Lung cancer early screening detection device based on TCR sequencing as well as construction method and application of lung cancer early screening detection device

Abstract

The invention discloses a method for constructing a lung cancer early screening model based on TCR sequencing, which comprises the following steps of: firstly, constructing a TCR-enriched sequence database based on TCR sequencing data from samples of lung cancer patients and healthy people, then performing TCR sequencing on each sample, and acquiring TCR characteristics of each sample; comprising proportion characteristics of V gene and J gene, statistical characteristics of a TCR immune repertoire, convergence sequence characteristics, TCR sequence amino acid proportion characteristics, proportion characteristics of TCR sequences with different lengths, proportion characteristics of TCR sequences with different frequencies and TCR enriched sequence characteristics; screening the TCR features of each sample by using a feature screening method, and constructing a machine learning model in combination with other information of the samples to obtain a lung cancer early screening model; a lung cancer early screening detection device is constructed based on the lung cancer early screening model, and lung cancer early screening can be achieved based on TCR sequencing information of a sample.

Description

Technical Field

[0001] The present invention relates to the technical field of gene detection, and in particular to a lung cancer early screening detection device based on TCR sequencing, and a construction method and application thereof. Background Art

[0002] Lung cancer is the malignant tumor with the highest incidence in my country. Early detection of tumors and taking corresponding measures are crucial for the prevention and treatment of lung cancer. The 5-year survival rate of stage IA lung cancer can reach more than 70%, while the 5-year survival rate of stage IVA lung cancer is only about 10%. However, due to the lack of effective tumor screening methods, in current clinical practice, the proportion of tumors discovered and diagnosed at an early stage is relatively small, and the proportion of lung cancer discovered in stages I and II is only about 30%. Early detection of tumors can also help reduce the costs of diagnosing and treating tumors. The diagnosis and treatment costs required for tumors discovered in the cancer metastasis stage are generally 2 to 3 times that of the early stage.

[0003] Currently, low-dose spiral CT is commonly used in clinical practice to screen specific populations for lung cancer. However, this method has a high false-positive rate, which can easily lead to overdiagnosis and treatment, increase patient risks and medical burdens, and also has radiation problems.

[0004] T cell receptor (TCR) refers to the specific antibody on the surface of T cells, which is responsible for recognizing antigens presented by the major histocompatibility complex (MHC) and plays a key role in recognizing tumor-related antigens. TCR is a heterodimer that can be composed of αβ chains or γδ chains, of which αβ chains account for more than 90% of TCRs.

[0005] The TCR sequence is generated by rearrangement and insertion of gene fragments from the V gene (variable), D gene (diversity) and J gene (joining) regions, and has rich diversity. The complementary-determining region (CDR) of the TCR is used to bind to the antigen-MHC (pMHC) complex; the CDR3 region of the β chain of the TCR is mainly responsible for the specific recognition of the antigenic determinant (epitope).

[0006] About 10 per individual 11 T cells, consisting of 10 6 ~10 8The T cell immune repertoire refers to the entire TCR sequence of an individual. By analyzing the T cell immune repertoire, we can understand the individual's immune status, previous infection history, immune response and other information, which may play a key role in predicting the occurrence and development of tumors. TCR sequencing refers to the analysis of an individual's T cell immune repertoire to confirm the type (clonotype) and number (clone counts) of TCR sequences. Currently, most TCR sequencing methods mainly refer to sequencing the CDR3 region of the TCRβ chain.

[0007] At present, some studies have used TCR sequencing for early cancer screening, but most of these studies use statistical indicators of the TCR immune repertoire and features such as kmer diversity of TCR sequences to build machine learning models for early cancer screening. Models based on statistical indicators of the TCR immune repertoire use features that are too simple and fail to fully utilize the rich information provided by TCR sequencing, and may find it difficult to capture early tumor-related signals. Models based on kmer diversity of TCR sequences often have too many input features and are prone to overfitting models.

[0008] Therefore, there is an urgent need for a method that can use TCR sequencing information for early tumor screening, so as to achieve accurate early screening of tumors through the rich information provided by TCR sequencing. Summary of the invention

[0009] The purpose of the present invention is to overcome the above-mentioned deficiencies in the prior art and to provide a lung cancer early screening detection device based on TCR sequencing and a construction method and application thereof.

[0010] The first objective of the present invention is to provide a method for constructing an early screening model for lung cancer based on TCR sequencing.

[0011] The second objective of the present invention is to provide a lung cancer early screening detection device based on TCR sequencing.

[0012] In order to achieve the above object, the present invention is implemented by the following scheme:

[0013] A method for constructing a lung cancer early screening model based on TCR sequencing comprises the following steps:

[0014] S1. Establishment of enriched TCR sequence database:

[0015] S11. Obtaining an experimental group TCR immune repertoire and a control group TCR immune repertoire, and respectively calculating the number of occurrences of each TCR sequence in the experimental group TCR immune repertoire and the control group TCR immune repertoire; the experimental group contains samples from lung cancer patients, and the control group contains samples from healthy people;

[0016] S12. removing TCR sequences with a number of occurrences ≤ 5 in the control group TCR immune library obtained in step S11, to obtain a control group TCR sequence data set;

[0017] Removing low-abundance TCR sequences in the control group TCR sequence data set to obtain a control group enriched TCR sequence database; removing TCR sequences that appeared in the control group TCR sequence data set in the experimental group TCR immune repertoire obtained in step S11, and removing low-abundance TCR sequences to obtain an experimental group enriched TCR sequence database;

[0018] S2. Construction of lung cancer early screening model:

[0019] The machine learning model is cross-validated and trained using the features of the samples in the training set to obtain a lung cancer early screening model; the training set includes positive samples and negative samples, the positive samples are blood and / or tissue samples of malignant lung nodule samples; the negative samples are blood and / or tissue samples of benign lung nodule samples, or blood samples of healthy people; the features include TCR features, clinical features and / or imaging features; the TCR features include: V gene and J gene proportion features, statistical features of the TCR immune repertoire, convergent sequence features, TCR sequence amino acid proportion features, TCR sequence proportion features of different lengths, TCR sequence proportion features of different frequencies, and TCR enrichment sequence features;

[0020] The method for obtaining the TCR feature is specifically as follows:

[0021] The TCR sequencing data of each sample in the training set was quality controlled, and then the TCRs aligned with the CDR3 region of TCRβ were converted into clonotypes files to obtain the data to be analyzed for the training set samples;

[0022] V gene and J gene ratio features: Based on the data to be analyzed of the samples in the training set, the corrected clonecounts of each V gene and J gene in each TCR sequence of the sample are obtained as the V gene and J gene ratio features of the sample;

[0023] Convergent sequence features: Based on the data to be analyzed of the samples in the training set, TCR sequence reads greater than or equal to 1 to 100, and TCR sequences that do not contain stop codons and / or have frameshift mutations are retained as functional TCR sequences of the samples, and the features of the convergent sequences are calculated; the convergent sequences are TCR sequences with the same amino acid sequence but different nucleotide sequences;

[0024] The characteristics of convergent sequences are: the proportion of the number of convergent sequences, the proportion of clonecounts of convergent sequences, and the proportion of clonecounts after correction of convergent sequences;

