the target gene structure sample of the target object.

5. The method for predicting the mutation position of a lung cancer gene according to claim 1, It is characterized in that The optimization network includes the following structures: an identification layer, a self-positioning layer, and an optimization layer; The step of inputting the mutation location feature and basic data of the target object into an optimization network, performing self-location optimization on the mutation location feature, and obtaining a target mutation location feature of the target object comprises: Through the recognition layer, based on the basic data, the candidate mutation type to which the target object belongs and the applicable application scenario are identified; Through the self-positioning layer, the mutation positioning feature is self-positioned and reconstructed to obtain the candidate mutation positioning feature of the target object; wherein the construction mechanism of the dynamic configuration in the self-positioning and reconstruction process is determined based on the candidate mutation type and the application scenario; Through the optimization layer, the candidate mutation location features of the target object are optimized to obtain the target mutation location features of the target object.

6. The method for predicting the mutation position of a lung cancer gene according to claim 5, It is characterized in that The construction mechanism at least includes gene positioning microstructure parameters and mutation gene position offset correction parameters; Among them, the dynamic configuration factor T of the gene positioning microstructure parameter or the mutation gene position offset correction parameter i It is calculated by the following function; wherein the function is expressed as the following formula: Among them, T i represents the dynamic configuration factor of the ith mutation gene type, k represents the total number of mutation structures of the ith mutation gene type, yi1 represents the factor to be configured of the ith mutation gene type, y i2 represents the actual dynamic configuration factor of the ith mutant gene type, e (yi1-yi2)2 represents the variable characteristic distance of the dynamic configuration factor, β and α are the adjustable empirical thresholds corresponding to the i-th mutant gene type, and p i The gene structure preset parameters corresponding to the i-th mutant gene type are determined by the stage of the gene mutation, the gene mutation family factors and / or the gene mutation expression performance.

7. The method for predicting the mutation position of a lung cancer gene according to claim 1, It is characterized in that The generating mutation prediction analysis information of the target object based on the target mutation location feature includes: Constructing according to the target mutation location characteristics to obtain a mutation location space model; Inputting the mutation localization space model into the mutation analysis model, analyzing the mutation distribution law between the various mutation gene position points of the target object in the mutation localization space model, and determining whether the mutation distribution law meets the pre-configured gene mutation detection conditions to obtain the mutation analysis result; Based on the mutation analysis results, candidate genes of the target object that may have mutations in the current state are predicted and output as the mutation prediction analysis information.

8. A system for predicting the mutation location of lung cancer genes. It is characterized in that The system comprises the following units, wherein: An acquisition unit is configured to acquire multi-dimensional clinical test data of a target object; the multi-dimensional clinical test data at least includes: gene sequence data and clinical test index data of the target object; The detection unit is configured to input the multidimensional clinical detection data into a target positioning network, perform mutation positioning prediction processing on a target object through the target positioning network, and obtain a mutation positioning feature of the target object; input the mutation positioning feature and basic data of the target object into an optimization network, perform self-positioning optimization on the mutation positioning feature, and obtain a target mutation positioning feature of the target object; the basic data at least includes: case data and basic physical attribute information of the target object; The analysis unit is configured to generate a defect analysis report of the target object based on the target defect feature to assist in optimizing the detection of the target object.

9. An electronic device, It is characterized in that include: Memory for storing computer software programs; A processor is used to read and execute the computer software program, thereby implementing the method for predicting the mutation position of a lung cancer gene according to any one of claims 1 to 7.

10. A computer-readable storage medium, It is characterized in that The storage medium stores a computer software program, and when the computer software program is executed by the processor, the method for predicting the mutation position of a lung cancer gene according to any one of claims 1 to 7 is implemented.

Citation Information

Cited By

  • Chromosome abnormality detection method based on image analysis

    CN120221094A