Thyroid cancer combined diagnosis marker screening method based on plasma exosome lncRNA and miRNA
By sequencing and analyzing lncRNA and miRNA in plasma exosomes, a thyroid cancer diagnosis model was constructed and the landmark lncRNA and miRNA were screened out, which solved the problem of insufficient early diagnosis sensitivity and specificity of thyroid cancer in the prior art, and achieved a more efficient and accurate diagnosis.
Patent Information
- Application Number
- CN202411905050.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-23
- Publication Date
- 2025-04-11
AI Technical Summary
The prior art has not been able to effectively utilize lncRNA and miRNA in plasma exosomes as specific diagnostic markers for thyroid cancer, resulting in limited sensitivity and specificity of early diagnosis.
By sequencing plasma exosomes with lncRNA and miRNA, a thyroid cancer diagnosis model was constructed, variable selection was used using the LASSO algorithm, a multigene regression model was established, and the landmark lncRNA and miRNA were screened out through bioinformatics analysis and ceRNA regulation network construction.
It improves the screening efficiency and accuracy of thyroid cancer diagnostic markers, enhances the sensitivity and specificity of the diagnosis, and provides a more reliable and effective early diagnosis method.
Smart Images
Figure CN119811473A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of gene detection, and in particular to a method for screening a combined diagnostic marker for thyroid cancer based on plasma exosome lncRNA and miRNA. Background Art
[0002] Thyroid cancer, as a malignant tumor originating from thyroid tissue, is one of the most common malignant tumors in the endocrine system. In the past three decades, the incidence of thyroid cancer has continued to rise worldwide, especially in women. Thyroid cancer has become the fourth most common malignant tumor, and this trend is still growing rapidly. The aggressiveness of thyroid cancer is not only reflected in the malignant behavior of the primary lesion, but is also often accompanied by metastasis to the lateral cervical lymph nodes, which further increases the difficulty of surgery and the economic burden. Therefore, finding specific early diagnostic markers and developing effective early diagnostic methods are of great significance to improving the diagnosis and treatment of thyroid cancer.
[0003] Traditional diagnostic methods for thyroid cancer mainly rely on imaging examinations and pathological histological examinations, but these methods have limited sensitivity and specificity in early diagnosis. In recent years, with the rapid development of molecular biology, more and more studies have begun to focus on the role of non-coding RNA (such as lncRNA and miRNA) in tumor occurrence and development. As two important types of non-coding RNA, lncRNA (long non-coding RNA) and miRNA (microRNA) play a key role in the occurrence and development of thyroid cancer by participating in gene expression regulation, cell proliferation, apoptosis and other processes.
[0004] It is worth noting that lncRNA and miRNA not only exist in cells, but can also be encapsulated in extracellular vesicles, exosomes, and transmit information between cells through exosomes. Exosomes are extracellular vesicles with a diameter of 30-150nm, containing lipids, proteins, nucleic acids and other components, with a double-layer membrane structure that can protect the stability of its contents. Since exosomes exist in a variety of body fluids (such as blood, urine, saliva, etc.) and are easy to obtain through minimally invasive or even non-invasive methods, they have become a hot topic in the study of tumor diagnostic markers.
[0005] However, although lncRNA and miRNA in exosomes have many advantages as tumor diagnostic markers, there is currently no specific lncRNA and miRNA biomarker combination for thyroid cancer detection using plasma exosome samples. Summary of the invention
[0006] The purpose of the present invention is to propose a method for screening a combined diagnostic marker for thyroid cancer based on plasma exosome lncRNA and miRNA. This technical solution can improve the screening efficiency and accuracy of thyroid cancer diagnostic markers.
[0007] To achieve the above objectives, in a first aspect, the present disclosure provides a method for screening a combined diagnostic marker for thyroid cancer based on plasma exosome lncRNA and miRNA, comprising: Sequencing of lncRNA and miRNA in plasma exosomes; Construct a diagnostic model for thyroid cancer to identify differentially expressed lncRNA and miRNA; use the LASSO algorithm to select variables for the screened differentially expressed lncRNA and miRNA; establish a multi-gene regression model to screen for iconic miRNA / IncRNA; Conduct bioinformatics analysis to find key lncRNAs and miRNAs and their biological functions; Construct ceRNA regulatory network, predict the interaction between lncRNA and miRNA, verify the targeted mRNA of miRNA, screen out significant ceRNA relationships, and perform functional analysis on key lncRNA and miRNA in ceRNA network; The results of the polygenic regression model were cross-validated with those of the bioinformatics analysis.