[0025] TCR enrichment sequence features: (1) For the TCR sequences with the same length in the training set sample data to be analyzed and the control group enrichment TCR sequence database in step S1, retain the TCR sequence pairs with an edit distance ≤ 0.3 × b, where b is the TCR sequence length, and select the TCR sequence pairs with alignment_score ≥ 4.5 × b as the TCR sequence combination;

[0026] Based on the TCR sequence combination, the TCR enrichment sequence feature enrich_score1 is calculated according to formula II, and the TCR enrichment sequence feature enrich_type1 is calculated according to formula III;

[0027] Formula II:

[0028] Among them, alignment_score ij is the alignment_score of TCR sequence i and TCR sequence j of the same length; n ki is the number of TCR sequences i in sample k, sample k belongs to The samples in Enrich the TCR sequence database for the control group; query_seq_score j =ln(n j ), n j is the number of TCR sequence j in the TCR sequence combination, and TCR sequence j comes from the data to be analyzed of the sample; query_seqs is the number of TCR sequences in the data to be analyzed of the sample;

[0029] Formula III:

[0030] Where enriched_seqs is the number of TCR sequences from the sample in the TCR sequence combination; query_seqs is the number of TCR sequences in the sample data to be analyzed

[0031] (2) According to the method shown in (1), based on the sample to be analyzed and the enriched TCR sequence database of the experimental group shown in step S1, the TCR enrichment features enrich_score2 and enrich_type2 are calculated.

[0032] Preferably, the patient in step S11 is a lung cancer patient.

[0033] Preferably, the experimental group shown in step S11 contains no less than 50 patient samples, and the control group shown contains no less than 50 healthy person samples.

[0034] Preferably, the step S11 of obtaining the experimental group TCR immune repertoire and the control group TCR immune repertoire is specifically as follows: performing TCR sequencing, data analysis and quality control on the experimental group containing patient samples and the control group containing healthy human samples, respectively, to obtain the TCR immune repertoire of each sample in the experimental group after quality control and the TCR immune repertoire of each sample in the control group after quality control, and merge them respectively to obtain the experimental group TCR immune repertoire and the control group TCR immune repertoire.

[0035] More preferably, the TCR sequencing is to construct a library and sequence according to the primers and multiplex PCR amplification method described in the prior art CN114774517A.

[0036] More preferably, the data analysis specifically includes: quality control, followed by comparison with the CDR3 region of TCRβ to obtain the compared TCR sequence.

[0037] Further preferably, the quality control specifically includes: filtering the fastq file downloaded from the TCR sequencing machine using a data filtering tool, removing the linker and low-quality sequences, and removing the TCR sequence that does not contain the TCR sequencing primer.

[0038] More preferably, the data filtering tool is Fastp, FastQC, flash and / or seqkit.

[0039] More preferably, Fastp is used to filter, remove linkers and remove low-quality sequences based on default parameters, and then seqkit is used to remove TCR sequences that do not contain TCR sequencing primers.

[0040] Further preferably, the CDR3 region of the TCRβ is the CDR3 region of human TCRβ in the IMGT database.

[0041] More preferably, the CDR3 region of human TCRβ in the IMGT database specifically refers to: https: / / www.imgt.org / download / V-QUEST / IMGT_V-QUEST_reference_directory / Homo_sapies / TR / TRBV.fasta.

[0042] More preferably, the quality control is to remove TCR sequences with a number of less than 1-100.

[0043] Further preferably, TCR sequences with a number of TCR sequences less than 10 are removed.

[0044] More preferably, removing low-abundance TCR sequences in the control group TCR sequence data set in step S12 specifically includes: removing TCR sequences whose occurrence times in the control group TCR sequence data set are ≤ n*a; n is 0.001 to 0.1, and a is the number of samples in the control group.

[0045] More preferably, removing low-abundance TCR sequences in step S12 specifically includes: removing TCR sequences with a number of occurrences ≤ n1*b; n1 is 0.001 to 0.1, and b is the number of samples in the experimental group.

[0046] Preferably, in step S2, k-fold cross validation and / or leave-one-out cross validation are used to perform cross validation training on the machine learning model.

[0047] More preferably, k=10 in the k-fold crosss validation.

[0048] Preferably, the machine learning model in step S2 is a random forest, a generalized linear model, a gradient boosting model, a linear support vector machine and / or a radial basis function kernel support vector machine.

[0049] Preferably, during the training process in step S2, the expected output of positive samples in the training set is ≥ 0.5, and the expected output of negative samples is < 0.5.

[0050] Preferably, the clinical characteristics in step S2 include whether there is a history of smoking, whether smoking has been quit, whether lung cancer has been suffered, whether COPD has been suffered and / or whether there is a family history of lung cancer; the imaging characteristics include the size of the lung nodules, the number of lung nodules, whether it is a solid nodule, whether it is a partially solid nodule, whether it is a ground glass nodule, whether there are burrs and / or the location of the lung nodules.

[0051] Among them, the numerical features in clinical characteristics and influence characteristics use their numerical values ​​as feature values, and the binary features are converted to 0 / 1 feature values.

[0052] Preferably, before cross-training in step S2, features of samples in the training set are screened using Ensemble Feature selection and / or boruta feature screening methods.

[0053] More preferably, the Ensemble Feature selection method includes recursive feature elimination, random forest and / or glmnet.

[0054] Further preferably, the Ensemble Feature selection method includes glmnet.

[0055] More preferably, after screening the features of the samples in the training set, data balancing is performed.

[0056] Further preferably, data balancing is performed using oversampling and / or undersampling methods.

[0057] More preferably, the SMOTE algorithm is used to balance the features of the samples in the training set, that is, the data from the positive samples and the data from the negative samples in the features of the samples in the training set are balanced.

[0058] Preferably, the TCR sequencing data in step S2 is TCR sequencing data obtained by constructing a library and sequencing according to the primers and multiplex PCR amplification method described in the prior art CN114774517A.

[0059] Preferably, the CDR3 region of TCRβ in step S2 is the CDR3 region of human TCRβ in the IMGT database.

[0060] More preferably, the CDR3 region of human TCRβ in the IMGT database can be specifically referred to: https: / / www.imgt.org / download / V-QUEST / IMGT_V-QUEST_reference_directory / Homo_sapies / TR / TRBV.fasta.

[0061] Preferably, the corrected clonecounts in step S2 is the ln value of clonecounts.

[0062] Preferably, the statistical characteristics of the TCR immune repertoire in step S2 include Shannon index, d50Index index, reverse Simpson index and homogeneity index.

[0063] More preferably, the method for obtaining the TCR immune repertoire is: using the TCR sequence of the data to be analyzed of the sample in the training set as the TCR immune repertoire of the sample, and calculating the statistical characteristics of the TCR immune repertoire.

[0064] Further preferably, when calculating the statistical characteristics of the TCR immune repertoire, when the number of TCR sequences of the sample data to be analyzed is greater than m, m TCR sequences are randomly selected as the TCR immune repertoire of the sample; m=1E6~1E7.

[0065] More preferably, m=1E6.

