Methylation markers for detecting benign and malignant pulmonary nodules and their applications

Abstract

The present invention provides a methylation marker for detecting benign and malignant pulmonary nodules and its application. The methylation marker includes a fragment having a sequence of at least one of the sequences shown in Seq ID No: 1 to Seq ID No: 43 or its fully complementary sequence, or a fragment including the sequence and its upstream and downstream 1kb sequence, a DNA methylation site in the region of the sequence, a variant having at least 90% sequence identity with the sequence and having the same methylation site as the sequence, or a DNA methylation haplotype covered in the region of the sequence and the abundance of the DNA methylation haplotype. Based on the methylation sequencing data of benign and malignant samples of pulmonary nodules, the present invention screens out markers with obvious differences in methylation levels in benign pulmonary nodules and malignant pulmonary nodules, which can be used to effectively distinguish the benign and malignant nature of pulmonary nodules, and also uses these markers to establish a risk prediction and assessment model for benign and malignant pulmonary nodules / lung cancer.

Description

Technical Field

[0001] The present invention relates to the field of biomedical technology, and in particular to a methylation marker for detecting benign or malignant pulmonary nodules and an application thereof. Background Art

[0002] Low-dose computed tomography (LDCT) has been widely used as the most effective method for detecting early lung cancer. The landmark National Lung Screening Trial (NLST) showed that LDCT has a high sensitivity for early small malignant pulmonary nodules (<3 cm) when screening high-risk populations for lung cancer, and can reduce lung cancer mortality by 20% compared with single-view posteroanterior chest radiographs. (National Lung Screening Trial Research et al, 2011). However, LDCT has a high false positive rate (96.4% in the NLST study) by detecting benign pulmonary nodules (National Lung Screening Trial Research et al, 2011).

[0003] In order to minimize the misdiagnosis of benign nodules as malignant, it is necessary to accurately stratify the malignant risk of detected pulmonary nodules. Recently, liquid biopsy has been proposed as an alternative method for differential diagnosis of pulmonary nodules. It samples and analyzes non-solid biological tissues (mainly blood) to identify and quantify lung cancer-derived biomarkers, such as circulating tumor cells, nucleic acids or proteins, to assess the presence and / or status of tumor detection, diagnosis, prognosis and monitoring (Wan, JCM, et al. Liquid biopsies come of age: Towards implementation of circulating tumour DNA. Nature Reviews Cancer, 2017). Current studies have confirmed that body fluids (blood, urine, bronchoalveolar lavage fluid, etc.) contain a large amount of free DNA (cfDNA), which has tissue-specific DNA methylation characteristics and can be used for tissue tracing and cancer detection (Lo YMD et al., 2021). cfDNA methylation detection has now been applied to the detection of various cancers such as colorectal cancer, liver cancer, etc., and can sensitively detect early cancers with high specificity. Some literature has found that cfDNA methylation can be used in the early diagnosis and screening of lung cancer (AUC = 0.90) (Liang N et al., 2021), but it performs poorly in the diagnosis of benign and malignant lung nodules (AUC = 0.72-0.81) (Liang W et al., 2021). Therefore, there is an urgent need to find liquid biopsy methylation biomarkers with higher stability and specificity for the diagnosis of benign and malignant lung nodules.

[0004] In view of this, the present invention is proposed. Summary of the invention

[0005] The purpose of the present invention is to provide a methylation marker for detecting benign and malignant pulmonary nodules and its application. The present invention can distinguish patients with benign pulmonary nodules and malignant pulmonary nodules by detecting the methylation level of the methylation marker in plasma, thereby achieving the purpose of non-invasive and precise diagnosis of pulmonary nodules with higher accuracy and lower cost.

[0006] In order to achieve the above-mentioned purpose of the present invention, the following technical solutions are particularly adopted:

[0007] In one aspect, the present invention provides methylation markers for detecting benign and malignant pulmonary nodules, wherein the methylation markers include the following markers:

[0008] (a1), a fragment having a sequence of at least one of the sequences shown in Seq ID No: 1 to Seq ID No: 43 or their completely complementary sequences;

[0009] (a2), a fragment including the sequence described in (a1) and 1 kb upstream and downstream sequences;

[0010] DNA methylation sites in the region of the sequence shown in (a3), (a1) or (a2);

[0011] (a4), a variant having at least 90% sequence identity with the sequence of (a1) or (a2) and having the same methylation site as (a1) or (a2); or

[0012] DNA methylation haplotypes covered in the region of the sequence shown in (a5), (a1) or (a2) and the abundance of the DNA methylation haplotypes.