[0008] Beneficial effects of the basic scheme: This application is different from the traditional scheme in that the lncRNA and miRNA contained in the plasma exosomes are extracted, which have higher stability than these molecules directly free in the plasma. The lncRNA and miRNA in the exosomes are a class of small molecules that are stable, easy to measure, and can reflect the characteristics of early tumors.
[0009] By integrating the sequencing data of lncRNA and miRNA, the molecular characteristics of thyroid cancer can be more comprehensively reflected, thereby improving the accuracy of diagnosis. As important non-coding RNAs, lncRNA and miRNA play a key role in the occurrence and development of thyroid cancer, and their abnormal expression can provide a strong molecular basis for the diagnosis of thyroid cancer.
[0010] Using the LASSO algorithm to select variables for differentially expressed lncRNA and miRNA can effectively avoid the overfitting problem and screen out the genes that contribute most to the diagnosis of thyroid cancer, thereby improving the stability and predictive ability of the model.
[0011] The above-mentioned multi-level marker screening strategy enables us to identify more potential diagnostic markers for thyroid cancer and increase the sensitivity and specificity of diagnosis.
[0012] Through bioinformatics analysis, we can further explore the biological functions of lncRNA and miRNA and reveal their potential mechanisms of action in thyroid cancer. Through multi-angle analysis, we can ensure the biological relevance and functional importance of the screened markers, which will help understand the pathogenesis of thyroid cancer and provide new targets and ideas for treatment.
[0013] Constructing the ceRNA regulatory network can predict the interaction between lncRNA and miRNA and verify the targeted mRNA of miRNA, which can reveal the complex regulatory relationship between lncRNA, miRNA and mRNA, providing a new perspective to understand the role of lncRNA and miRNA in thyroid cancer, and also providing a theoretical basis for subsequent functional verification and mechanism research.
[0014] The results of the multi-gene regression model were cross-validated with the results of bioinformatics analysis to ensure the reliability and practicality of the selected lncRNA and miRNA in the diagnosis of thyroid cancer, and to effectively reduce the redundancy of the analysis results, thereby improving the accuracy and clinical application value of the markers. The results of the mutual validation provided multiple evidence support for the selected markers, enhancing their credibility in practical applications.
[0015] Through the above comprehensive analysis and mutual verification, the selected lncRNA and miRNA markers not only have higher accuracy and reliability in the diagnosis of thyroid cancer, but also provide a solid foundation for future clinical applications. This scheme effectively reduces false positive and false negative results and improves the clinical practicality and operability of the diagnostic model. By identifying specific lncRNA and miRNA markers, more personalized treatment plans can be provided for thyroid cancer patients, improving treatment effects and patients' quality of life.
[0016] As an implementable preferred solution, sequencing of lncRNA and miRNA in plasma exosomes includes the following: lncRNA and miRNA were extracted from plasma exosome samples; the extracted lncRNA and miRNA samples were reverse transcribed to generate cDNA. The cDNA was then sequenced using high-throughput sequencing technology to obtain the original sequencing data.
[0017] As an implementable preferred solution, identifying differentially expressed lncRNA and miRNA also includes the following: The expression levels of the preprocessed sequencing data were calculated to obtain the expression matrix of lncRNA and miRNA; the lncRNA and miRNA expression data of thyroid cancer patients and normal subjects were differentially expressed to identify the differentially expressed lncRNA and miRNA; the differential expression results were visualized using heat maps, volcano maps, cluster analysis or Venn diagrams.
[0018] As an implementable preferred solution, differentially expressed genes were screened by DEseq2; DEseq2 uses a negative binomial distribution model for differential analysis, and the formula for negative binomial distribution is:
[0019] Among them, k represents the number of reads of a specific gene observed under given conditions under the definition based on the number of failures, which is equivalent to the expression level of a gene in a certain sample; r is related to the variability of gene expression and is used to measure the magnitude of variability between samples; p refers to the probability that a specific gene is correctly assigned to the gene in a test.
