TsRNA biomarker combination for diagnosing or aiding in the diagnosis of gastric cancer and application thereof

Abstract

The application discloses a tsRNA biomarker combination for diagnosing or assisting in diagnosing gastric cancer, and provides the tsRNA biomarker combination as a gastric cancer diagnosis marker, which has good specificity, can accurately identify gastric cancer patients, has high sensitivity, can early detect cancer patients and has high diagnosis efficiency, provides a new clinical means for diagnosing gastric cancer, has a good translational medicine prospect, and the application constructs a gastric cancer diagnosis model based on the tsRNA biomarker, the model can be used for early screening and auxiliary diagnosis of gastric cancer, can provide accurate evaluation on the risk of gastric cancer, and thus promotes early treatment and management decision.

Description

Technical Field

[0001] This invention relates to the field of biomedicine, and more particularly to a combination of tsRNA biomarkers for the diagnosis or auxiliary diagnosis of gastric cancer and their applications. Background Technology

[0002] According to data from the World Health Organization, in 2023, my country had approximately 503,000 new cases of gastric cancer, ranking fourth in incidence among malignant tumors, and 389,000 deaths, ranking third in mortality. my country accounts for approximately 44.8% of global gastric cancer cases and 49.2% of global deaths. Early symptoms of gastric cancer are often subtle, leading to a low diagnosis rate and a high proportion of advanced-stage cancer. The 5-year relative survival rate for gastric cancer patients in my country is approximately 35.9%, significantly lower than in Japan (60.3%) and South Korea (77.0%). This is mainly because most patients are diagnosed at an advanced stage, resulting in limited effectiveness of conventional treatments and high rates of metastasis and recurrence. Therefore, early diagnosis and prevention of gastric cancer remain a major challenge in the control of malignant tumors in my country.

[0003] Currently, clinical guidelines recommend gastric cancer screening methods including serum tumor marker testing, imaging examinations (such as barium meal X-ray, CT, MRI, etc.), and endoscopy and tissue biopsy. These methods play an important role in the diagnosis of gastric cancer, but each has certain limitations. Traditional serological markers, such as carcinoembryonic antigen (CEA) and carbohydrate antigen 19-9 (CA19-9), although widely used in the auxiliary diagnosis and efficacy monitoring of gastric cancer, have low sensitivity and specificity for early gastric cancer. For example, studies have shown that the sensitivities of CEA (cutoff value 3 ng / mL) and CA19-9 (cutoff value 37 U / mL) for the diagnosis of gastric cancer are 20%-30% and 30%-40%, respectively, and the specificities are 70%-80% and 80%-90%, respectively. While imaging examinations and endoscopy can detect gastric lesions relatively accurately, the former carries the risk of radiation exposure, and the latter is an invasive procedure with poor patient compliance and high cost, making it difficult to use as a routine means for large-scale population screening. Therefore, in order to meet clinical needs, there is an urgent need to develop a new method for the early diagnosis of gastric cancer that is highly sensitive, specific, non-invasive, convenient, and economical.

[0004] Liquid biopsy is a breakthrough technology for tumor screening and diagnosis. It has unique advantages such as being non-invasive, providing real-time dynamic detection, and overcoming tumor heterogeneity. It enables gastric cancer patients to receive treatment early, achieve radical treatment through surgery, improve patient survival rates, and reduce treatment costs, which has significant clinical significance and social benefits.

[0005] Liquid biopsy, as an emerging tumor screening and diagnostic technology, has advantages such as being non-invasive, repeatable, allowing for real-time dynamic monitoring, and overcoming tumor heterogeneity. It has shown great application potential in early tumor diagnosis, efficacy evaluation, and prognostic monitoring.