[0013] In one embodiment, the benign and malignant methylation markers of pulmonary nodules include one or more or all of the following sequences, or one or more of the complementary sequences of the following sequences: Seq ID No: 1, Seq ID No: 2, Seq ID No: 3, Seq ID No: 4, Seq ID No: 5, Seq ID No: 6, Seq ID No: 7, Seq ID No: 8, Seq ID No: 9, Seq ID No: 10, Seq ID No: 11, Seq ID No: 12, Seq ID No: 13, Seq ID No: 14, Seq ID No: 15, Seq ID No: 16, Seq ID No: 17, Seq ID No: 18, Seq ID No: 19, Seq ID No: 20, Seq ID No: 21, Seq ID No: 22, Seq ID No: 23, Seq ID No: 24, Seq ID No: 25, Seq ID No: 26, Seq ID No: 27, Seq ID No: 28 No: 28, Seq ID No: 29, Seq ID No: 30, Seq ID No: 31, Seq ID No: 32, Seq ID No: 33, Seq ID No: 34, Seq ID No: 35, Seq ID No: 36, Seq ID No: 37, Seq ID No: 38, Seq ID No: 39, Seq ID No: 40, Seq ID No: 41, Seq ID No: 42, Seq ID No: 43. The positions of the sequences in the genome are: chr1:29586580:29586780, chr1:44015730:44015930, chr1:110334699:110334899, chr1:111813478:111813678, chr1:119526811:119527011, chr1:156092393:156092593, chr10:123923943:123924143, chr10: 131223428:131223628、chr10:131223438:131223638、chr10:134598574:134598774、chr11:63687159:63687359、c hr12:108169405:108169605、chr13:25320480:25320680、chr15:29395897:29396097、chr16:22825689:22825889、<h2 style=";text-align:left;direction:ltr">chr16:22825962:22826162、chr17:62775624:62775824、chr17:73483989:73484189、chr17:7738 6068:77386268、chr17:80535356:80535556、chr2:20869406:20869606、chr2:31805332:31805532 、chr2:45227849:45228049、chr2:66671569:66671769、chr2:106690309:106690509、chr2:17183 0634:171830834、chr2:177024290:177024490、chr2:198650929:198651129、chr22:51016376:510 16576、chr3:52828464:52828664、chr3:181444089:181444289、chr4:190940255:190940455、chr 5:1876269:1876469、chr5:2749400:2749600、chr5:134827504:134827704、chr6:19691753:19691 953, chr6:36253042:36253242, chr6:106429583:106429783, chr7:29605883:29606083, chr7:150655260:150655460, chr8:687377:687577, chr8:20375580:20375780, chr8:99986572:99986772.<h2 style=";text-align:left;direction:ltr"> <h2 style=";text-align:left;direction:ltr">

[0014] Preferably, the benign and malignant methylation markers of pulmonary nodules include Seq ID No: 33 and one or more or all of the following sequences: Seq ID No: 1, Seq ID No: 2, Seq ID No: 3, Seq ID No: 4, Seq ID No: 5, Seq ID No: 6, Seq ID No: 7, Seq ID No: 8, Seq ID No: 9, Seq ID No: 10, Seq ID No: 11, Seq ID No: 12, Seq ID No: 13, Seq ID No: 14, Seq ID No: 15, Seq ID No: 16, Seq ID No: 17, Seq ID No: 18, Seq ID No: 19, Seq ID No: 20, Seq ID No: 21, Seq ID No: 22, Seq ID No: 23, Seq ID No: 24, Seq ID No: 25, Seq ID No: 26, Seq ID No: 27, Seq ID No: 28, Seq ID No: 29, No: 29, Seq ID No: 30, Seq ID No: 31, Seq ID No: 32, Seq ID No: 34, Seq ID No: 35, Seq ID No: 36, Seq ID No: 37, Seq ID No: 38, Seq ID No: 39, Seq ID No: 40, Seq ID No: 41, Seq ID No: 42, Seq ID No: 43.

[0015] More preferably, the benign and malignant methylation markers of pulmonary nodules include the sequences shown in Seq ID No: 1 to Seq ID No: 43 or their completely complementary sequences.