[0066] Preferably, when calculating the convergent sequence features in step S2, TCR sequences with a number of TCR sequence reads ≥ 10 are retained.

[0067] Among them, the proportion of the number of convergent sequences (Convergence_type) is the proportion of the number of TCR sequence types of convergent sequences in the functional TCR sequence to the number of TCR sequence types in the functional TCR sequence; the proportion of convergent sequence clonecounts (Convergence_freq) is the ratio of the number of convergent sequences in the functional TCR sequence to the number of TCR sequences in the functional TCR sequence; the proportion of convergent sequence corrected clonecounts (Convergence_score) is the ratio of the corrected number of convergent sequences in the functional TCR sequence to the corrected number of TCR sequences in the functional TCR sequence; the corrected number of convergent sequences in the functional TCR sequence is the number of convergent sequences in the functional TCR sequence taken as the ln value; the corrected number of TCR sequences in the functional TCR sequence is the number of TCR sequences in the functional TCR sequence taken as the ln value.

[0068] Preferably, the method for obtaining the amino acid proportion feature of the TCR sequence in step S2 is: based on the functional TCR sequence of the sample, 1 to 4 amino acids at both ends of each TCR sequence are removed, and the proportion of each amino acid in the TCR sequence after the amino acids are deleted is calculated.

[0069] More preferably, 3 amino acids at both ends of each TCR sequence are removed.

[0070] Preferably, the method for acquiring the characteristics of the proportion of TCR sequences of different lengths and the characteristics of the proportion of TCR sequences of different frequencies in step S2 is: based on the functional TCR sequence of the sample, it is grouped by length, and the quantitative ratio of TCR sequences in each length group is calculated to obtain the proportion of TCR sequences of each length; based on the functional TCR sequence of the sample, it is grouped by frequency, and the quantitative ratio of TCR sequences in each frequency group is calculated to obtain the proportion of TCR sequences of each frequency.

[0071] More preferably, when calculating the proportion of TCR sequences of each length, the sequences are grouped according to lengths of ≤10aa, 11-13aa, 14-16aa, 17-19aa, and ≥20aa.

[0072] More preferably, when calculating the proportion of TCR sequences of each frequency, grouping is performed according to the occurrence frequency of ≥1E-2, 1E-3 to 1E-4, 1E-4 to 1E-5, 1E-5 to 1E-6, 1E-6 to 1E-7 and ≤1E-7.

[0073] Wherein, when calculating the TCR enriched sequence features in step S2, the TCR sequence combination shown contains multiple TCR sequence pairs, each TCR sequence pair contains 1 TCR sequence from the data to be analyzed of the sample and 1 TCR sequence from the enriched TCR sequence database of the control group.

[0074] Preferably, in step S2, the Needleman-Wunsch algorithm is used to calculate the alignment_score.

[0075] More preferably, in the process of calculating alignment_score, the amino acid substitution scoring matrix is ​​Blosum62, the open penalty is -10, and the extend penalty is -1.

[0076] The present invention also seeks to protect the lung cancer early screening model constructed by the above construction method.

[0077] The present invention also seeks to protect a lung cancer early screening detection device based on TCR sequencing, comprising a sequencing module, an analysis module, a prediction module and a result output module;

[0078] The sequencing module is used to perform TCR sequencing on the sample to be tested and obtain the fastq file in the sequencing file;

[0079] The analysis module performs the steps shown in step S2 in the above construction method on the fastq file obtained by the sequencing module to obtain the TCR characteristics of the sample to be tested, and obtains the input characteristics of the sample to be tested after screening;

[0080] The prediction module is a lung cancer early screening model constructed by the above construction method, and the input features of the sample to be tested obtained by the analysis module after screening are used as input variables, which are input into the lung cancer early screening model to obtain the lung cancer prediction score of the sample to be tested;

[0081] The result output module is used to output the lung cancer prediction score of the sample to be tested obtained by the prediction module.

[0082] Preferably, in the analysis module, the input features of the sample to be tested after screening include: the proportion features of TRBJ2-6 genes, the proportion features of TRBV18 genes, the proportion features of TRBV19 genes, the proportion features of TRBV2 genes, the proportion features of TRBV27 genes, the proportion features of TRBV28 genes, the proportion features of TRBV3-1 genes, the proportion features of TRBV5-1 genes, the proportion features of TRBV5-4 genes, the proportion features of TRBV5-5 genes, the proportion features of TRBV7-8 genes Characteristics, the proportion of TRBV9 genes, the proportion of glycine, the proportion of leucine, the proportion of asparagine, the proportion of tryptophan, the proportion of TCR sequences with a frequency of 1E-5 to 1E-6, the enriched sequence feature enrich_score of TCR, the Shannon index, the reverse Simpson index, the homogeneity index, the proportion of clonecounts after convergence sequence correction Convergence_score, whether to quit smoking, whether there is a history of smoking, whether it is a ground glass nodule, whether it is a solid nodule, whether there are spicules and the size of the lung nodule.

[0083] Among them, the TRBJ2-6 gene (Gene ID: 28623) in the input features after screening of the sample to be tested belongs to the J gene of the TCR sequence; the TRBV18 gene (Gene ID: 28569), TRBV19 gene (Gene ID: 28568), TRBV2 gene (GeneID: 28620), TRBV27 gene (Gene ID: 28560), TRBV28 gene (Gene ID: 28559), TRBV3-1 gene (Gene ID: 28619), TRBV5-1 gene (Gene ID: 28614), TRBV5-4 gene (GeneID: 28611), TRBV5-5 gene (Gene ID: 28610), TRBV7-8 gene (Gene ID: 28590) and TRBV9 gene (Gene ID: 28586) belong to the V gene of the TCR sequence.

[0084] Preferably, the lung cancer prediction score is [0, 1], the input result is 0, the probability that the sample to be tested is lung cancer is 0, and the output result is 1, the probability that the sample to be tested is lung cancer is 1.

[0085] More preferably, the cutoff value of the lung cancer prediction score is 0.5, and if the output result is ≥0.5, the sample to be tested is determined to be a positive sample (i.e., a malignant lung nodule); if the output result is <0.5, the sample is determined to be a negative sample (i.e., a benign lung nodule or a healthy sample).