[0006] In recent years, various liquid biopsy biopsy markers, such as circulating tumor cells (CTCs), circulating tumor DNA (ctDNA), and exosomes, have been used in gastric cancer research. Among them, circulating small non-coding RNAs (sncRNAs) have attracted much attention due to their stability in the blood. sncRNAs, including microRNAs (miRNAs), have been shown to be closely related to tumor development and progression, and have the potential to become novel tumor markers. tRNA-derived small RNAs (tsRNAs) are a class of non-coding RNA molecules with a length of 14-30 nucleotides, produced by the cleavage of mature tRNA or tRNA precursors by specific nucleases (such as Dicer and RNase Z). Studies have shown that tsRNAs are not random degradation products of tRNA, but functional molecules that play important regulatory roles in cells, participating in various biological processes such as gene expression regulation, cell proliferation, apoptosis, and stress responses. More importantly, studies have found that tsRNAs are abnormally expressed in various tumors (including gastric cancer), and this abnormal expression often precedes the appearance of clinical symptoms of the tumor. tsRNA exhibits good stability in body fluids (such as blood and urine), is not easily degraded by RNases, can be accurately quantified using techniques such as qRT-PCR, and has a wide variety of types with tissue- and disease-specific expression patterns. These characteristics make tsRNA a highly promising non-invasive tumor biomarker.

[0007] In view of this, the present invention is proposed: a combination of tsRNA biopsy markers based on liquid biopsy for early screening of gastric cancer in the Chinese population. Summary of the Invention

[0008] The purpose of this invention is to provide a combination of tsRNA biomarkers for the diagnosis or auxiliary diagnosis of gastric cancer and their applications, so as to solve the problem of facilitating early screening of gastric cancer mentioned in the background art.

[0009] To achieve the above objectives, the present invention provides the following technical solution: A combination of tsRNA biomarkers for the diagnosis or auxiliary diagnosis of gastric cancer, wherein the tsRNA molecules comprising the combination of said tsRNA biomarkers include one or more: mature-tRNA-Arg-TCG_5_end, mature-tRNA-Gly-GCC_CCA_end, mature-tRNA-Gly-TCC_5_end, mature-tRNA-Pro-TGG_CCA_end, mature-tRNA-Ser-GCT_5_end.

[0010] The application of the combination of target tsRNA biomarkers in the detection sample in the preparation of a kit for the early diagnosis or auxiliary diagnosis of gastric cancer, wherein the nucleotide sequences of the tsRNA molecules in the combination of tsRNA biomarkers are shown in SEQ ID No. 1-5.

[0011] Preferably, the reagents for detecting the target tsRNA in the sample include primers and probes for detecting the target tsRNA.

[0012] A kit for diagnosing or assisting in the diagnosis of gastric cancer, the kit comprising primers and probes for detecting target tsRNA; the nucleotide sequences of the tsRNA molecules in the combination of the tsRNA biomarkers are shown sequentially in SEQ ID No. 1-5.

[0013] A method for gastric cancer tumor screening and risk calculation includes the following steps: S1. Blood sample collection and processing; S2, Small RNA sequencing library construction and tsRNA quantification; S3, differential analysis and classification performance evaluation and screening of characteristic tsRNAs; S4. Construct a stacked integration model; S5, independent validation set performance verification and risk calculation.

[0014] As a preferred option, step S1 is divided into two stages: CF1 stage and CF2 stage, which respectively obtain the experimental set and the independent validation set. The CF1 stage: Blood samples were collected using EDTA anticoagulant tubes, plasma was separated by centrifugation, aliquoted into RNase-free cryopreservation tubes, and stored at low temperature as an experimental set until RNA extraction. In the CF2 stage: blood samples were collected using the same method as an independent validation set and stored until RNA extraction.