[0016] In one embodiment, the methylation markers of benign and malignant pulmonary nodules include fragments including the sequences shown in Seq ID No: 1 to Seq ID No: 43 and 1 kb sequences upstream and downstream thereof; or DNA methylation sites in the region of the shown sequences; or variants having at least 90% sequence identity with the shown sequences and the same methylation sites as the shown sequences; or DNA methylation haplotypes covered in the region of the shown sequences and the abundance of the DNA methylation haplotypes.

[0017] Preferably, (a4) is a variant having at least 95%, 98%, 99% or 99.9% sequence identity with the sequence of (a1) or (a2) and having the same methylation sites as (a1) or (a2).

[0018] In the present invention, the methylation biomarker uses the methylation state / level of the region / fragment / methylation site containing the CpG island involved in the present invention as the basis for distinguishing patients with benign pulmonary nodules and malignant pulmonary nodules. The methylation level of the methylation marker can be obtained by a variety of detection methods known in the art, including but not limited to obtaining the methylation level by second-generation sequencing.

[0019] The numbers of the positions of the sequences in the genome in the present invention correspond to the positions on the UCSC (http: / / hgdownload.soe.ucsc.edu / goldenPath / hg19 / bigZips / hg19.fa.gz) HG19 genome.

[0020] The methylation biomarkers in the present invention can be used alone or in combination as lung cancer-related methylation molecular markers for detecting or assisting in detecting or identifying benign and malignant lung nodules and / or lung cancer.

[0021] On the other hand, the present invention also provides reagents for detecting the methylation degree of the aforementioned methylation markers; for example, primers or probes for the aforementioned methylation markers, wherein the primers use the nucleotide sequence where the methylation marker is located as the target sequence and are used for specific amplification of the target sequence; and the probes specifically capture the nucleotide sequence where the methylation marker is located.

[0022] On the other hand, the present invention also provides a kit for detecting benign and malignant lung nodules, wherein the kit contains reagents for detecting the above-mentioned methylation markers of benign and malignant lung nodules.

[0023] On the other hand, the present invention also provides the use of a reagent for detecting the methylation level of the above-mentioned benign and malignant methylation marker for lung nodules in the preparation of a diagnostic kit for identifying benign and malignant lung nodules in a sample.

[0024] Further, the detection reagent also includes reagents used in any one of the PCR amplification method, fluorescent quantitative PCR method, digital PCR method, liquid phase chip method, second generation sequencing method, third generation sequencing method, bisulfite sequencing method, whole genome methylation sequencing method, methylation chip method or a combination thereof. For example, the detection reagent can be selected from the following: bisulfite and its derivatives, PCR buffer, polymerase, dNTP, primer, probe, methylation sensitive or insensitive restriction endonuclease, enzyme digestion buffer, fluorescent dye, fluorescence quencher, fluorescence reporter, exonuclease, alkaline phosphatase, internal standard, control, etc.

[0025] In one embodiment, the biomarker is obtained from a liquid sample, which is blood of a mammal and other liquid samples of biological origin, such as peripheral blood, serum, plasma, ascites, urine, cerebrospinal fluid, sputum, saliva, etc. The mammals include rats, mice and humans; preferably humans.

[0026] In one embodiment, the sample is a fine needle aspiration biopsy or plasma. The sample includes genomic DNA or cfDNA. In one embodiment, the sample preferably refers to plasma cfDNA. cfDNA refers to circulating free DNA or cell-free DNA, degraded DNA fragments released into the plasma.

[0027] In one embodiment, the present invention provides markers for screening benign and malignant pulmonary nodules based on cfDNA methylation.

[0028] In the present invention, "benign" and "malignant" of pulmonary nodules refer to the properties of pulmonary nodules. The pulmonary nodules include solid or partially solid or ground glass nodules, preferably partially solid or ground glass nodules. "Malignant" pulmonary nodules usually refer to those with cancerous changes.

[0029] On the other hand, the present invention also provides a method for constructing a prediction and evaluation model for benign and malignant pulmonary nodules, comprising the following steps:

[0030] (a) Collect benign and malignant lung nodule samples and divide them into training sets and test sets;

[0031] (b) Extract cfDNA from samples, build libraries, and sequence them;

[0032] (c) Perform methylation conversion processing on the sequences, compare the data, and calculate the proportion of discordant reads (PDR) value of the samples;

[0033] (d) constructing a feature matrix and an algorithm model for the sample data, and screening out methylation markers for benign and malignant pulmonary nodules based on the training set sample data;

[0034] (e) Use the test set sample data to verify the effect of the model;

[0035] (f) To determine the methylation markers that can be used to predict the benign and malignant nature of lung nodules.