[0086] Compared with the prior art, the present invention has the following beneficial effects:

[0087] The present invention provides a method for constructing a lung cancer early screening model based on TCR sequencing. First, an enriched TCR sequence database is constructed based on TCR sequencing data from samples of lung cancer patients and healthy people. Then, TCR sequencing is performed on each sample to obtain the TCR characteristics of each sample, including V gene and J gene proportion characteristics, statistical characteristics of the TCR immune repertoire, convergent sequence characteristics, TCR sequence amino acid proportion characteristics, TCR sequence proportion characteristics of different lengths, TCR sequence proportion characteristics of different frequencies, and TCR enrichment sequence characteristics. After the TCR characteristics of the sample are screened by a feature screening method, a machine learning model is constructed in combination with other information of the sample to obtain a lung cancer early screening model. A lung cancer early screening detection device is constructed based on the lung cancer early screening model, which can realize lung cancer early screening based on the TCR sequencing information of the sample. BRIEF DESCRIPTION OF THE DRAWINGS

[0088] Figure 1 This is a flow chart for constructing the lung cancer early screening model shown in Example 3;

[0089] Figure 2 The ROC-AUC curve diagram of the five candidate lung cancer early screening models trained in Example 3;

[0090] Figure 3 : is the output result distribution diagram of the five candidate lung cancer early screening models trained in Example 3; A is the output result distribution diagram of the trained random forest, B is the output result distribution diagram of the trained gradient boosting model, C is the output result distribution diagram of the trained generalized linear model, D is the output result distribution diagram of the trained svmLinear, and E is the output result distribution diagram of the trained svmRadial;

[0091] Figure 4 is a ROC-AUC curve diagram of the lung cancer early screening detection device in Example 4;

[0092] Figure 5 It is the output result distribution diagram of the lung cancer early screening detection device in Example 4; A is the output result distribution diagram of the lung cancer early screening detection device 1, B is the output result distribution diagram of the lung cancer early screening detection device 2, C is the output result distribution diagram of the lung cancer early screening detection device, D is the output result distribution diagram of the lung cancer early screening detection device 3, and E is the output result distribution diagram of the lung cancer early screening detection device 4. DETAILED DESCRIPTION

[0093] The present invention is further described in detail below in conjunction with the accompanying drawings and specific examples of the specification. The examples are only used to explain the present invention and are not used to limit the scope of the present invention. The test methods used in the following examples are conventional methods unless otherwise specified; the materials and reagents used are reagents and materials that can be obtained from commercial channels unless otherwise specified.

[0094] The samples used in the embodiments of the present invention are all from Shenzhen Hypros Biotechnology Co., Ltd., and informed consent has been obtained from the patients before using the samples for testing.

[0095] Example 1 A method for obtaining and analyzing lung cancer enriched TCR sequences

[0096] 1. Experimental Methods

[0097] 1. TCR Sequencing

[0098] DNA was extracted from each sample in experimental group 1, experimental group 2 and control group using a DNA extraction kit, and a TCR sequencing library was constructed by multiplex PCR amplification using 40 V gene-specific primers and 12 J gene-specific primers for the DNA of each sample in each group (primers and multiplex PCR amplification method are as shown in the prior art CN114774517A). The constructed TCR library was sequenced using a Novasek-2000 sequencer, and the TCR sequencing fastq file of each sample in each group was obtained.

[0099] The sample types and sample numbers of experimental group 1, experimental group 2 and control group are shown in Table 1.

[0100] Table 1 Sample types and sample numbers of experimental group 1, experimental group 2 and control group

[0101] Group Sample Type Sample size Control group Healthy Blood 2699 Experimental Group 1 Lung Cancer Blood 3360 Experimental Group 2 Lung Cancer Tissue 988

[0102] 2. Data analysis and quality control

[0103] (1) The Fastp software was used to filter, remove the linker, and remove the low-quality sequences of the TCR sequencing fastq files of each sample in each group based on the default parameters, and the seqkit software was used to remove the TCR sequences that did not contain the TCR sequencing primers (i.e., the specific primers shown in the technical CN114774517A) based on the preset parameters. Then, the mixcr software was used to align the CDR3 region of the human TCRβ gene sequence in the IMGT database (https: / / www.imgt.org / download / V-QUEST / IMGT_V-QUEST_reference_directory / Homo_sapiens / TR / TRBV.fasta), and the alignment results were sorted with vdjtools to obtain the TCR immune repertoire of each sample in each group;

[0104] The TCR immune repertoire contains the types (clonotypes) of TCR sequences in the CDR3 region and the number (clone counts) of each TCR sequence.

[0105] (2) TCR sequences with clone counts < 10 in the TCR immune repertoire of each sample in each group obtained in step (1) are removed respectively to obtain the TCR immune repertoires of each sample in experimental group 1, experimental group 2 and control group after quality control.

[0106] (3) The TCR immune repertoires of the samples in the experimental group 1 obtained in step (2) after quality control are combined to obtain the TCR immune repertoire of the experimental group 1; the TCR immune repertoires of the samples in the experimental group 2 obtained in step (2) after quality control are combined to obtain the TCR immune repertoire of the experimental group 2; the TCR immune repertoires of the samples in the control group obtained in step (2) after quality control are combined to obtain the TCR immune repertoire of the control group;

[0107] Count the number of times each TCR sequence in each TCR immune repertoire appears in the immune repertoire (db_seq_count) and calculate the score of each TCR sequence (db_seq_score);

[0108] where db_seq_score i =∑ k∈S ln(n ki ), n ki represents the number of TCR sequence i in sample k, S represents experimental group 1, experimental group 2 or control group, k∈S represents sample k comes from experimental group 1, experimental group 2 or control group;

[0109] (4) Based on the control group TCR immune repertoire obtained in step (3), TCR sequences with a number of appearances (db_seq_count) ≤ 5 in the control group TCR immune repertoire are removed to obtain a control group TCR sequence data set; based on the experimental group 1 TCR immune repertoire obtained in step (3), TCR sequences that have appeared in the control group TCR sequence data set are removed, and TCR sequences that have appeared a number of times ≤ 17 in the experimental group 1 TCR immune repertoire are removed to obtain an enriched TCR sequence database for experimental group 1; based on the experimental group 2 TCR immune repertoire obtained in step (3), TCR sequences that have appeared in the control group TCR sequence data set are removed, and TCR sequences that have appeared a number of times ≤ 15 in the experimental group 2 TCR immune repertoire are removed to obtain an enriched TCR sequence database for experimental group 2.

[0110] TCR sequences with a number of occurrences ≤ 150 in the control group TCR sequence data set were removed to obtain a control group enriched TCR sequence database.

[0111] 2. Experimental Results

[0112] The numbers of enriched sequences in the control group enriched TCR sequence database, the experimental group 1 enriched TCR sequence database, and the experimental group 2 enriched TCR sequence database are shown in Table 2.

[0113] Table 2 Number of enriched sequences

[0114] Group Sample Type Enriched TCR sequence number Control group enriched TCR sequence database Healthy Blood 5480 Experimental group 1 enrichment of TCR sequence database Lung Cancer Blood 3840 Experimental group 2 enrichment of TCR sequence database Lung Cancer Tissue 3652

[0115] The TCR sequences in each enriched TCR database were sorted by count. The top 30 enriched sequences in the enriched TCR sequence database of the control group are shown in Table 3, the top 30 enriched sequences in the enriched TCR sequence database of the experimental group 1 are shown in Table 4, and the top 30 enriched sequences in the enriched TCR sequence database of the experimental group 2 are shown in Table 5.

