Kit and use thereof

Abstract

Disclosed in the present invention are a kit and use thereof. The kit comprises a detection reagent for determining an expression level of one or more RNA molecules, which can achieve accurate diagnosis of endometriosis. In addition, also disclosed in the present invention is a method for diagnosing endometriosis.

Description

A reagent kit and its application Technical Field

[0001] This invention belongs to the fields of molecular biology and medical diagnostics, specifically relating to the application of one or more RNA molecules in human serum in the preparation of a kit for detecting or diagnosing endometriosis. Background Technology

[0002] Endometriosis is a common disease defined as the implantation and persistent presence of endometrial tissue outside the uterine cavity. Its incidence rate in women of reproductive age can be as high as 10-15%, with approximately 50% of infertile patients having endometriosis. 70%–80% of endometriosis patients experience varying degrees of pelvic pain symptoms, including dysmenorrhea, chronic pelvic pain, and dyspareunia, significantly impacting their quality of life. Currently, clinical diagnosis of endometriosis primarily relies on physical examination, imaging studies, and laparoscopy. However, the results of gynecological examinations are greatly influenced by the physician's experience and skill; the sensitivity of imaging examinations varies depending on the location of the endometrial lesions; and while laparoscopy is an invasive procedure, although considered minimally invasive, it causes significant damage to the endometrium. Elevated CA125 levels, commonly used in clinical gynecology, are more frequently seen in patients with severe endometriosis, significant pelvic inflammation, ruptured endometriotic cysts, or adenomyosis. However, its specificity for endometriosis detection is poor, rendering it of limited diagnostic value. Therefore, identifying molecular markers for endometriosis and developing them into early diagnostic kits would undoubtedly have significant clinical and social value for patients, enabling accurate diagnosis of endometriosis in its early stages via liquid biopsy.

[0003] To date, research on the relationship between non-coding RNA and the early diagnosis of endometriosis is constantly evolving. However, current endometriosis diagnostic kits suffer from low accuracy and precision. Furthermore, the poor sensitivity and specificity of current endometriosis diagnostic kits also urgently need to be addressed. Summary of the Invention

[0004] The first aspect of this invention aims to provide a reagent kit.

[0005] The second aspect of this invention aims to provide an application of a reagent kit in the diagnosis of endometriosis.

[0006] A third aspect of the present invention aims to provide a method for diagnosing endometriosis, the method comprising the following steps:

[0007] (1) Collect human serum samples;

[0008] (2) Detect the expression levels of one or more RNA molecules as described in claim 1 in serum samples;

[0009] (3) Input the expression level of RNA molecules in step (2) into the trained machine learning model, output the model evaluation results, and determine whether the patient has endometriosis.

[0010] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0011] This invention provides 10 RNA molecules, namely tRF-59:75-Ala-AGC-2-M4, tRF-1:16-Leu-AAG-4-M2, tRF-1:16-Leu-AAG-1-M4, tRF-+1:T33-Gln-CTG-3, tiRNA-1:31-chrM.Met-CAT, tRF-1:16-Gly-CCC-2, tiRNA-1:34-Val-AAC-1-M3, tRF-54:75-His-GTG-1-M2, tRF-58:75-Ala-AGC-1, and tRF-1:28-chrM.Ser-TGA. The nucleotide sequence of tRF-59:75-Ala-AGC-2-M4 is shown in SEQ ID NO.1; the nucleotide sequence of tRF-1:16-Leu-AAG-4-M2 is shown in SEQ ID NO.2; the nucleotide sequence of tRF-1:16-Leu-AAG-1-M4 is shown in SEQ ID NO.3; the nucleotide sequence of tRF-+1:T33-Gln-CTG-3 is shown in SEQ ID NO.4; the nucleotide sequence of tiRNA-1:31-chrM.Met-CAT is shown in SEQ ID NO.5; the nucleotide sequence of tRF-1:16-Gly-CCC-2 is shown in SEQ ID NO.6; the nucleotide sequence of tiRNA-1:34-Val-AAC-1-M3 is shown in SEQ ID NO.7; the nucleotide sequence of tRF-54:75-His-GTG-1-M2 is shown in SEQ ID NO. As shown in NO.8; the nucleotide sequence of tRF-58:75-Ala-AGC-1 is shown in SEQ ID NO.9; the nucleotide sequence of tRF-1:28-chrM.Ser-TGA is shown in SEQ ID NO.10.