[0015] Preferably, step S2 includes the following steps: S21. The 3' end of RNA is ligated using a 5'-adenosylated / 3'-blocked single-stranded DNA adapter. An RT primer containing UMI is introduced to bind the residual adapter, followed by 5' end ligation. RNA is then extracted from the blood sample in step S1. S22. The extracted RNA was reverse transcribed into cDNA using a high-sensitivity polymerase and amplified. The 110-140bp PCR product was recovered by PAGE gel. S23. Further circularize into a single-stranded DNA library, and after quality control, perform single-end 50bp high-throughput sequencing based on the DNBSEQ platform; S24. After sequencing, the quality control process filters out low-quality reads, trims adapter sequences, filters reads that are not 15-44bp in length, removes all-PolyA sequences, limits high adenine content, and removes reads containing non-standard bases, and outputs the sequence data. S25. Quantify tsRNA using the SPORTS tool based on the output sequence data; S26. The quantified tsRNA is preferentially mapped to the GtRNAdb database via Bowtie to screen out reads that match the tRNA sequence. S27. The matched reads of the screened tRNA sequences are analyzed by the tsRNA cleavage site to distinguish whether they originate from the 5' end, 3' end, or 3' CCA end of the tRNA. Based on the number of matched reads, the expression abundance of each tsRNA is calculated, and the feature set of tsRNAs with an average expression level greater than 5 is retained.

[0016] Preferably, step S3 includes the following steps: S31. Differentially expressed tsRNAs were screened based on empirical Bayesian methods, p-values ​​were corrected using the Benjamini-Hochberg method, and candidate biomarkers were screened using the following criteria: AUC>0.75, |LogFC|>1, and average expression level>3CPM. Lasso regression was used to further screen for characteristic tsRNAs, and the final tsRNA characteristic markers were obtained.

[0017] Preferably, in step S4, the construction of the stacked ensemble model involves first generating a prediction probability matrix based on the base learner, and then using a generalized linear model as a meta-learner to integrate the classification output of the base model through a weighted fusion strategy.

[0018] Preferably, in step S5, the performance verification of the independent validation set is performed using the CF2 stage independent sample queue for blind verification, which comprehensively evaluates the detection performance of the ensemble model for gastric cancer.

[0019] The beneficial effects of this invention are: 1. The tsRNA biomarker combination provided by this invention, as a diagnostic biomarker for gastric cancer, has good specificity and can accurately identify gastric cancer patients; it also has high sensitivity, can detect cancer patients at an early stage, and has high diagnostic efficiency, providing a new clinical means for diagnosing gastric cancer and has good prospects for translational medicine.

[0020] 2. This invention constructs a gastric cancer diagnostic model based on tsRNA biomarkers. This model can be used for early screening and auxiliary diagnosis of gastric cancer, and can provide accurate assessment of gastric cancer risk, thereby promoting early treatment and management decisions. Attached Figure Description

[0021] Figure 1 This is a schematic diagram illustrating the expression of tsRNA biomarkers in CF1 samples according to an embodiment of the present invention. Figure 2 This is a schematic diagram of the performance of the tsRNA biomarker gastric cancer diagnostic model based on ensemble learning in the cross-validation set according to an embodiment of the present invention. Figure 3 This is a schematic diagram of the performance of the gastric cancer diagnostic model based on tsRNA biomarkers according to an embodiment of the present invention on an independent validation set. Figure 4 This is a schematic diagram of the risk score distribution of the model in an independent validation set according to an embodiment of the present invention. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] Please see Figures 1-4 This invention provides a combination of tsRNA biomarkers for the diagnosis or auxiliary diagnosis of gastric cancer and their applications.

[0024] Example 1: A combination of tsRNA biomarkers for diagnosing or assisting in the diagnosis of gastric cancer, wherein the tsRNA molecules comprising the combination of said tsRNA biomarkers include one or more: mature-tRNA-Arg-TCG_5_end, mature-tRNA-Gly-GCC_CCA_end, mature-tRNA-Gly-TCC_5_end, mature-tRNA-Pro-TGG_CCA_end, mature-tRNA-Ser-GCT_5_end.