[0020] As a feasible preferred solution, a polygenic regression model is established, including the following contents: The screened differentially expressed genes were used as input features, and the occurrence and development of thyroid cancer were used as output targets. Extract the expression levels of differentially expressed genes as input items of the Lasso regression model; The formula for constructing the objective function of the Lasso regression model is:
[0021] in, is the negative log-likelihood of observing the ith input, is the elastic-net regularization term, is the learning variable, representing the regression coefficient of each feature, =1 specifies the use of Lasso regression, and introduces w as the weight coefficient. The formula is:
[0022] in, is a non-negative regularization parameter that controls the strength of the regularization term.
[0023] As an implementable optimization scheme, draw the regularization path diagram and cross-validation curve of the model to assist in selection The regularization path diagram shows the change of the coefficient of each variable under different regularization degrees; the cross-validation curve shows the AUC value of the model under different λ values, and the cross-validation error under different values is calculated to select the value with the minimum error.
[0024] As an implementable optimization scheme, the first batch of data is predicted using the trained LASSO regression model to evaluate the prediction performance and accuracy of the model, including the following: Draw the ROC curve to show the relationship between the true positive rate and the false positive rate by comparing the predicted results with the true category of the sample, test the prediction effect of the model on different sample groups, and calculate the AUC score; Evaluation accuracy is calculated as follows:
[0025] The sensitivity of the evaluation is calculated as follows:
[0026] Among them, TP represents the number of samples correctly predicted as positive; TN represents the number of samples correctly predicted as negative; FP represents the number of samples incorrectly predicted as positive; FN represents the number of samples incorrectly predicted as negative. Draw a box plot of thyroid cancer scores to show the distribution of prediction results and further verify the reliability and stability of the model; Machine learning methods were used to perform stratified analysis on data from different groups to further optimize and verify the predictive performance and reliability of the model.
[0027] As a feasible preferred solution, bioinformatics analysis is performed, including the following contents: Perform differential expression analysis to identify and analyze differences in lncRNA and miRNA expression levels under different conditions; Predict the possible target genes of differentially expressed miRNAs and lncRNAs, including co-localization and co-expression predictions; Functional enrichment analysis was performed on the predicted lncRNA and miRNA target genes; Through literature search and biological analysis, the literature information was compared and organized with the results of functional analysis to analyze the functional categories, pathway participation, and biological significance of miRNA target genes related to thyroid cancer.
[0028] As an implementable preferred solution, the results of the polygenic regression model are cross-validated with the results of the bioinformatics analysis, including the following: Test whether the lncRNA and miRNA screened by LASSO also have significant expression changes in differential expression analysis, and whether they have significant biological functions in functional enrichment analysis; integrate the analysis results of lncRNA and miRNA, and verify their relationships and biological functions in the regulatory network by constructing a joint network.
[0029] As an implementable preferred solution, it also includes: The established LASSO regression model was verified and optimized using newly added batches of sequencing data from thyroid cancer patients and normal subjects. The predictive performance of the model was evaluated, the ROC curve and AUC score were calculated, and the model parameters were further optimized based on the verification results. BRIEF DESCRIPTION OF THE DRAWINGS
[0030] Figure 1 This is a logical schematic diagram of a method for screening combined diagnostic markers for thyroid cancer based on plasma exosomal lncRNA and miRNA.
[0031] Figure 2 This is the LASSO regularization path diagram.
[0032] Figure 3 This is a cross validation curve graph.
[0033] Figure 4 To construct a logical schematic diagram of the LncRNA-miRNA-mRNA regulatory network.
[0034] Figure 5 Schematic diagram of the structure of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION
[0035] In order to make the technical solution and advantages of the present application clearer, the technical solution of the present invention will be further described in detail below in conjunction with the accompanying drawings. It will be understood that the specific embodiments described herein are only partial embodiments of the present invention, which are only used to explain the present application, rather than to limit the present application. It should be noted that the technical features or combinations of technical features described in the following embodiments should not be considered isolated, and they can be combined with each other to achieve better technical effects. The same reference numerals appearing in the drawings of the following embodiments represent the same features or components, which can be applied to different embodiments.
[0036] In addition, unless otherwise defined, the technical terms or scientific terms used in the description of the present invention should have the common meanings understood by those skilled in the art in the art to which the present invention belongs.
[0037] The present invention is further described in detail below in conjunction with the accompanying drawings: Reference numerals: electronic device 500 , processor 501 , communication interface 502 , memory 503 , bus 504 .