[0116] Table 3 The top 30 enriched sequences in the control group enriched TCR sequence database

[0117]

[0118]

[0119] Table 4 The top 30 enriched sequences in the enriched TCR sequence database of experimental group 1

[0120]

[0121]

[0122] Table 5 The top 30 enriched sequences in the enriched TCR sequence database of experimental group 2

[0123]

[0124] Example 2: Method for extracting TCR features for early screening prediction of lung cancer

[0125] Pulmonary peripheral blood was used as a sample, and TCR sequencing was performed on the sample according to step 1 of Example 1 to obtain a TCR sequencing fastq file of the sample. The TCR sequencing fastq file of the sample was filtered, removed from the linker, and low-quality sequences were removed using Fastp software based on default parameters, and the TCR sequence that did not contain the TCR sequencing primer (i.e., the specific primer shown in the prior art CN114774517A) was removed using seqkit software based on preset parameters. Then, the CDR3 region of the human TCRβ gene sequence in the IMGT database (https: / / www.imgt.org / download / V-QUEST / IMGT_V-QUEST_reference_directory / Homo_sapiens / TR / TRBV.fasta) was used to align, and the alignment results were sorted using vdjtools to obtain the clonotypes file of the sample; the clonotypes file of the sample records the type (clonotype) of the TCR sequence in the CDR3 region of the sample and the number (clonecounts) of each TCR sequence.

[0126] The clonotypes file of the sample is used as the data to be analyzed for the sample, and the TCR characteristics of the sample are extracted, including the proportion characteristics of each V gene (variable region gene Variable gene) and J gene (joining region gene Joining gene) in the TCR sequence, the statistical characteristics of the TCR immune repertoire, the convergent sequence characteristics, the proportion characteristics of the TCR sequence amino acids, the proportion characteristics of TCR sequences of different lengths, the proportion characteristics of TCR sequences of different frequencies, and the characteristics of TCR enrichment sequences, as follows:

[0127] 1. Characteristics of the proportion of V gene and J gene:

[0128] Based on the data to be analyzed of the sample, the clonecounts of each V gene and J gene in each TCR sequence of the sample are obtained and the ln value is taken to obtain the adjusted clonecounts (adjustedclonecounts) of each V gene and J gene in the sample as the proportion feature of the V gene and the proportion feature of the J gene of the sample.

[0129] 2. Statistical characteristics of the TCR immune repertoire:

[0130] Based on the sample data to be analyzed, 10 6The TCR sequences of the samples (if the number is insufficient, all are selected) are used as the TCR immune repertoire of the samples. Based on the TCR immune repertoire of the samples, vdjtools is used to calculate the statistical characteristics of the TCR immune repertoire of the samples, including the Shannon index (Formula IV), d50Index index (TCR diversity index), reverse Simpson index (Formula V) and homogeneity index (Formula VI);

[0131] Formula IV:

[0132] Formula V:

[0133] Formula VI:

[0134] in is the set of TCR sequences in the TCR immune repertoire of the sample; C is the number of types of TCR sequences in the TCR immune repertoire of the sample; p i is the ratio of the number of TCR sequences i in the TCR immune repertoire of the sample to the number of TCR sequences in the TCR immune repertoire of the sample n i is the number of TCR sequence i, is the number of TCR sequences in the TCR immune repertoire of the sample.

[0135] 3. Convergent sequence characteristics:

[0136] Remove non-functional TCR sequences (TCR sequences containing stop codons or frameshift mutations) from the data to be analyzed, and remove TCR sequences with TCR sequence reads < 10 to obtain the functional TCR sequences of the samples, and calculate the features of convergent sequences based on the functional TCR sequences;

[0137] Among them, convergent sequences are TCR sequences with the same amino acid sequence but different nucleotide sequences;

[0138] The characteristics of convergent sequences include: the proportion of convergent sequence number Convergence_type, the proportion of convergent sequence clonecounts Convergence_freq, and the proportion of convergent sequence corrected clonecounts Convergence_score;

[0139] Convergence_type is the ratio of the number of TCR sequence types with convergent sequences in the functional TCR sequence to the number of TCR sequence types in the functional TCR sequence;

[0140] Convergence_freq is the ratio of the number of convergent sequences in the functional TCR sequence to the number of TCR sequences in the functional TCR sequence;

[0141] Convergence_score is the ratio of the corrected number of convergent sequences in the functional TCR sequence to the corrected number of TCR sequences in the functional TCR sequence;

[0142] The corrected number of convergent sequences in the functional TCR sequence is the number of convergent sequences in the functional TCR sequence taken as the ln value; the corrected number of TCR sequences in the functional TCR sequence is the number of TCR sequences in the functional TCR sequence taken as the ln value.

[0143] 4. Characteristics of amino acid ratios in TCR sequences:

[0144] Delete the three amino acids at both ends (5' and 3' ends) of each TCR sequence in the functional TCR sequence obtained in step 3 to obtain the TCR sequence of the sample after the amino acids are deleted, and calculate the proportion of each amino acid in the TCR sequence of the sample after the amino acids are deleted.

[0145] 5. Characteristics of the proportion of TCR sequences of different lengths and different frequencies:

[0146] For the functional TCR sequences obtained in step 3, group them according to Table 6 based on the length of the TCR sequences, and group them according to Table 7 based on the frequency of the TCR sequences, and calculate the proportion of the number of TCR sequences in each group when grouped by different lengths and different frequencies to the number of TCR sequences in the functional TCR sequence, and obtain the proportion of TCR sequences of different lengths and the proportion of TCR sequences of different frequencies.

[0147] Table 6 TCR sequence length grouping

[0148] TCR sequence length ≤10 11～13 14～16 17～19 ≥20 Group A B C D E

[0149] Table 7 TCR sequence frequency grouping

[0150] TCR sequence frequency ≥1E-2 1E-3~1E-4 1E-4~1E-5 1E-5~1E-6 1E-6~1E-7 ≤1E-7 Group 1 2 3 4 5 6

[0151] 6. TCR enrichment sequence characteristics:

[0152] A sequence similarity comparison is performed between each TCR sequence in the data to be analyzed of the sample and the TCR sequence in the enriched TCR sequence database of the control group obtained in Example 1, and TCR sequence pairs are obtained for the TCR sequences in the data to be analyzed of the sample and the TCR sequences in the enriched TCR sequence database of the control group, both of which have a length of b (b>0); each TCR sequence pair contains one TCR sequence derived from the data to be analyzed of the sample and one TCR sequence derived from the enriched TCR sequence database of the control group, and the TCR sequences in the TCR sequence pair satisfy an edit distance ≤ 0.3*b; the edit distance (Edit Distance) is the minimum number of editing operations required to convert one TCR sequence into another TCR sequence.

[0153] For each TCR sequence pair, the Needleman-Wunsch algorithm was used to calculate the alignment_score, and the TCR sequence pairs with alignment_score ≥ 4.5 × b were used to obtain the TCR sequence combination;

[0154] In the process of calculating alignment_score, the amino acid substitution scoring matrix is ​​Blosum62, the open penalty is -10, and the extend penalty is -1;

[0155] Based on the TCR sequence combination, the TCR enrichment sequence feature enrich_score1 of the sample is calculated according to Formula II, and the TCR enrichment sequence feature enrich_type1 of the sample is calculated according to Formula III;

[0156] Formula II:

[0157] Among them, alignment_score ij is the alignment_score of TCR sequence i and TCR sequence j of the same length; n ki is the number of TCR sequences i in sample k in the TCR sequence combination, sample k belongs to The samples in To enrich the TCR sequence database; query_seq_score j =ln(n j ), n j is the number of TCR sequences j in the TCR sequence group that originate from the data to be analyzed of the sample; query_seqs is the number of TCR sequences in the data to be analyzed of the sample;

[0158] Formula III:

[0159] Where enriched_seqs is the number of TCR sequences in the TCR sequence combination that originates from the data to be analyzed in the sample; query_seqs is the number of TCR sequences in the data to be analyzed in the sample.