[0025] The tsRNA biomarker combination includes, but is not limited to, the following specific tsRNA molecules: mature-tRNA-Arg-TCG_5_end (sequence: GACCGCGTGGCCTAATGGA) mature-tRNA-Gly-GCC_CCA_end (sequence: TCGATTCCCGGCCAATGCACCA) mature-tRNA-Gly-TCC_5_end (sequence: GCGTTGGTGGTATAGTGGTTAGCATA) mature-tRNA-Pro-TGG_CCA_end (sequence: GCGTTGGTGGTATAGTGGTTAGCATA) mature-tRNA-Ser-GCT_5_end(sequence:GACGAGGTGGCCGAGTGGTTAAGG) Example 2: Application of the combination of target tsRNA biomarkers in the sample in the preparation of a kit for early diagnosis or auxiliary diagnosis of gastric cancer, wherein the nucleotide sequences of the tsRNA molecules in the combination of tsRNA biomarkers are shown in SEQ ID No. 1-5.

[0026] The tsRNA nucleotide sequences of the tsRNA biomarker combination are shown in SEQ ID No. 1-5 as follows: mature-tRNA-Arg-TCG_5_end (sequence: GACCGCGTGGCCTAATGGA) mature-tRNA-Gly-GCC_CCA_end (sequence: TCGATTCCCGGCCAATGCACCA) mature-tRNA-Gly-TCC_5_end (sequence: GCGTTGGTGGTATAGTGGTTAGCATA) mature-tRNA-Pro-TGG_CCA_end (sequence: GCGTTGGTGGTATAGTGGTTAGCATA) mature-tRNA-Ser-GCT_5_end(sequence:GACGAGGTGGCCGAGTGGTTAAGG) Specifically, the reagents for detecting the target tsRNA in the sample include primers and probes for detecting the target tsRNA.

[0027] Example 3: A kit for diagnosing or assisting in the diagnosis of gastric cancer, the kit comprising primers and probes for detecting target tsRNA; the nucleotide sequences of the tsRNA molecules in the combination of the tsRNA biomarkers are shown sequentially as SEQ ID No. 1-5.

[0028] Example 4: A method for gastric cancer tumor screening and risk calculation, comprising the following steps: S1. Blood sample collection and processing; S2, Small RNA sequencing library construction and tsRNA quantification; S3, differential analysis and classification performance evaluation and screening of characteristic tsRNAs; S4. Construct a stacked integration model; S5, independent validation set performance verification and risk calculation.

[0029] Specifically, step S1 is divided into two stages: CF1 and CF2, which respectively yield the experimental set and the independent validation set. The CF1 stage: Blood samples were collected using EDTA anticoagulant tubes, and plasma was separated by centrifugation within 1 hour. The plasma was then aliquoted into RNase-free cryovials and stored at low temperature as an experimental set until RNA extraction. In the CF2 stage: blood samples were collected using the same method as an independent validation set and stored until RNA extraction.

[0030] The sample collection and processing include: After approval by the Ethics Committee of the First Affiliated Hospital of Wenzhou Medical University and obtaining informed consent from the subjects, clinical samples were collected in two phases (CF1 and CF2).

[0031] Phase CF1: Peripheral blood samples were collected preoperatively from 30 patients with pathologically confirmed gastric cancer and from 30 age / sex-matched healthy controls (without a history of gastric disease or malignancy). Blood samples were collected using EDTA anticoagulant tubes, centrifuged at 12,000g for 10 minutes within 1 hour to separate plasma, aliquoted into RNase-free cryovials, and stored at -80°C until RNA extraction.

[0032] CF2 phase: Blood samples were collected from 30 gastric cancer patients and 30 healthy controls using the same method to serve as an independent validation set.