[0038] Reference Figure 1 , a method for screening combined diagnostic markers for thyroid cancer based on plasma exosomal lncRNA and miRNA, comprising the following steps.
[0039] Step S100, collecting samples and performing processing and sequencing, including: Step S101, collect plasma samples from thyroid cancer patients and normal subjects to ensure that the quantity and quality of the samples meet the requirements of high-throughput sequencing. The collection of samples should follow standard clinical operating procedures to ensure that the quality and quantity of the samples meet the requirements of subsequent experiments.
[0040] Step S102, extracting exosomes from the sample. This includes centrifugation, filtration, washing and other steps to obtain a pure exosome sample. During the extraction process, care should be taken to avoid the rupture of the exosomes and the degradation of the RNA to ensure that the quality of the extracted RNA meets the sequencing requirements.
[0041] Exosomes are small vesicles secreted by cells that carry cell-specific biological molecules, including various RNAs. RNA extracted from exosomes can better reflect the state and function of the cells from which they come.
[0042] Exosomes have a double-layer membrane structure that can protect RNA from the action of RNases and have better stability and integrity, while traditional technical solutions use ENA in plasma which is easily degraded by RNases.
[0043] Step S103, extracting lncRNA and miRNA from the plasma exosome sample using an RNA extraction reagent, including lysis, centrifugation, precipitation, washing and other steps to obtain a high-quality RNA sample. During the extraction process, care should be taken to avoid RNA degradation and contamination to ensure that the extracted RNA concentration and purity meet the sequencing requirements.
[0044] The RNA extracted directly from plasma by traditional technical solutions may contain a large amount of plasma components, such as RNA produced by blood cell lysis, which may interfere with the detection of target RNA. However, the present technical solution extracts RNA from exosomes to reduce such interference.
[0045] At the same time, RNA molecules carried by exosomes may serve as biomarkers for certain diseases. By analyzing these RNAs, we can understand information about the progression, diagnosis or treatment effects of the disease.
[0046] Step S104, reverse transcribe the extracted lncRNA and miRNA samples to generate cDNA. Then use high-throughput sequencing technology to sequence the cDNA to obtain original sequencing data. During the sequencing process, attention should be paid to the depth and coverage of sequencing to ensure the accuracy and reliability of the sequencing data.
[0047] Step S200, constructing a thyroid cancer diagnosis model, including: Step S201, using bioinformatics tools to pre-process the original sequencing data, including removing low-quality sequences, removing adapter sequences, removing duplicate sequences, etc., to obtain high-quality sequencing data.
[0048] Step S202: Use bioinformatics tools to calculate the expression of the pre-processed sequencing data to obtain the expression matrix of lncRNA and miRNA. The expression calculation can be performed using methods such as FPKM (Fragments Per Kilobase of transcript per Million mapped reads) or TPM (Transcripts Per Million).
[0049] Step S203, using bioinformatics tools such as DESeq2 to perform differential expression analysis on the lncRNA and miRNA expression data of thyroid cancer patients and normal subjects, and identifying differentially expressed lncRNA and miRNA.
[0050] The miRNA and IncRNA obtained by sequencing were screened for differentially expressed genes through DEseq2. DEseq2 uses the negative binomial distribution model for differential analysis, and the formula for negative binomial distribution is:
[0051] Among them, k represents the number of reads of a specific gene observed under given conditions under the definition based on the number of failures, which is equivalent to the expression level of a gene in a certain sample; r is related to the variability of gene expression and is used to measure the magnitude of variability between samples; p refers to the probability that a specific gene is correctly assigned to the gene in a single trial (i.e., one read).
[0052] The differentially expressed genes screened by DEseq2 were further filtered to screen out the significantly expressed differentially expressed genes (p < 0.05).
[0053] Optionally, the differential expression results can be visualized using methods such as heat maps, volcano maps, cluster analysis, and Venn diagrams to intuitively present the expression differences of lncRNA and miRNA between different samples.
[0054] Step S204, constructing a LASSO (Least Absolute Shrinkage and Selection Operator) regression model, including: Step S204-1, using the LASSO algorithm to perform variable selection on the screened differentially expressed lncRNA and miRNA to reduce redundant variables and improve the generalization ability of the model.
[0055] Step S204-2, establish a multi-gene regression model to screen the core differential genes of miRNA / IncRNA, use the screened variables as input features, and the occurrence and development of thyroid cancer as output targets, The expression levels of significantly differentially expressed genes were extracted as input items of the Lasso regression model.