[0160] Each TCR sequence in the sample data to be analyzed is compared for sequence similarity with the TCR sequence in the enriched TCR sequence database of experimental group 1 obtained in Example 1 according to the above method, and the TCR enriched sequence feature enrich_score2 and TCR enriched sequence feature enrich_type2 of the sample are calculated.

[0161] Each TCR sequence in the sample data to be analyzed is compared for sequence similarity with the TCR sequence in the enriched TCR sequence database of experimental group 2 obtained in Example 1 according to the above method, and the TCR enriched sequence feature enrich_score3 and TCR enriched sequence feature enrich_type3 of the sample are calculated.

[0162] Example 3 A lung cancer early screening model based on TCR sequencing

[0163] 1. Model Training

[0164] Schematic diagram of the construction process of the lung cancer early screening model Figure 1 shown.

[0165] 1. Composition of training set and test set

[0166] A sample set containing 687 malignant pulmonary nodule samples (positive samples) and 202 benign pulmonary nodule samples (negative samples) was used as the test samples, and the test samples were divided into a training set and a test set at a ratio of 85:15, as shown in Table 8; during the division process, there was no significant difference in the proportion of benign and malignant pulmonary nodule samples in the training set and the test set.

[0167] Table 8 Training set and test set samples

[0168] Types of pulmonary nodules Training set Test Set total Malignant (positive) 583 104 687 Benign (negative) 172 30 202

[0169] At the same time, each sample in the training set and the test set also contains the clinical characteristics and imaging characteristics of each sample;

[0170] Clinical characteristics include: whether the patient has a history of smoking, whether the patient has quit smoking, whether the patient has had lung cancer, whether the patient has a family history of lung cancer, and whether the patient has COPD (chronic obstructive pulmonary disease);

[0171] Imaging features include: size of pulmonary nodules, number of pulmonary nodules, whether it is a solid nodule, whether it is a partially solid nodule, whether it is a ground glass nodule, whether there are burrs and the location of the pulmonary nodules.

[0172] 2. Screening of features for training models

[0173] (1) Extraction of TCR features of each sample

[0174] The samples in the training set and the test set were processed according to the method shown in Example 2, respectively, to obtain the TCR features of the samples in the training set and the test set (the proportion characteristics of each V gene and J gene in the TCR sequence, the statistical characteristics of the TCR immune repertoire, the convergent sequence characteristics, the TCR sequence amino acid proportion characteristics, the proportion characteristics of TCR sequences of different lengths, the proportion characteristics of TCR sequences of different frequencies, and the TCR enrichment sequence characteristics).

[0175] (2) Feature screening

[0176] The TCR features, clinical features and imaging features of the samples in the training set are combined as the input features of the samples in the training set; the numerical features in the clinical features and imaging features of the samples use their numerical values ​​as the corresponding feature values, and the binary features (yes / no, presence / absence) are converted into 1 / 0 input, 1 represents yes / presence, and 0 represents no / absence.

[0177] The positive and negative results of the samples in the training set are used as the prediction target to perform feature screening, as follows:

[0178] The caret software package is used to perform Ensemble Feature selection on the input features of the samples in the training set, and the feature importance scores of each input feature of the samples in the training set are obtained respectively; based on the feature importance scores of each input feature of the samples in the training set, the 30 input features with the highest feature importance scores are selected as the Ensemble features of the samples;

[0179] Ensemble Feature selection includes three feature screening methods: recursive feature elimination, random forest and glmnet. The feature importance score is obtained by adding the scores of each input feature of the sample obtained by each feature screening method after normalization.

[0180] The R boruta software package is used to perform boruta feature screening on the input features of the samples in the training set to obtain the boruta determined features and boruta undetermined features of the samples.

[0181] The Ensemble features of the samples, the boruta determined features of the samples, and the boruta pending features of the samples are combined to obtain the types of input features after screening as shown in Table 9, and the input features after screening of the samples in the training set are obtained.

[0182] Table 9 Types of input features after screening

[0183]

[0184]

[0185] Based on the types of filtered input features shown in Table 9, the filtered input features of each sample in the test set are obtained.

[0186] 3. Training of lung cancer early screening model

[0187] The SMOTE algorithm is used to oversample the screened input features of the samples in the training set obtained in step 2 until the screened input features of the malignant pulmonary nodule samples (positive samples) and the screened input features of the benign pulmonary nodule samples (negative samples) in the training set are balanced, and the screened input features of the samples in the training set after data balance are obtained;

[0188] After data balancing, the number of benign pulmonary nodule samples in the training set was 408, reaching 70% of the number of malignant pulmonary nodule samples.

[0189] Random Forest (rf), Generalized Linear Model (glmnet), Gradient Boosting Model (gbm), Linear Support Vector Machine (svmLinear) and Radial Basis Function Kernel Support Vector Machine (svmRadial) were used as candidate lung cancer early screening models respectively. The 10-fold cross validation method was used. The screened input features of the samples in the training set after data balance were used as input variables and were input into the 5 candidate lung cancer early screening models to train the candidate lung cancer early screening models. The grid search hyperparameter optimization method was used to optimize the hyperparameters of the candidate lung cancer early screening models until there was no significant difference between the output results of the candidate lung cancer early screening models and the actual lung nodule types (negative and positive results) of the samples in the training set. Five trained candidate lung cancer early screening models were obtained, including the trained random forest, the trained generalized linear model, the trained gradient boosting model, the trained linear support vector machine and the trained radial basis function kernel support vector machine.

[0190] Among them, during the training process, the expected output of the malignant lung nodule samples in the training set in the candidate lung cancer early screening model is ≥0.5, and the expected output of the benign lung nodule samples in the candidate lung cancer early screening model is <0.5;

[0191] The output result of the candidate lung cancer early screening model is [0, 1]. If the input result is 0, the probability that the sample is lung cancer is 0. If the output result is 1, the probability that the sample is lung cancer is 1.

[0192] The cutoff value is 0.5, and the output result is ≥0.5, and the sample is judged as a positive sample (i.e., malignant pulmonary nodule); the output result is <0.5, and the sample is judged as a negative sample (i.e., benign pulmonary nodule).

[0193] The input features after screening of each sample in the test set obtained in step 2 are used as input variables to be input into the 5 trained candidate lung cancer early screening models, and the output results of the 5 trained candidate lung cancer early screening models are obtained respectively. Based on the actual lung nodule type (negative and positive results) of each sample in the test set, the indicators (sensitivity, specificity, pos pred value, neg pred value, F1 value) of each trained lung cancer early screening model are calculated according to formula IV to formula VIII;

[0194] Formula VI: Sensitivity = TP / (TP+FN);

[0195] Formula V: Specificity = TN / (TN + FP);

[0196] Formula VI: Pos Pred Value = TP / (TP+FP);

[0197] Formula VII: Neg Pred Value=TN / (TN+FN);

[0198] Formula VIII: F1 = 2TP / (2TP+FP+FN);

[0199] Among them, TP is true positive, FP is false positive, TN is true negative, and FN is false negative.