[0036] In the present invention, 80 stable methylation markers were first screened out by comparing the Reduced Representative Bisulfite Sequencing (RRBS) data of benign and malignant nodule tissue samples, and then combined with the reported Panseer Panel (Chen et al., 2020) to form a panel containing 760 methylation markers. Then, the methylation sequencing data of sample plasma cfDNA was obtained by the Methyl-Titan (patent number: CN201910515830) method to obtain the methylation levels of the above 760 markers, and then the key methylation marker combinations were screened out to assist in the diagnosis of benign and malignant pulmonary nodules.

[0037] In one embodiment, the test set includes a validation set, a single-center test set, and a multi-center test set.

[0038] In step (d), the best combination of classifier and parameters was found by grid search method in the training set. First, logistic regression, support vector machine (SVM) and random forest models were pre-selected for classifier selection. For all features, a 3-fold cross-validation grid search method was used to select classifiers, and this process selected logistic regression as the optimal classifier. Then, grid search was performed for different combinations of modeling parameters and thresholds, and the features with the best classification ability were identified using the logistic regression model. Finally, 43 features were selected for modeling; these processes were implemented using python (version 3.6.13) and scikit-learn (version 0.24.2).

[0039] Furthermore, the algorithm model includes a machine learning model; the machine learning model includes any one of a principal component analysis model, a logistic regression analysis model, a nearest neighbor analysis model, a support vector machine and a neural network model.

[0040] Preferably, the machine learning model is a logistic regression analysis model.

[0041] On the other hand, the present invention also provides a risk assessment model for benign and malignant pulmonary nodules, the risk assessment model is obtained according to the above-mentioned construction method, and comprises:

[0042] A data acquisition module, at least used to acquire a sample data set;

[0043] A sequencing module, at least used to obtain sequencing data;

[0044] A data comparison module, at least used to compare the sequencing data with a reference sequence, and determine the methylation results of the markers in the sequencing data based on the comparison results;

[0045] The result determination module is at least used to calculate the prediction score threshold through statistical model analysis to determine whether the sample to be tested is a benign or malignant pulmonary nodule.

[0046] The risk assessment model may further include a methylation processing module, which is at least used to perform methylation processing to obtain data after methylation processing.

[0047] The present invention performs regression analysis on the methylation levels of the screened markers and the benign and malignant nodules to construct a regression model, thereby obtaining a risk assessment model or diagnosis model for benign and malignant pulmonary nodules. The methylation levels of the 43 methylation markers in the present invention are used to construct a machine learning model of logistic regression, which can be used to distinguish benign pulmonary nodules from malignant pulmonary nodules.

[0048] On the other hand, the present invention further provides an information data processing terminal, the information data processing terminal comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the following steps when executing the computer program:

[0049] (a) obtaining the methylation level of at least one region of the sequence from Seq ID No: 1 to Seq ID No: 43 in the sample to be tested; (b) calculating a score by constructing a logistic regression diagnostic model; (c) identifying the benign or malignant nature of the pulmonary nodule according to the score.

[0050] On the other hand, the present invention further provides a computer-readable storage medium, wherein a computer program is stored on the storage medium, and when the computer program is executed by a processor, the following steps are implemented:

[0051] (a) obtaining the methylation level of at least one region of the sequence from Seq ID No: 1 to Seq ID No: 43 in the sample to be tested; (b) calculating a score by constructing a logistic regression diagnostic model; and (c) identifying benign and malignant pulmonary nodules according to the score.

[0052] A method for detecting benign or malignant pulmonary nodules for non-disease diagnosis purposes, comprising the following steps:

[0053] S1: Detecting the methylation level of at least one region of the sequence Seq ID No: 1 to Seq ID No: 43 in the sample;

[0054] S2: The score is calculated based on the aforementioned risk assessment model for benign and malignant pulmonary nodules;

[0055] S3: Identify benign or malignant thyroid nodules based on the score.