[0056] The core of the Lasso regression model lies in the shrinkage characteristic of the weight coefficient in the regularization process. By introducing the penalty term to control the complexity of the model, the features that contribute less to the prediction are automatically excluded, that is, the core features are selected. The formula for constructing the objective function is:
[0057] in, is the negative log-likelihood of observing the ith input, is the elastic-net regularization term, is the learning variable, representing the regression coefficient of each feature, =1 specifies the use of Lasso regression, and introduces w as the weight coefficient. The formula is:
[0058] in, is a non-negative regularization parameter that controls the strength of the regularization term.
[0059] The operation steps are: Initialize the weight coefficients to a zero vector.
[0060] Iteration process, in each iteration, each weight coefficient is updated in turn, while other weight coefficients remain unchanged. The result of the previous iteration is used as a constant for the next calculation.
[0061] Update the weight coefficients, add a new feature vector in each iteration, and adjust the weight coefficients for the next iteration.
[0062] Determine convergence and iteration termination. If the change of all weight coefficients is very small, or the preset maximum number of iterations is reached, stop the iteration.
[0063] Choose the right The value is one of the key points of Lasso regression.
[0064] Draw the regularization path diagram of the model, refer to Figure 2 , which shows the changes of each coefficient under different λ (regularization parameter) values, and shows the changes of the coefficient of each variable under different regularization degrees, so as to assist in selecting the best λ value to balance the complexity and performance of the model.
[0065] Draw a cross validation curve, refer to Figure 3 , using 10-fold cross validation, the AUC (Area Under Curve) value of the model under different λ (regularization parameter) values is displayed, the cross validation error under different values is calculated, and the value with the smallest error is selected. This helps to select the appropriate λ value to balance the complexity and performance of the model, thereby obtaining the best model prediction effect.
[0066] After establishing the polygenic regression model using the LASSO algorithm, the model parameters and coefficients were extracted and the The values are extracted, the features of non-zero coefficients are extracted, and the expression profiles of these characteristic genes are derived to determine the final marker lncRNA / miRNA. These parameters and coefficients will be used to explain the model and serve as an important basis for predicting the occurrence and development of thyroid cancer.
[0067] Step S204-3, using the trained LASSO regression model to predict the first batch of data, by predicting the first batch of data and evaluating the prediction performance and accuracy of the model, the constructed polygenic regression model is evaluated.
[0068] ROC curve and AUC score were used for evaluation.
[0069] Draw the ROC curve, compare the predicted results with the true category of the sample to show the relationship between the true positive rate and the false positive rate, and test the prediction effect of the model on different sample groups. Use the pROC package of R language to draw and calculate the AUC score. The AUC value ranges from 0 to 1. The closer the result is to 1, the better the classification effect. The calculation method is: estimate the area under the ROC curve by numerical integration.
[0070] Evaluation accuracy is calculated as follows:
[0071] The sensitivity of the evaluation is calculated as follows:
[0072] Among them, TP represents the number of samples correctly predicted as positive; TN represents the number of samples correctly predicted as negative; FP represents the number of samples incorrectly predicted as positive; and FN represents the number of samples incorrectly predicted as negative.
[0073] At the same time, the distribution of the prediction results was displayed by drawing a box plot of thyroid cancer scores to further verify the reliability and stability of the model.
[0074] Step S205, based on the regression model results and sample data, the data is stratified and analyzed to identify specific thyroid cancer diagnostic markers in different subgroups (such as age, gender, etc.). In this embodiment, the age is currently considered to be stratified by 35 years old, and the gender is divided into male and female groups.
[0075] Machine learning methods are used to perform stratified analysis on data from different groups to further optimize and verify the predictive performance and reliability of the model. AUC scores and ROC curves for different stratified groups are generated, and the analysis results of important variables and model parameters are output.
[0076] Step S206, investigating the marker lncRNA / miRNA screened out in the multi-gene regression model to understand its function and its role in the occurrence and development of thyroid cancer, including: Step S206-1, collecting marker lncRNA / miRNA information, extracting the marker lncRNA / miRNA list screened out in the multi-gene regression model; confirming the name and sequence information of the marker.