[0012] In a first aspect, the present invention provides a kit for detecting the expression levels of one or more of the said RNA molecules.

[0013] In some embodiments of the present invention, the nucleotide sequence of the RNA molecule is SEQ ID NO.1, SEQ ID NO., SEQ ID NO.3, SEQ ID NO.4, SEQ ID NO.5, SEQ ID NO.6, SEQ ID NO.7, SEQ ID NO.8, SEQ ID NO.9 or SEQ ID NO.10.

[0014] In some embodiments of the present invention, the kit is used to diagnose endometriosis, and the kit contains a detection reagent for detecting the expression level of RNA molecules in serum.

[0015] In some embodiments of the present invention, the detection reagents include reverse transcription reagents and quantitative real-time PCR reaction reagents.

[0016] In some embodiments of the present invention, the reverse transcription reagent includes 5x RT buffer, dNTP mixture, RNase inhibitor, and specific reverse transcription primers.

[0017] In some embodiments of the present invention, the real-time PCR reaction reagent includes specific primer pairs, PCR amplification buffer, and DNA polymerase mixture.

[0018] The second aspect of this invention aims to provide an application of a reagent kit in the diagnosis of endometriosis.

[0019] A third aspect of the present invention is to provide a method for diagnosing endometriosis, the method comprising the following steps:

[0020] (1) Collect human serum samples;

[0021] (2) Detect the expression levels of one or more RNA molecules as described in claim 1 in serum samples;

[0022] (3) Input the expression level of RNA molecules in step (2) into the trained machine learning model, output the model evaluation results, and determine whether the patient has endometriosis.

[0023] In some embodiments of the present invention, Trizol reagent can be used to extract RNA molecules from human serum samples.

[0024] In some embodiments of the present invention, cDNA is obtained by reverse transcription of RNA, and the expression level of RNA molecules is obtained by quantitative real-time PCR.

[0025] In some embodiments of the present invention, the reverse transcription reagent includes 5x RT buffer, dNTP mixture, RNase inhibitor, and specific reverse transcription primers.

[0026] In some embodiments of the present invention, the reverse transcription reaction step is incubation at 42°C for 60 minutes and incubation at 70°C for 10 minutes.

[0027] In some embodiments of the present invention, a real-time PCR instrument can be used to perform quantitative PCR reactions.

[0028] In some embodiments of the present invention, the real-time PCR reaction reagent includes specific primer pairs, a real-time PCR amplification buffer, and a DNA polymerase mixture.

[0029] In some embodiments of the present invention, one or more RNA molecules as described in claim 1 and the external reference RNA molecule cel-miR-39-3p in serum samples are detected by real-time fluorescence quantitative PCR, and the number of cycles (Cycle Threshold, or CT value for short) when the fluorescence signal during the exponential growth phase of PCR amplification reaches a set threshold is recorded.

[0030] In some embodiments of the present invention, the average cycle number of RNA molecules and the external reference cel-miR-39-3p is calculated, and the relative expression level of RNA molecules, i.e. the difference between the average cycle number of RNA molecules and cel-miR-39-3p, is calculated. The calculation formula is: ΔCT = CT RNA molecules - CT cel-miR-39-3p;

[0031] The expression level of RNA molecules is then input into a trained machine learning model, which outputs the model evaluation results to determine whether the patient has endometriosis.

[0032] In some embodiments of the present invention, each RNA molecule is subjected to diagnostic testing separately, and 14 different machine learning models are used for analysis and modeling to evaluate the diagnostic accuracy of each RNA molecule.

[0033] In some embodiments of the present invention, the diagnostic efficacy of ten RNA molecules—tRF-59:75-Ala-AGC-2-M4, tRF-1:16-Leu-AAG-4-M2, tRF-1:16-Leu-AAG-1-M4, tRF-+1:T33-Gln-CTG-3, tiRNA-1:31-chrM.Met-CAT, tRF-1:16-Gly-CCC-2, tiRNA-1:34-Val-AAC-1-M3, tRF-54:75-His-GTG-1-M2, tRF-58:75-Ala-AGC-1, and tRF-1:28-chrM.Ser-TGA—was evaluated.