[0056] Specifically, when the methylation level of the target sequence of the sample to be tested meets a certain threshold, it is identified as a malignant nodule. For example, for the sample to be tested, when the score is greater than the threshold, the result is positive, that is, a malignant nodule, otherwise it is negative, that is, a benign nodule.

[0057] The present invention has the following beneficial effects:

[0058] 1. This application and the present invention provide 43 methylation markers with a very high correlation with benign and malignant lung nodules based on high-throughput methylation sequencing of plasma cfDNA. These markers can effectively distinguish benign and malignant lung nodules with high sensitivity and specificity. These markers can be used to establish a benign and malignant lung nodule / lung cancer risk prediction and assessment model for distinguishing benign and malignant lung nodules;

[0059] 2. The present application also provides a risk prediction assessment model or a diagnostic model established based on the relationship between the methylation level of the marker and the benign and malignant nature of lung nodules and a specific construction method thereof. The prediction score of the model can be used to distinguish the benign and malignant nature of lung nodules. The model has the advantages of non-invasive detection, safe and convenient detection, high throughput, and high detection accuracy;

[0060] 3. Using the methylation markers obtained in this application and their risk prediction and assessment model to distinguish between benign and malignant lung nodules can effectively control the detection cost while achieving better detection performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0062] Figure 1 Predict the score distribution for the AllModel diagnostic model training set and validation set;

[0063] Figure 2 The ROC curve shows the effect of the AllModel diagnostic model on the training set and validation set;

[0064] Figure 3 The distribution of predicted scores for the AllModel diagnostic model in single-center and multi-center test sets;

[0065] Figure 4 The ROC curve shows the effect of the AllModel diagnostic model in single-center and multi-center test sets. DETAILED DESCRIPTION

[0066] In order to more clearly explain the overall concept of the present application, the following is described in detail in conjunction with the accompanying drawings of the specification by way of embodiment. In the following description, a large number of specific details are provided to provide a more thorough understanding of the present invention. However, it is obvious to those skilled in the art that the present invention can be implemented without one or more of these details. In other examples, in order to avoid confusion with the present invention, some technical features well known in the art are not described.

[0067] If no specific conditions are specified in the examples, the experiments were carried out according to conventional conditions or conditions recommended by the manufacturer.

[0068] Unless otherwise specified, in the following embodiments, the reagents or instruments used without indicating the manufacturer are all conventional products that can be purchased commercially.

[0069] In this application, by comparing the Reduced Representative Bisulfite Sequencing (RRBS) data of benign and malignant nodule tissue samples, 80 stable methylation markers were screened out, and then combined with the reported Panseer Panel (Chen et al., 2020) to form a panel that ultimately contains 760 methylation markers.

[0070] This application collects blood samples from patients with pulmonary nodules. The sample collection criteria are as follows: adult patients aged 40-80 years, pulmonary nodules found by LDCT screening (especially between 5 and 30 mm). Exclusion criteria include patients with pulmonary nodules accompanied by hilar or mediastinal lymphadenopathy or suspected intrapulmonary metastasis of lung cancer or other malignant tumors, as well as other situations that the researchers consider inappropriate for this study. The samples are divided into training sets (Training), validation sets (Validation), single-center test sets (Single-center Test) and multi-center test sets (Multi-center Test). The specific sample information is as follows (see Table 1):

[0071] Table 1 Sample information statistics

[0072]

[0073] This application uses the Methyl-Titan (patent number: CN201910515830) method to obtain methylation sequencing data of sample plasma cfDNA and screen out methylation markers. The specific screening process is as follows:

[0074] 1. Extraction of plasma cfDNA samples:

[0075] A 10 ml whole blood sample was collected from the patient using a Streck blood collection tube (Streck, 218962) and transported at ambient temperature within 12 hours after collection, and centrifuged to separate plasma in time (within 3 days). To separate plasma, the blood sample was first centrifuged at 1600 g for 10 minutes at 4°C, and then the supernatant was extracted and centrifuged again at 16000 g for another 10 minutes at 4°C. All cfDNA was then extracted using the HiPure Circulating DNA Midi Kit C (Magen).

[0076] 2. Sequencing and data preprocessing:

[0077] a) The library was sequenced with an Illumina Nextseq 500 sequencer for 150 bp paired-end sequencing, with a sequencing volume of no less than 5M.

[0078] b) Pear (v0.6.0) software merges the double-end sequencing data of the same fragment of 150bp sequencing from the Illumina Hiseq X10 / Nextseq 500 / Novaseq sequencer into one sequence with a minimum overlap length of 20bp and a minimum overlap length of 30bp after merging.