[0077] Step S206-2, query the bioinformatics database, use the database (such as NCBI Gene, miRBase, lncRNA database) to query the annotation information and known functions of each marker lncRNA / miRNA, and obtain their participation in different biological processes and their expression and function in cancer (especially thyroid cancer).
[0078] Step S206-3, literature search, using academic databases such as PubMed and Google Scholar to search for the latest research literature related to each marker lncRNA / miRNA. Collect the relevant research progress and experimental results of the marker in thyroid cancer to understand its biological significance.
[0079] Step S206-4, functional verification, predicting the targets of marker lncRNA / miRNA through target prediction tools (such as TargetScan, miRanda, lncTar, etc.); constructing the lncRNA-miRNA-mRNA interaction network using interaction network analysis tools (such as STRING, Cytoscape); analyzing the expression correlation between marker lncRNA and marker miRNA, and verifying their role in thyroid cancer.
[0080] Step S300, performing bioinformatics analysis, including: Step S301, miRNA-seq bioinformatics analysis, including: Step S301 - 1 , differential expression analysis, performs differential expression analysis on miRNA, identifies and analyzes the differences in miRNA expression levels under different conditions, and generates differential expression analysis results.
[0081] Step S301-2, target gene prediction, using software such as TargetScan, miRanda, miRDB, etc. to predict genes that may be targeted by differentially expressed miRNAs, and further understand the role of miRNAs in gene regulation.
[0082] Step S301-3, functional enrichment analysis, performs functional enrichment analysis (such as GO, KEGG, etc.) on the predicted miRNA target genes to understand the functional characteristics of these genes in biological processes.
[0083] Step S301-4, through literature search and biological analysis, the literature information is compared and sorted with the results of functional analysis, and the functional categories, pathway participation, and biological significance of miRNA target genes related to thyroid cancer are analyzed.
[0084] Step S302, lncRNA-seq bioinformatics analysis, including: Step S302-1, differential expression analysis, performing differential expression analysis on lncRNA to identify and analyze differences in lncRNA expression levels under different conditions.
[0085] Step S302-2, target gene prediction, using software such as lncTar, LncBase, miRcode, etc. to predict genes that may be targeted by differentially expressed lncRNAs, including co-localization and co-expression predictions.
[0086] Step S302-3, functional enrichment analysis, performs functional enrichment analysis on the predicted lncRNA target genes to understand the functional characteristics of these genes in biological processes.
[0087] Step S302-4, through literature search and biological analysis, the literature information is compared and sorted with the results of functional analysis, and the functional categories, pathway participation, and biological significance related to thyroid cancer of lncRNA target genes are analyzed.
[0088] Step S400, refer to Figure 4 , construct the LncRNA-miRNA-mRNA regulatory network (ceRNA network), including: Step S401, network construction, includes: Step S401-1, using algorithms such as miRcode to predict the interaction relationship between lncRNA and miRNA.
[0089] Step S401-2, verifying the target mRNA of miRNA from databases such as TarBase.
[0090] Step S401-3, screen out significant ceRNA relationships by resampling, and construct a lncRNA-miRNA-mRNA regulatory network (ceRNA network, Competitive Endogenous RNA, competitive endogenous RNA network).
[0091] Step S402, network analysis, includes: Step S402-1, using software such as Cytoscape to visualize the ceRNA network to intuitively present the interaction relationship between lncRNA, miRNA and mRNA.
[0092] Step S402-2, perform functional analysis on key nodes (such as lncRNA, miRNA, etc.) in the ceRNA network to understand their potential role and significance in the occurrence and development of thyroid cancer.
[0093] Step S500, performing mutual verification and integrated analysis of results, including: Step S501, cross-validate the results of the polygenic regression model with the results of the basic bioinformatics analysis to test whether the lncRNA and miRNA screened by LASSO also have significant expression changes in the differential expression analysis, and whether they have significant biological functions in the functional enrichment analysis.
[0094] Step S502, integrating the analysis results of lncRNA and miRNA, and verifying their mutual relationships and biological functions in the regulatory network by constructing a joint network.
[0095] Through the above steps, the results of basic bioinformatics analysis and multi-gene regression model can be verified with each other, ensuring the reliability and practicality of the screened lncRNA and miRNA in the diagnosis of thyroid cancer. The redundancy of analysis results can be effectively reduced, and the accuracy and clinical application value of markers can be improved.