[0034] In some embodiments of the present invention, two RNA molecules are used for combined diagnostic detection, and 14 different machine learning models are used for analysis and modeling to evaluate the diagnostic accuracy of different combinations. For example, combined detection of tRF-1:16-Leu-AAG-1-M4 and tRF-59:75-Ala-AGC-2-M4; combined detection of tRF-1:28-chrM.Ser-TGA and tRF-1:16-Leu-AAG-4-M2; combined detection of tRF-1:16-Leu-AAG-1-M4 and tRF-1:28-chrM.Ser-TGA; a total of 45 combinations were evaluated for diagnostic efficacy.

[0035] In some embodiments of the present invention, three RNA molecules are used for combined diagnostic detection, and 14 different machine learning models are used for analysis and modeling to evaluate the diagnostic accuracy of different combinations. For example, combined detection of tRF-1:16-Leu-AAG-1-M4, tRF-59:75-Ala-AGC-2-M4, and tRF-1:28-chrM.Ser-TGA; combined detection of tRF-54:75-His-GTG-1-M2, tRF-1:16-Leu-AAG-1-M4, and tRF-59:75-Ala-AGC-2-M4; and combined detection of tRF-58:75-Ala-AGC-1, tRF-1:16-Leu-AAG-1-M44, and tRF-59:75-Ala-AGC-2-M4, for a total of 120 combinations, is evaluated for diagnostic efficacy.

[0036] In some embodiments of the present invention, four RNA molecules are used for joint diagnostic detection, and 14 different machine learning models are used for analysis and modeling to evaluate the diagnostic accuracy of different combinations. For example, combined detection of tRF-58:75-Ala-AGC-1, tRF-1:16-Leu-AAG-1-M4, tRF-59:75-Ala-AGC-2-M4, and tRF-1:28-chrM.Ser-TGA was performed; combined detection of tRF-58:75-Ala-AGC-1, tRF-1:16-Leu-AAG-1-M4, tRF-59:75-Ala-AGC-2-M4, and tRF-1:16-Gly-CCC-2 was performed; combined detection of tRF-1:16-Leu-AAG-1-M4, tRF-59:75-Ala-AGC-2-M4, tRF-1:28-chrM.Ser-TGA, and tiRNA-1:31-chrM.Met-CAT was performed; a total of 210 combinations were evaluated for diagnostic efficacy.

[0037] In some embodiments of the present invention, the combined diagnostic detection of five RNA molecules is analyzed and modeled using 14 different machine learning models to evaluate the diagnostic accuracy of different combinations. For example: combined detection of tRF-58:75-Ala-AGC-1, tRF-1:16-Leu-AAG-1-M4, tRF-59:75-Ala-AGC-2-M4, tRF-1:28-chrM.Ser-TGA, and tiRNA-1:31-chrM.Met-CAT; and combined detection of tRF-58:75-Ala-AGC-1, tRF-1:16-Leu-AAG-1-M4, and tRF-59:75-Ala-AGC-2-M4. M4, tRF-1:28-chrM.Ser-TGA, and tRF-1:16-Gly-CCC-2 were detected in combination; tRF-58:75-Ala-AGC-1, tRF-54:75-His-GTG-1-M2, tRF-1:16-Leu-AAG-1-M4, tRF-59:75-Ala-AGC-2-M4, and tRF-1:28-chrM.Ser-TGA were detected in combination; a total of 252 combinations were evaluated for diagnostic efficacy.