[0033] Specifically, step S2 includes the following steps: S21. The 3' end of RNA is ligated using a 5'-adenosylated / 3'-blocked single-stranded DNA adapter. An RT primer containing UMI is introduced to bind the residual adapter, followed by 5' end ligation. RNA is then extracted from the blood sample in step S1. S22. The extracted RNA was reverse transcribed into cDNA using a high-sensitivity polymerase and amplified. The 110-140bp PCR product was recovered by PAGE gel. S23. Further circularize into a single-stranded DNA library, and after quality control, perform single-end 50bp high-throughput sequencing based on the DNBSEQ platform; S24. After sequencing, the quality control process filters out low-quality reads, trims adapter sequences, filters reads that are not 15-44bp in length, removes all-PolyA sequences, limits high adenine content, and removes reads containing non-standard bases, and outputs the sequence data. S25. Quantify tsRNA using the SPORTS tool based on the output sequence data; S26. The quantified tsRNA is preferentially mapped to the GtRNAdb database via Bowtie to screen out reads that match the tRNA sequence. S27. The matched reads of the screened tRNA sequences are analyzed by the tsRNA cleavage site to distinguish whether they originate from the 5' end, 3' end, or 3' CCA end of the tRNA. Based on the number of matched reads, the expression abundance of each tsRNA is calculated, and the feature set of tsRNAs with an average expression level greater than 5 is retained.

[0034] The small RNA sequencing library was constructed as follows: Total RNA was extracted from the first-stage sample (CF1) and the second-stage sample (CF2). The 3' end of the RNA was ligated using a 5'-adenosylation / 3'-blocking single-stranded DNA adapter. RT primers containing UMI were introduced to bind the residual adapter, followed by 5' end ligation. cDNA was synthesized and amplified using high-sensitivity polymerase reverse transcription. The 110-140bp PCR product was recovered by PAGE gel. The cDNA was further circularized into a single-stranded DNA library, and after quality control, 50bp single-end high-throughput sequencing was performed on the DNBSEQ platform.

[0035] The tsRNA quantification: High-quality data output was ensured by filtering low-quality reads (Phred<13 base percentage >0.1%), trimming adapter sequences, filtering reads longer than 15-44 bp, removing all-PolyA sequences, limiting high adenine content, and eliminating reads containing non-standard bases through quality control. The SPORTS tool was used to quantify the tsRNAs of the preprocessed high-quality reads (as shown in Table 1). Reads matching the tRNA sequence were selected by preferential mapping to the GtRNAdb database via Bowtie. The tsRNA cleavage sites were analyzed to accurately distinguish whether they originated from the 5', 3', or 3' CCA ends of the tRNA. Based on the number of matching reads, the expression abundance of each tsRNA was calculated, retaining features with an average expression level (represented by read count) greater than 5. Features with an average expression level below this threshold were generally considered to have high sequencing noise or low biological significance and were therefore discarded.

[0036] Table 1. tsRNA quantification results of the two-stage samples Specifically, step S3 includes the following steps: S31. Differentially expressed tsRNAs were screened based on empirical Bayesian methods, p-values ​​were corrected using the Benjamini-Hochberg method, and candidate biomarkers were screened using the following criteria: AUC>0.75, |LogFC|>1, and average expression level>3CPM. Lasso regression was used to further screen for characteristic tsRNAs, and the final tsRNA characteristic markers were obtained.

[0037] Among them, the difference analysis and classification performance evaluation Differentially expressed tsRNAs were screened using the R language limma package based on an empirical Bayesian method. The p-value was corrected using the Benjamini-Hochberg method (significance threshold p < 0.05), and candidate biomarkers were selected based on the following criteria: AUC > 0.75, |LogFC| > 1, and mean expression level > 3 CPM. The ROC curve AUC value and 95% confidence interval (1000 bootstrap resampling) were calculated using the pROC package. The optimal threshold was determined using the Youden index, and sensitivity, specificity, positive / negative predictive value, precision, and accuracy were calculated simultaneously (as shown in Table 2).