[0096] Step S600: Use the newly added batches of sequencing data of thyroid cancer patients and normal subjects to verify and optimize the established LASSO regression model. Evaluate the prediction performance of the model, calculate the ROC curve and AUC score, and further optimize the model parameters according to the verification results to improve the generalization ability of the model.
[0097] The disclosed embodiments also provide a system for screening a combined diagnostic marker for thyroid cancer based on plasma exosome lncRNA and miRNA, which utilizes the above-mentioned method for screening a combined diagnostic marker for thyroid cancer based on plasma exosome lncRNA and miRNA.
[0098] The disclosed embodiments also provide a storage medium, wherein a computer program is stored in the storage medium. When the computer program is executed by a processor, all steps of the above-mentioned method for screening a combined diagnostic marker for thyroid cancer based on plasma exosome lncRNA and miRNA can be implemented.
[0099] Those of ordinary skill in the art can understand that the implementation of all or part of the process in a method for screening a combined diagnostic marker for thyroid cancer based on plasma exosome lncRNA and miRNA can be completed by instructing the relevant hardware through a computer program, and the program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include a process of each embodiment of a combined diagnostic marker screening method for thyroid cancer based on plasma exosome lncRNA and miRNA. Among them, any reference to memory, storage, database or other medium used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.
[0100] Reference Figure 5The electronic device 500 includes: at least one processor 501, at least one communication interface 502, at least one memory 503 and at least one bus 504. The bus 504 is used to realize the connection and communication between these components, the communication interface 502 is used to communicate with other node devices for signaling or data, and the memory 503 stores machine-readable instructions executable by the processor 501. When the electronic device 500 is running, the processor 501 communicates with the memory 503 through the bus 504, and when the machine-readable instructions are called by the processor 501, the steps of the above-mentioned method for optical detection of appearance defects of a flexible production line are executed.
[0101] The above contents are only embodiments of the present invention. The common sense such as the known specific structures and characteristics in the scheme is not described in detail here. The ordinary technicians in the relevant field know all the common technical knowledge in the technical field to which the invention belongs before the application date or the priority date, can obtain all the existing technologies in the field, and have the ability to apply the conventional experimental means before that date. The ordinary technicians in the relevant field can improve and implement this scheme in combination with their own abilities under the enlightenment given by this application. Some typical known structures or known methods should not become obstacles for the ordinary technicians in the relevant field to implement this application. It should be pointed out that for the technicians in this field, without departing from the structure of the present invention, several deformations and improvements can be made, which should also be regarded as the protection scope of the present invention, which will not affect the effect of the implementation of the present invention and the practicality of the patent. The protection scope required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to explain the content of the claims.
Claims
1. A method for screening combined diagnostic markers for thyroid cancer based on plasma exosome lncRNA and miRNA, characterized in that: include: Sequencing of lncRNA and miRNA in plasma exosomes; Construct a diagnostic model for thyroid cancer to identify differentially expressed lncRNA and miRNA; use the LASSO algorithm to select variables for the screened differentially expressed lncRNA and miRNA; establish a multi-gene regression model to screen for iconic miRNA / IncRNA; Conduct bioinformatics analysis to find key lncRNAs and miRNAs and their biological functions; Construct ceRNA regulatory network, predict the interaction between lncRNA and miRNA, verify the targeted mRNA of miRNA, screen out significant ceRNA relationships, and perform functional analysis on key lncRNA and miRNA in ceRNA network; The results of the polygenic regression model were cross-validated with those of the bioinformatics analysis.
2. The method for screening combined diagnostic markers for thyroid cancer based on plasma exosome lncRNA and miRNA according to claim 1, characterized in that: Plasma exosomes were sequenced for lncRNA and miRNA, including the following: lncRNA and miRNA were extracted from plasma exosome samples; the extracted lncRNA and miRNA samples were reverse transcribed to generate cDNA. The cDNA was then sequenced using high-throughput sequencing technology to obtain the original sequencing data.
3. The method for screening combined diagnostic markers for thyroid cancer based on plasma exosome lncRNA and miRNA according to claim 1, characterized in that: Identification of differentially expressed lncRNAs and miRNAs also includes the following: The expression level of the preprocessed sequencing data was calculated to obtain the expression level matrix of lncRNA and miRNA; Differential expression analysis was performed on the lncRNA and miRNA expression data of thyroid cancer patients and normal subjects to identify differentially expressed lncRNA and miRNA; heat maps, volcano maps, cluster analysis or Venn diagrams were used to visualize the differential expression results.