[0079] c) Trim_galore v 0.6.0 and cutadapt v1.8.1 software were used to remove the adapters from the merged sequencing data. The adapter sequence at the 5' end of the sequence was "AGATCGGAAGAGCAC", and the bases with sequencing quality values ​​less than 20 at both ends were removed.

[0080] 3. Sequencing data comparison

[0081] The reference genome data used in this article comes from the UCSC database (UCSC: HG19, http: / / hgdownload.soe.ucsc.edu / goldenPath / hg19 / bigZips / hg19.fa.gz).

[0082] a) First, HG19 was converted from cytosine to thymine (CT) and adenine to guanine (GA) using Bismark software, and the index was constructed for the converted genome using Bowtie2 software.

[0083] b) The pre-processed data are also transformed into CT and GA.

[0084] c) Use Bowtie2 software to align the transformed sequences to the transformed HG19 reference genome, with a minimum seed sequence length of 20 and no mismatch in the seed sequence.

[0085] 4. Calculation of PDR for each sample

[0086] The proportion of discordant reads (PDR) describes the degree of consistency of methylation in the target region. In short, if all CpG sites on the read are methylated or unmethylated, the read is classified as consistent; otherwise, it is inconsistent. At each CpG site, PDR is equal to the number of inconsistent reads covering the position divided by the total number of reads covering the position:

[0087]

[0088] Where l represents the target methylation interval, N l，d Indicates the number of reads with inconsistent methylation in the target region, N l Indicates the total number of reads located in the target methylation interval.

[0089] 5. Feature Matrix Construction

[0090] a) The PDR values ​​of each target methylation interval of each sample in the training set and the test set are merged into the feature matrix of the training set and the test set, and the target methylation interval with less than 100 reads is regarded as a missing value.

[0091] b) Remove target methylation intervals with missing value ratios higher than 10%.

[0092] c) Use the KNN algorithm to train the transformer on the training set matrix, and use the transformer to interpolate missing data on the training set and test set feature matrices.

[0093] 6. Find out the methylation markers and modeling process of benign and malignant lung nodules based on the training set samples

[0094] The present application uses a grid search method to find the best combination of classifiers and parameters in the training set. First, logistic regression, support vector machine (SVM) and random forest models are pre-selected for classifier selection. For all features, a 3-fold cross-validation grid search method is used to select the classifier, and the process selects logistic regression as the optimal classifier. Then a grid search is performed for the combination of different modeling parameters and thresholds, and the features with the best classification ability are identified using the logistic regression model. Finally, 43 feature models are selected and screened out by the model. These 43 methylation feature regions and adjacent genes are shown in Table 2; These processes are implemented with python (version 3.6.13) and scikit-learn (version 0.24.2).

[0095] Table 2 Genomic locations and associated genes of 43 methylation markers

[0096]

[0097]

[0098] 7. Modeling effect evaluation

[0099] The above-mentioned training set (Training), validation set (Validation), single-center test set (Single-center Test) and multi-center test set (Multi-center Test) were used for effect evaluation.

[0100] a) Evaluation of training set and validation set effects:

[0101] The methylation levels of the above 43 methylation markers were used to construct a logistic regression machine learning model to distinguish benign and malignant pulmonary nodules, and then validated on other data sets. The specific steps are as follows:

[0102] Use the logistic regression model in the sklearn package in Python: AllModel = LogisticRegr ession(C = 10, penalty = 'l2', random_state = 99)

[0103] Use the samples in the training set for training: AllModel.fit(Traindata, TrainPheno), where TrainData is the data in the training set, TrainPheno is the characteristics of the samples in the training set (malignant nodules are 1, benign nodules are 0), and determine the relevant thresholds of the model based on the samples in the training set.

[0104] Use samples from the validation set, single-center test set, and multi-center test set for testing: TestPred = AllModel.predict_proba(TestData)[:,1], where TestData is the validation set or test set data, TestPred is the model prediction score, and use the prediction score and the above threshold to determine whether the sample is a malignant nodule.