[0038] In some embodiments of the present invention, six RNA molecules are used for combined diagnostic detection, and 14 different machine learning models are used for analysis and modeling to evaluate the diagnostic accuracy of different combinations. For example, combined detection of tRF-58:75-Ala-AGC-1, tRF-1:16-Leu-AAG-1-M4, tRF-59:75-Ala-AGC-2-M4, tRF-1:28-chrM.Ser-TGA, tRF-1:16-Gly-CCC-2, and tRF-1:16-Leu-AAG-4-M2; and combined detection of tRF-58:75-Ala-AGC-1, tRF-1:16-Leu-AAG-1-M4, tRF-59:75-Ala-AGC-2-M4, and tRF-1:28-chrM.Ser-TGA, tRF-1:16-Gly-CCC-2, and tRF-1:16-Leu-AAG-4-M2; and combined detection of tRF-58:75-Ala-AGC-1, tRF-1:16-Leu-AAG-1-M4, tRF-59:75-Ala-AGC-2-M4, and tRF-1:28-chrM.Ser-TGA, tRF-1:16-Gly-CCC-2, and tRF-1:16-Leu-AAG-4-M2. M.Ser-TGA, tRF-1:16-Leu-AAG-4-M2, and tiRNA-1:31-chrM.Met-CAT were detected in combination; tRF-58:75-Ala-AGC-1, tRF-1:16-Leu-AAG-1-M4, tRF-59:75-Ala-AGC-2-M4, tRF-1:28-chrM.Ser-TGA, tRF-1:16-Gly-CCC-2, and tiRNA-1:31-chrM.Met-CAT were also detected in combination; a total of 210 combinations were evaluated for diagnostic efficacy.

[0039] In some embodiments of the present invention, seven RNA molecules are used for combined diagnostic detection, and 14 different machine learning models are used for analysis and modeling to evaluate the diagnostic accuracy of different combinations. For example: combined detection of tRF-58:75-Ala-AGC-1, tRF-54:75-His-GTG-1-M2, tRF-1:16-Leu-AAG-1-M4, tRF-59:75-Ala-AGC-2-M4, tRF-1:28-chrM.Ser-TGA, tRF-1:16-Gly-CCC-2, and tRF-1:16-Leu-AAG-4-M2; combined detection of tRF-58:75-Ala-AGC-1, tRF-54:75-His-GTG-1-M2, tRF-1:16-Leu-AAG-1-M4, and tRF-59:75-Ala-AGC-2-M4, t... RF-1:28-chrM.Ser-TGA, tRF-1:16-Gly-CCC-2, and tiRNA-1:31-chrM.Met-CAT2 were detected in combination; tRF-58:75-Ala-AGC-1, tRF-54:75-His-GTG-1-M2, tRF-1:16-Leu-AAG-1-M4, tRF-59:75-Ala-AGC-2-M4, tRF-1:28-chrM.Ser-TGA, tRF-1:16-Leu-AAG-4-M2, and tiRNA-1:31-chrM.Met-CAT were detected in combination; a total of 120 combinations were evaluated for diagnostic efficacy.

[0040] In some embodiments of the present invention, eight RNA molecules are combined for diagnostic detection, and 14 different machine learning models are used for analysis and modeling to evaluate the diagnostic accuracy of different combinations. For example: Combined detection of tRF-58:75-Ala-AGC-1, tRF-54:75-His-GTG-1-M2, tRF-1:16-Leu-AAG-1-M4, tRF-59:75-Ala-AGC-2-M4, tRF-1:28-chrM.Ser-TGA, tRF-1:16-Gly-CCC-2, tiRNA-1:34-Val-AAC-1-M3, and tiRNA-1:31-chrM.Met-CAT; tRF-58:75-Ala-AGC-1, tRF-54:75-His-GTG-1-M2, tRF-1:16-Leu-AAG-1-M4, tRF-59:75-Ala-AGC-2-M4, and tRF-+1:T3. 3-Gln-CTG-3, tRF-1:28-chrM.Ser-TGA, tRF-1:16-Leu-AAG-4-M2, and tiRNA-1:31-chrM.Met-CAT were detected in combination; tRF-58:75-Ala-AGC-1, tRF-1:16-Leu-AAG-1-M4, tRF-59:75-Ala-AGC-2-M4, tRF-1:28-chrM.Ser-TGA, tRF-1:16-Gly-CCC-2, tRF-1:16-Leu-AAG-4-M2, tiRNA-1:34-Val-AAC-1-M3, and tiRNA-1:31-chrM.Met-CAT were detected in combination; a total of 45 combinations were evaluated for diagnostic efficacy.