[0200] The AUC value was calculated based on the output results of the five trained candidate lung cancer early screening models, and the ROC-AUC curve and output result distribution diagram were drawn.

[0201] 2. Experimental Results

[0202] The index results of each trained candidate lung cancer early screening model are shown in Table 10, and the AUC values ​​are shown in Table 11.

[0203] Table 10 Index results of each trained candidate lung cancer early screening model

[0204] Model randomForest gbm glmnet svmLinear svmRadial Sensitivity 0.88 0.89 0.89 0.88 0.86 Specificity 0.60 0.60 0.70 0.57 0.63 Pos Pred Value 0.88 0.89 0.91 0.88 0.89 Neg Pred Value 0.60 0.62 0.66 0.57 0.56 F1 0.88 0.89 0.90 0.88 0.87

[0205] Table 11 AUC values ​​of the trained candidate lung cancer early screening models

[0206] Model randomForest gbm glmnet svmLinear svmRadial AUC 0.82 0.86 0.83 0.82 0.84

[0207] The ROC-AUC curves of the five trained candidate lung cancer early screening models are shown below: Figure 2 As shown; the output distribution diagram of the trained 5 candidate lung cancer early screening models is as follows Figure 3 As shown, A is the output result distribution diagram of the trained random forest, B is the output result distribution diagram of the trained gradient boosting model, C is the output result distribution diagram of the trained generalized linear model, D is the output result distribution diagram of the trained svmLinear, and E is the output result distribution diagram of the trained svmRadial.

[0208] The results showed that among the five trained lung cancer early screening models, the trained generalized linear model (glmnet) had the highest F1 value (≥0.90), indicating that the trained generalized linear model could detect samples more accurately and more accurately identify samples as benign or malignant lung nodules; and the ROC-AUC curves of the five trained lung cancer early screening models showed that the AUC value of the trained generalized linear model (glmnet) was 0.83>0.8, indicating that the prediction results of the model had excellent reliability.

[0209] Figure 3 The results show that the p-value of the output result distribution of the trained random forest is 7.33E-0.8, the p-value of the output result distribution of the trained gradient boosting model is 2.66E-09, the p-value of the output result distribution of the trained generalized linear model is 3.73E-08, the p-value of the output result distribution of the trained svmLinear is 1.06E-07, and the p-value of the output result distribution of the trained svmRadial is 1.27E-08.

[0210] Therefore, the trained generalized linear model (glmnet) is the lung cancer early screening model based on TCR sequencing.

[0211] Example 4 A lung cancer early screening detection device based on TCR sequencing

[0212] 1. Composition of the lung cancer early screening detection device

[0213] A lung cancer early screening detection device based on TCR sequencing, comprising a sequencing module, an analysis module, a prediction module and a result output module;

[0214] The sequencing module is used to perform TCR sequencing on the sample to be tested and obtain the fastq file in the sequencing file;

[0215] The analysis module is used to perform the steps shown in Example 2 on the fastq obtained by the sequencing module to obtain the TCR characteristics of the sample to be tested, and obtain the input characteristics of the sample to be tested after screening;

[0216] The types of input features after screening of the samples to be tested are shown in Table 9;

[0217] The prediction module is the lung cancer early screening model based on TCR sequencing as shown in Example 3, and the input features of the sample to be tested obtained by the analysis module after screening are used as input variables, which are input into the lung cancer early screening model to obtain the prediction score of the sample to be tested;

[0218] The result output module is used to output the prediction score of the sample to be tested obtained by the prediction module;

[0219] The value of the prediction score is [0, 1]. If the prediction score is 0, the probability that the sample to be tested is lung cancer is 0. If the prediction score is 1, the probability that the sample to be tested is lung cancer is 1.

[0220] The cutoff value is 0.5, the prediction score is ≥0.5, and the sample to be tested is judged to be positive (i.e., malignant lung nodules); the output result is <0.5, and the sample to be tested is judged to be negative (i.e., benign lung nodules or healthy people).

[0221] The lung cancer early screening model based on TCR sequencing shown in Example 3 in the prediction module is replaced with the trained random forest shown in Example 3 to obtain a lung cancer early screening detection device 1;

[0222] The lung cancer early screening model based on TCR sequencing shown in Example 3 in the prediction module is replaced with the trained gradient boosting model shown in Example 3 to obtain a lung cancer early screening detection device 2;

[0223] The lung cancer early screening model based on TCR sequencing shown in Example 3 in the prediction module is replaced with the trained svmLinear shown in Example 3 to obtain a lung cancer early screening detection device 3;

[0224] The lung cancer early screening model based on TCR sequencing shown in Example 3 in the prediction module is replaced with the trained svmRadial shown in Example 3 to obtain a lung cancer early screening detection device 4.

[0225] 2. Verification of the lung cancer early screening detection device

[0226] A sample set containing 166 malignant pulmonary nodule samples (positive samples) and 52 benign pulmonary nodule samples (negative samples) is used as an independent validation set, and each sample in the independent validation set is used as a test sample. The lung cancer early screening detection device shown in step one is used for detection to obtain the prediction score of each sample in the independent validation set, and the indicators (sensitivity, specificity, pos pred value, neg pred value, F1 value) of the lung cancer early screening detection device are calculated. Based on the prediction score of each sample in the independent validation set, the AUC value is calculated, and the ROC-AUC curve and the output result distribution diagram are drawn.

[0227] According to the above method, the lung cancer early screening detection device 1 to the lung cancer early screening detection device 4 shown in step 1 are respectively verified using an independent verification set and the results are recorded.

[0228] 3. Experimental Results

[0229] The index results of the lung cancer early screening detection device are shown in Table 12, and the AUC values ​​are shown in Table 13.

[0230] Table 12 Index results of lung cancer early screening detection device

[0231]

[0232] Table 13 AUC values ​​of lung cancer early screening detection device

[0233]

[0234] The ROC-AUC curve of the lung cancer early screening detection device is as follows Figure 4 As shown; the output result distribution diagram of the lung cancer early screening detection device is as follows Figure 5 As shown, A is the output result distribution diagram of the lung cancer early screening detection device 1, B is the output result distribution diagram of the lung cancer early screening detection device 2, C is the output result distribution diagram of the lung cancer early screening detection device, D is the output result distribution diagram of the lung cancer early screening detection device 3, and E is the output result distribution diagram of the lung cancer early screening detection device 4.

[0235] The results showed that the F1 value of the lung cancer early screening detection device was 0.87 and the AUC value was 0.84, indicating that the lung cancer early screening detection device can accurately detect lung cancer in the tested samples.

[0236] Figure 5The results show that the p-value of the output result of the early screening detection device for lung cancer 1 is 8.36E-11, the p-value of the output result distribution of the early screening detection device for lung cancer 2 is 1.80E-12, the p-value of the output result distribution of the early screening detection device for lung cancer is 2.03E-13, the p-value of the output result distribution of the early screening detection device for lung cancer 3 is 7.75E-13, and the p-value of the output result distribution of the early screening detection device for lung cancer 4 is 6.86E-12.