[0038] The characteristic tsRNA screening Lasso regression was used to further screen for feature tsRNAs. Expression profiles of differentially expressed tsRNAs were used as independent variables, and the gastric cancer group vs. control group were used as dependent variables. Five-fold cross-validation was employed to select the optimal penalty parameter λ. A Lasso regression model was constructed based on the optimal λ value to obtain the regression coefficient for each tsRNA. tsRNAs with non-zero regression coefficients were selected as the final feature tsRNAs. To ensure stability, the screening was repeated 100 times, and tsRNAs selected more than 95% of the time were retained as feature biomarkers. The performance of the ensemble learning-based tsRNA biomarker gastric cancer diagnostic model on the cross-validation set is as follows: Figure 1 As shown.

[0039] Table 2. Summary of the classification performance of tsRNA biomarkers Specifically, in step S4, the ensemble model is constructed by first generating a prediction probability matrix based on the base learner, and then using a generalized linear model as a meta-learner to integrate the classification output of the base model through a weighted fusion strategy.

[0040] Among them, ensemble model construction To improve the generalization ability and prediction accuracy of the gastric cancer diagnostic model, Random Forest, Naïve Bayes, XGBoost, Deep Neural Network, and Logistic Regression were used as base learners, and the hyperparameters of each model were optimized through five-fold cross-validation. Based on this, a stacked ensemble model was constructed: first, a prediction probability matrix was generated based on the base learners; then, a generalized linear model (GLM) was used as a meta-learner, and the classification outputs of the base models were integrated through a weighted fusion strategy, significantly improving the model's decision robustness and clinical interpretability. Results showed that the ensemble model achieved an AUC of 0.999 (95% CI: 0.995-1.000) in cross-validation, a 2.5% improvement over the best-performing single base model (AUC=0.974). The performance of the tsRNA biomarker-based gastric cancer diagnostic model based on ensemble learning in the cross-validation set is shown below. Figure 2 As shown.

[0041] Specifically, in step S5, the performance verification of the independent validation set is performed by blind verification using the CF2 stage independent sample queue to comprehensively evaluate the detection performance of the ensemble model for gastric cancer.

[0042] Among them, independent validation set performance verification To validate the clinical applicability of the model, a blinded validation process was conducted using a phase 2 (CF2) independent sample cohort to comprehensively evaluate the ensemble model's performance in detecting gastric cancer. For example... Figure 4 As shown, the ensemble model achieved an AUC of 0.795 (95% CI: 0.673–0.917) on the independent validation set, higher than any single model. This result demonstrates that the ensemble model, through multi-algorithm collaborative optimization, effectively overcomes the performance degradation problem of single models on cross-cohort data, exhibiting high robustness in distinguishing gastric cancer patients from healthy individuals in real-world clinical scenarios, and possessing the potential for large-scale screening applications. The performance of the tsRNA biomarker-based gastric cancer diagnostic model based on ensemble learning on the independent validation set is shown below. Figure 3 As shown.

[0043] It should be noted that the tsRNA biomarker combinations of the present invention include, but are not limited to, the following specific tsRNA molecules: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions or improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A combination of tsRNA biomarkers for the diagnosis or auxiliary diagnosis of gastric cancer, characterized in that: The tsRNA molecules in the combination of tsRNA biomarkers include one or more: mature-tRNA-Arg-TCG_5_end, mature-tRNA-Gly-GCC_CCA_end, mature-tRNA-Gly-TCC_5_end, mature-tRNA-Pro-TGG_CCA_end, mature-tRNA-Ser-GCT_5_end.

2. The application of a combination of target tsRNA biomarkers in a sample in the preparation of a kit for the early diagnosis or auxiliary diagnosis of gastric cancer, characterized in that: The tsRNA nucleotide sequences of the tsRNA biomarker combination are shown in SEQ ID No. 1-5.