[0041] In some embodiments of the present invention, nine RNA molecules are used for combined diagnostic detection, and 14 different machine learning models are used for analysis and modeling to evaluate the diagnostic accuracy of different combinations. For example, the diagnostic accuracy of tRF-58:75-Ala-AGC-1, tRF-54:75-His-GTG-1-M2, tRF-1:16-Leu-AAG-1-M4, tRF-59:75-Ala-AGC-2-M4, tRF-+1:T33-Gln-CTG-3, tRF-1:28-chrM.Ser-TGA, tRF-1:16-Gly-CCC-2, and tRF-1... :16-Leu-AAG-4-M2 and tiRNA-1:31-chrM.Met-CAT were detected in combination; tRF-58:75-Ala-AGC-1, tRF-54:75-His-GTG-1-M2, tRF-1:16-Leu-AAG-1-M4, tRF-59:75-Ala-AGC-2-M4, tRF-1:28-chrM.Ser-TGA, tRF-1:16-Gly-CCC-2, tRF-1:16-Leu-AAG-4-M2, tiRNA-1:34-Val-AAC-1-M3, and tiRNA-1:31-chrM.Met-CAT were detected in combination; tRF-54:75-His-GTG-1-M2, tRF-1:16-Leu-AAG-1-M4, and tRF-59:75-Ala-AGC were also detected. The following 10 combinations were used for combined detection: tRF-2-M4, tRF-+1:T33-Gln-CTG-3, tRF-1:28-chrM.Ser-TGA, tRF-1:16-Gly-CCC-2, tRF-1:16-Leu-AAG-4-M2, tiRNA-1:34-Val-AAC-1-M3, and tiRNA-1:31-chrM.Met-CAT; their diagnostic efficacy was evaluated.

[0042] In some embodiments of the present invention, ten RNA molecules are used for combined diagnostic detection, and 14 different machine learning models are used for analysis and modeling to evaluate the diagnostic accuracy of different combinations. For example, the following RNA molecules are used for combined detection: tRF-58:75-Ala-AGC-1, tRF-54:75-His-GTG-1-M2, tRF-1:16-Leu-AAG-1-M4, tRF-59:75-Ala-AGC-2-M4, tRF-+1:T33-Gln-CTG-3, tRF-1:28-chrM.Ser-TGA, tRF-1:16-Gly-CCC-2, tRF-1:16-Leu-AAG-4-M2, tiRNA-1:34-Val-AAC-1-M3, and tiRNA-1:31-chrM.Met-CAT.

[0043] In some embodiments of the present invention, the machine learning model is a support vector machine (SVM) based on a sigmoid kernel function, a support vector machine (SVM) based on a radial basis function kernel function, a support vector machine (SVM) based on a polynomial kernel function, a support vector machine (SVM) based on a linear kernel function, a random forest, a quadratic discriminant analysis (QDA), a neural network, naive Bayes, logistic regression, linear discriminant analysis (LDA), a k-nearest neighbor algorithm (KNN), a decision tree, an adaptive boosting algorithm, or a bootstrap aggregating algorithm (bagging).

[0044] In some embodiments of the present invention, the machine learning model is a support vector machine (SVM with Radial Kernel), a neural network, or a logistic regression model; preferably.

[0045] In some embodiments of the present invention, the machine learning model is a support vector machine model based on radial basis kernel functions.

[0046] In some embodiments of the present invention, the machine learning model is a support vector machine model based on radial basis function kernel function, and the RNA molecule used to detect the expression level of RNA molecules is tRF-1:16-Leu-AAG-1-M4.

[0047] In some embodiments of the present invention, the machine learning model is a support vector machine model based on radial basis function kernel function, and the RNA molecule combination for detecting the expression level of RNA molecules is tRF-1:16-Leu-AAG-1-M4 and tRF-59:75-Ala-AGC-2-M4.

[0048] In some embodiments of the present invention, the machine learning model is a support vector machine model based on radial basis function kernel function, and the RNA molecule combination for detecting the expression level of RNA molecules is tRF-1:16-Leu-AAG-1-M4, tRF-59:75-Ala-AGC-2-M4 and tRF-+1:T33-Gln-CTG-3.

[0049] In some embodiments of the present invention, the machine learning model is a support vector machine model based on radial basis function kernel function, and the RNA molecule combination for detecting the expression level of RNA molecules is tRF-1:16-Leu-AAG-1-M4, tRF-59:75-Ala-AGC-2-M4, tRF-+1:T33-Gln-CTG-3, and tRF-1:28-chrM.Ser-TGA.