[0237] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit the protection scope of the present invention. For ordinary technicians in this field, other different forms of changes or modifications can be made based on the above descriptions and ideas. It is not necessary and impossible to list all the implementation methods here. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the claims of the present invention.

Claims

1. A method for constructing a lung cancer early screening model based on TCR sequencing, characterized in that: The following steps are involved: S1. Establishment of enriched TCR sequence database: S11. Obtaining an experimental group TCR immune repertoire and a control group TCR immune repertoire, and respectively calculating the number of occurrences of each TCR sequence in the experimental group TCR immune repertoire and the control group TCR immune repertoire; the experimental group contains patient samples, and the control group contains healthy human samples; S12. removing TCR sequences with a number of occurrences less than 5 in the control group TCR immune library obtained in step S11, to obtain a control group TCR sequence data set; Removing low-abundance TCR sequences from the control group TCR sequence data set to obtain a control group enriched TCR sequence database; removing TCR sequences contained in the control group TCR sequence data set from the experimental group TCR immune repertoire obtained in step S11, and removing low-abundance TCR sequences to obtain an experimental group enriched TCR sequence database; S2. Construction of lung cancer early screening model: The machine learning model is cross-validated and trained using the features of the samples in the training set to obtain a lung cancer early screening model; the training set includes positive samples and negative samples, the positive samples are blood and / or tissue samples of malignant lung nodule samples; the negative samples are blood and / or tissue samples of benign lung nodule samples, or blood samples of healthy people; The characteristics include TCR characteristics, clinical characteristics and / or imaging characteristics; the TCR characteristics include: V gene and J gene proportion characteristics, TCR immune repertoire statistical characteristics, convergent sequence characteristics, TCR sequence amino acid proportion characteristics, TCR sequence proportion characteristics of different lengths, TCR sequence proportion characteristics of different frequencies and TCR enrichment sequence characteristics; The method for obtaining the TCR feature is specifically as follows: The TCR sequencing data of each sample in the training set was quality controlled, and then the TCRs aligned with the CDR3 region of TCRβ were converted into clonotypes files to obtain the data to be analyzed for the training set samples; V gene and J gene ratio features: Based on the data to be analyzed of the samples in the training set, the corrected clonecounts of each V gene and J gene in each TCR sequence of the sample are obtained as the V gene and J gene ratio features of the sample; Convergent sequence features: Based on the data to be analyzed of the samples in the training set, TCR sequence reads greater than or equal to 1 to 100, and TCR sequences that do not contain stop codons and / or have frameshift mutations are retained as functional TCR sequences of the samples, and the features of the convergent sequences are calculated; the convergent sequences are TCR sequences with the same amino acid sequence but different nucleotide sequences; The characteristics of convergent sequences are: the proportion of the number of convergent sequences, the proportion of clonecounts of convergent sequences, and the proportion of clonecounts after correction of convergent sequences; TCR enrichment sequence features: (1) For the TCR sequences with the same length in the training set sample data to be analyzed and the control group enrichment TCR sequence database in step S1, retain the TCR sequence pairs with an edit distance ≤ 0.3 × b, where b is the TCR sequence length, and select the TCR sequence pairs with alignment_score ≥ 4.5 × b as the TCR sequence combination; Based on the TCR sequence combination, the TCR enrichment sequence feature enrich_score1 is calculated according to formula II, and the TCR enrichment sequence feature enrich_type1 is calculated according to formula III; Formula II: Among them, alignment_score ij is the alignment_score of TCR sequence i and TCR sequence j of the same length; n ki is the number of TCR sequences i in sample k, sample k belongs to The samples in Enrich the TCR sequence database for the control group; query_seq_score j =lm(n j ), n j is the number of TCR sequence j in the TCR sequence combination, and TCR sequence j comes from the data to be analyzed of the sample; query_seqs is the number of TCR sequences in the data to be analyzed of the sample; Formula III: Where enriched_seqs is the number of TCR sequences from the sample in the TCR sequence combination; query_seqs is the number of TCR sequences in the sample data to be analyzed (2) According to the method shown in (1), based on the sample to be analyzed and the enriched TCR sequence database of the experimental group shown in step S1, the TCR enrichment features enrich_score2 and enrich_type2 are calculated.

2. The construction method according to claim 1, characterized in that: In step S2, k-fold cross validation and / or leave-one-out cross validation are used to perform cross validation training on the machine learning model.

3. The construction method according to claim 1, characterized in that: The clinical characteristics described in step S2 include whether there is a history of smoking, whether to quit smoking, whether to have lung cancer, whether to have COPD and / or whether there is a family history of lung cancer; the imaging characteristics include the size of the lung nodules, the number of lung nodules, whether it is a solid nodule, whether it is a partially solid nodule, whether it is a ground glass nodule, whether there are burrs and / or the location of the lung nodules.

4. The construction method according to claim 1, characterized in that: Before cross-training in step S2, the features of the samples in the training set are screened using Ensemble Feature selection and / or Boruta feature screening methods.

5. The construction method according to claim 1, characterized in that: The Ensemble Feature selection feature screening method includes recursive feature elimination, random forest and / or glmnet.

6. The construction method according to claim 4, characterized in that: After screening the features of the samples in the training set, data balancing is performed.

7. The construction method according to claim 1, characterized in that: The statistical characteristics of the TCR immune repertoire in step S2 include the Shannon index, d50Index index, reverse Simpson index and homogeneity index.

8. The construction method according to claim 1, characterized in that: The method for obtaining the amino acid proportion feature of the TCR sequence in step S2 is: based on the functional TCR sequence of the sample, 1 to 4 amino acids at both ends of each TCR sequence are removed, and the proportion of each amino acid in the TCR sequence after the amino acids are deleted is calculated.

9. The construction method according to claim 1, characterized in that: The method for obtaining the characteristics of the proportion of TCR sequences of different lengths and the characteristics of the proportion of TCR sequences of different frequencies in step S2 is: based on the functional TCR sequence of the sample, it is grouped by length, and the quantitative ratio of TCR sequences in each length group is calculated to obtain the proportion of TCR sequences of each length; based on the functional TCR sequence of the sample, it is grouped by frequency, and the quantitative ratio of TCR sequences in each frequency group is calculated to obtain the proportion of TCR sequences of each frequency.

10. A lung cancer early screening detection device based on TCR sequencing, characterized in that: It includes sequencing module, analysis module, prediction module and result output module; The sequencing module is used to perform TCR sequencing on the sample to be tested and obtain the fastq file in the sequencing file; The analysis module performs the step S2 in the construction method of claim 1 on the fastq file obtained by the sequencing module to obtain the TCR characteristics of the sample to be tested, and obtains the input characteristics after screening of the sample to be tested; The prediction module is a lung cancer early screening model constructed by the construction method of claim 1, and the input features of the sample to be tested obtained by the analysis module after screening are used as input variables, which are input into the lung cancer early screening model to obtain the lung cancer prediction score of the sample to be tested; The result output module is used to output the lung cancer prediction score of the sample to be tested obtained by the prediction module.

Images