3. The application of the combination of target tsRNA biomarkers in the detection sample as described in claim 2 in the preparation of a kit for early diagnosis or auxiliary diagnosis of gastric cancer, characterized in that: The reagents for detecting the target tsRNA in the sample include primers and probes for detecting the target tsRNA.

4. A reagent kit for diagnosing or assisting in the diagnosis of gastric cancer, characterized in that: The kit includes primers and probes for detecting target tsRNA; the nucleotide sequences of the tsRNA molecules in the tsRNA biomarker combination are shown in SEQ ID No. 1-5.

5. A method for gastric cancer tumor screening and risk calculation, characterized in that, Includes the following steps: S1. Blood sample collection and processing; S2, Small RNA sequencing library construction and tsRNA quantification; S3, differential analysis and classification performance evaluation and screening of characteristic tsRNAs; S4. Construct a stacked integration model; S5, independent validation set performance verification and risk calculation.

6. The method for gastric cancer tumor screening and risk calculation according to claim 5, characterized in that: Step S1 is divided into two stages: CF1 and CF2, which yield the experimental set and the independent validation set, respectively. The CF1 stage: Blood samples were collected using EDTA anticoagulant tubes, and plasma was separated by centrifugation. The plasma was then aliquoted into RNase-free cryopreservation tubes and stored at low temperatures as an experimental set until RNA extraction. In the CF2 stage: blood samples were collected using the same method as an independent validation set and stored until RNA extraction.

7. The method for gastric cancer tumor screening and risk calculation according to claim 5, characterized in that: Step S2 includes the following steps: S21. The 3' end of RNA is ligated using a 5'-adenosylated / 3'-blocked single-stranded DNA adapter. An RT primer containing UMI is introduced to bind the residual adapter, followed by 5' end ligation. RNA is then extracted from the blood sample in step S1. S22. The extracted RNA was reverse transcribed into cDNA using a high-sensitivity polymerase and amplified. The 110-140bp PCR product was recovered by PAGE gel. S23. Further circularize into a single-stranded DNA library, and after quality control, perform single-end 50bp high-throughput sequencing based on the DNBSEQ platform; S24. After sequencing, the quality control process filters out low-quality reads, trims adapter sequences, filters reads that are not 15-44bp in length, removes all-PolyA sequences, limits high adenine content, and removes reads containing non-standard bases, and outputs the sequence data. S25. Quantify tsRNA using the SPORTS tool based on the output sequence data; S26. The quantified tsRNA is preferentially mapped to the GtRNAdb database via Bowtie to screen out reads that match the tRNA sequence. S27. The matched reads of the screened tRNA sequences are analyzed by the tsRNA cleavage site to distinguish whether they originate from the 5' end, 3' end, or 3' CCA end of the tRNA. Based on the number of matched reads, the expression abundance of each tsRNA is calculated, and the feature with an average expression level greater than 5 is retained to obtain the tsRNA feature set.

8. The method for gastric cancer tumor screening and risk calculation according to claim 5, characterized in that: Step S3 includes the following steps: S31. Differentially expressed tsRNAs were screened based on empirical Bayesian methods, p-values ​​were corrected using the Benjamini-Hochberg method, and candidate biomarkers were screened using the following criteria: AUC>0.75, |LogFC|>1, and average expression level>3CPM. Lasso regression was used to further screen for characteristic tsRNAs, and the final tsRNA characteristic markers were obtained.

9. The method for gastric cancer tumor screening and risk calculation according to claim 5, characterized in that: In step S4, the construction of the stacked ensemble model first generates a prediction probability matrix based on the base learner, and then uses the generalized linear model as the meta-learner to integrate the classification output of the base model through a weighted fusion strategy.

10. The method for gastric cancer tumor screening and risk calculation according to claim 6, characterized in that: In step S5, the performance verification of the independent validation set is performed by blind verification using the CF2 stage independent sample queue to comprehensively evaluate the detection performance of the ensemble model for gastric cancer.

Images