4. The method for screening combined diagnostic markers for thyroid cancer based on plasma exosome lncRNA and miRNA according to claim 3, characterized in that: Differentially expressed genes were screened out by DEseq2; DEseq2 uses a negative binomial distribution model for differential analysis. The formula for negative binomial distribution is: Among them, k represents the number of reads of a specific gene observed under given conditions under the definition based on the number of failures, which is equivalent to the expression level of a gene in a certain sample; r is related to the variability of gene expression and is used to measure the magnitude of variability between samples; p refers to the probability that a specific gene is correctly assigned to the gene in a test.
5. The method for screening combined diagnostic markers for thyroid cancer based on plasma exosome lncRNA and miRNA according to claim 4, characterized in that: A polygenic regression model was established, including the following: The screened differentially expressed genes were used as input features, and the occurrence and development of thyroid cancer were used as output targets. Extract the expression levels of differentially expressed genes as input items of the Lasso regression model; The formula for constructing the objective function of the Lasso regression model is: in, is the negative log-likelihood of observing the ith input, is the elastic-net regularization term, is the learning variable, representing the regression coefficient of each feature, =1 specifies the use of Lasso regression, and introduces w as the weight coefficient. The formula is: in, is a non-negative regularization parameter that controls the strength of the regularization term.
6. The method for screening combined diagnostic markers for thyroid cancer based on plasma exosome lncRNA and miRNA according to claim 5, characterized in that: Draw the regularization path diagram and cross-validation curve of the model to assist in selection The regularization path diagram shows the change of the coefficient of each variable under different regularization degrees; the cross-validation curve shows the AUC value of the model under different λ values, and the cross-validation error under different values is calculated to select the value with the minimum error.
7. The method for screening combined diagnostic markers for thyroid cancer based on plasma exosome lncRNA and miRNA according to claim 6, characterized in that: Use the trained LASSO regression model to predict the first batch of data and evaluate the model's prediction performance and accuracy, including the following: Draw the ROC curve to show the relationship between the true positive rate and the false positive rate by comparing the predicted results with the true category of the sample, test the prediction effect of the model on different sample groups, and calculate the AUC score; Evaluation accuracy is calculated as follows: The sensitivity of the evaluation is calculated as follows: Among them, TP represents the number of samples correctly predicted as positive; TN represents the number of samples correctly predicted as negative; FP represents the number of samples incorrectly predicted as positive; FN represents the number of samples incorrectly predicted as negative. Draw a box plot of thyroid cancer scores to show the distribution of prediction results and further verify the reliability and stability of the model; Machine learning methods were used to perform stratified analysis on data from different groups to further optimize and verify the predictive performance and reliability of the model.
8. The method for screening combined diagnostic markers for thyroid cancer based on plasma exosome lncRNA and miRNA according to claim 1, characterized in that: Perform bioinformatics analysis including the following: Perform differential expression analysis to identify and analyze differences in lncRNA and miRNA expression levels under different conditions; Predict the possible target genes of differentially expressed miRNAs and lncRNAs, including co-localization and co-expression predictions; Functional enrichment analysis was performed on the predicted lncRNA and miRNA target genes; Through literature search and biological analysis, the literature information was compared and organized with the results of functional analysis to analyze the functional categories, pathway participation, and biological significance of miRNA target genes related to thyroid cancer.
9. The method for screening combined diagnostic markers for thyroid cancer based on plasma exosome lncRNA and miRNA according to claim 1, characterized in that: The results of the polygenic regression model were cross-validated with the results of the bioinformatics analysis, including the following: Test whether the lncRNA and miRNA screened by LASSO also have significant expression changes in differential expression analysis, and whether they have significant biological functions in functional enrichment analysis; integrate the analysis results of lncRNA and miRNA, and verify their relationships and biological functions in the regulatory network by constructing a joint network.
10. The method for screening combined diagnostic markers for thyroid cancer based on plasma exosome lncRNA and miRNA according to claim 1, characterized in that: Also includes: The established LASSO regression model was verified and optimized using newly added batches of sequencing data from thyroid cancer patients and normal subjects. The predictive performance of the model was evaluated, the ROC curve and AUC score were calculated, and the model parameters were further optimized based on the verification results.