[0105] The model prediction scores of the training set, validation set, single-center test set, and multi-center test set are shown in Figure 1 , Figure 3 From the figure, we can see that there is a significant difference in the scores of benign pulmonary nodule samples and malignant pulmonary nodule samples. The corresponding ROC curves are shown in Figure 2 , Figure 4The training set AUC was 0.973, the validation set AUC was 0.810, and the single-center and multi-center test set AUCs were 0.815 and 0.761, respectively. The prediction score threshold was set to 0.63. If the score was greater than this value, it was predicted to be a malignant pulmonary nodule, otherwise it was predicted to be a benign pulmonary nodule. Under this threshold, the single-center and multi-center test set specificities were 0.857 and 0.846, and the sensitivities were 0.6 and 0.391, respectively (see Table 3).

[0106] Table 3 Effect of machine learning diagnosis model

[0107] AUC Threshold accuracy Specificity Sensitivity Training set 0.973 0.63 0.918 0.949 0.903 Validation set 0.810 0.63 0.710 0.789 0.674 Single-center test set 0.815 0.63 0.689 0.857 0.600 Multicenter test set 0.761 0.63 0.516 0.846 0.391

[0108] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A methylation marker for detecting benign or malignant pulmonary nodules, characterized in that: The methylation marker sequences are shown in Seq ID No: 1 to Seq ID No:

43.

2. A primer or probe for detecting the methylation marker according to claim 1, characterized in that: The primer uses the nucleotide sequence where the methylation marker is located as the target sequence and is used for specific amplification of the target sequence; the probe specifically captures the nucleotide sequence where the methylation marker is located.

3. A kit for detecting benign or malignant pulmonary nodules, characterized in that: The kit contains reagents for detecting the benign and malignant methylation markers of pulmonary nodules as claimed in claim 1.

4. Use of a reagent for detecting the methylation level of the benign and malignant methylation marker of pulmonary nodules according to claim 1 in the preparation of a diagnostic kit for identifying benign and malignant pulmonary nodules in a sample.

5. The use according to claim 4, characterized in that: The detection reagents also include reagents used in any one of PCR amplification method, fluorescent quantitative PCR method, digital PCR method, liquid phase chip method, second-generation sequencing method, third-generation sequencing method, bisulfite sequencing method, whole genome methylation sequencing method, and methylation chip method, or a combination thereof.

6. A method for constructing a prediction and evaluation model for benign and malignant pulmonary nodules, characterized in that: The steps include: (a) Collect benign and malignant lung nodule samples and divide them into training sets and test sets; (b) Extract cfDNA from samples, build libraries, and sequence them; (c) Perform methylation conversion processing on the sequences, compare the data, and calculate the proportion of discordant reads (PDR) value of the samples; (d) constructing a feature matrix and an algorithm model for the sample data, and screening out methylation markers for benign and malignant pulmonary nodules based on the training set sample data; (e) Use the test set sample data to verify the effect of the model; (f) Determine the methylation markers ultimately used for the prediction and evaluation of benign and malignant characteristics of pulmonary nodules; the methylation marker sequences are shown in Seq ID No: 1 to Seq ID No:

43.

7. The construction method according to claim 6, characterized in that: The algorithm model includes a machine learning model; the machine learning model includes any one of a principal component analysis model, a logistic regression analysis model, a nearest neighbor analysis model, a support vector machine and a neural network model.

8. A risk assessment model for benign and malignant pulmonary nodules, characterized in that: The risk assessment model is obtained according to the construction method according to claim 6 or 7, and comprises: a data acquisition module, at least for acquiring a sample data set; a sequencing module, at least for obtaining sequencing data of the sequence region from Seq ID No: 1 to Seq ID No: 43; a data comparison module, at least for comparing the sequencing data with a reference sequence, and determining the methylation result of the marker in the sequencing data based on the comparison result; and a result determination module, at least for calculating a prediction score threshold through statistical model analysis, and determining whether the sample to be tested is a benign or malignant pulmonary nodule.

9. An information data processing terminal, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the following steps are implemented: (a) obtaining the methylation level of the sequence region from Seq ID No: 1 to Seq ID No: 43 in the sample to be tested; (b) The score was calculated by constructing a logistic regression diagnostic model; (c) The benign and malignant nature of pulmonary nodules was identified based on the score.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the following steps are implemented: (a) Obtaining the methylation level of the sequence region from Seq ID No: 1 to Seq ID No: 43 in the sample to be tested: (b) The score is calculated by constructing a logistic regression diagnostic model; (c) The benign and malignant pulmonary nodules are identified based on the score.

Images