[0050] In some embodiments of the present invention, the machine learning model is a support vector machine model based on radial basis function kernel function, and the RNA molecule combination for detecting the expression level of RNA molecules is tRF-1:16-Leu-AAG-1-M4, tRF-59:75-Ala-AGC-2-M4, tRF-+1:T33-Gln-CTG-3, tRF-1:28-chrM.Ser-TGA, and tRF-1:16-Leu-AAG-4-M2.

[0051] In some embodiments of the present invention, the machine learning model is a support vector machine model based on radial basis function kernel function, and the RNA molecule combination for detecting the expression level of RNA molecules is tRF-58:75-Ala-AGC-1, tRF-1:16-Leu-AAG-1-M4, tRF-59:75-Ala-AGC-2-M4, tRF-+1:T33-Gln-CTG-3, tRF-1:28-chrM.Ser-TGA, and tRF-1:16-Leu-AAG-4-M2.

[0052] In some embodiments of the present invention, the machine learning model is a support vector machine model based on radial basis function kernel function, and the RNA molecule combination for detecting the expression level of RNA molecules is tRF-58:75-Ala-AGC-1, tRF-1:16-Leu-AAG-1-M4, tRF-59:75-Ala-AGC-2-M4, tRF-+1:T33-Gln-CTG-3, tRF-1:28-chrM.Ser-TGA, tRF-1:16-Leu-AAG-4-M2, and tiRNA-1:31-chrM.Met-CAT.

[0053] In some embodiments of the present invention, the machine learning model is a support vector machine model based on radial basis function kernel function, and the RNA molecule combination for detecting RNA molecule expression levels is tRF-58:75-Ala-AGC-1, tRF-1:16-Leu-AAG-1-M4, tRF-59:75-Ala-AGC-2-M4, tRF-+1:T33-Gln-CTG-3, tRF-1:28-chrM.Ser-TGA, tRF-1:16-Leu-AAG-4-M2, tiRNA-1:34-Val-AAC-1-M3, and tiRNA-1:31-chrM.Met-CAT.

[0054] In some embodiments of the present invention, the machine learning model is a support vector machine model based on radial basis function kernel function, and the RNA molecule combination for detecting the expression level of RNA molecules is tRF-58:75-Ala-AGC-1, tRF-54:75-His-GTG-1-M2, tRF-1:16-Leu-AAG-1-M4, tRF-59:75-Ala-AGC-2-M4, tRF-+1:T33-Gln-CTG-3, tRF-1:28-chrM.Ser-TGA, tRF-1:16-Leu-AAG-4-M2, tiRNA-1:34-Val-AAC-1-M3, and tiRNA-1:31-chrM.Met-CAT.

[0055] In some embodiments of the present invention, the machine learning model is a support vector machine model based on radial basis function kernel function, and the RNA molecule combination for detecting the expression level of RNA molecules is tRF-58:75-Ala-AGC-1, tRF-54:75-His-GTG-1-M2, tRF-1:16-Leu-AAG-1-M4, tRF-59:75-Ala-AGC-2-M4, tRF-+1:T33-Gln-CTG-3, tRF-1:28-chrM.Ser-TGA, tRF-1:16-Gly-CCC-2, tRF-1:16-Leu-AAG-4-M2, tiRNA-1:34-Val-AAC-1-M3+, and tiRNA-1:31-chrM.Met-CAT.

[0056] This invention utilizes high-throughput sequencing technology to detect RNA molecules in the endometrial tissues of patients with endometriosis and normal controls, and screens for significantly differentially expressed RNA molecules. Expanding the sample size, qRT-PCR data was collected from serum samples of 160 patients with endometriosis and 163 healthy individuals, and machine learning models were constructed and validated. The models demonstrated high accuracy in single and combined diagnosis using 1 to 10 RNA molecules, with high AUC and accuracy scores, indicating high accuracy across 14 machine learning models. In particular, the combined use of multiple RNA molecules significantly improved the model's detection accuracy. The inventors selected the highest AUC and Accuracy scores from 14 different models and RNA molecule combinations for validation. The results showed that the AUC and Accuracy scores of Support Vector Machine (SVM) with Radial Kernal, Neural Network, and Logistic Regression based on Radial Kernel Function were high and stable, indicating good reliability and accuracy of the models. In particular, the Support Vector Machine (SVM) with Radial Kernel Function model was selected. This invention ultimately identified the RNA molecule composition with the highest AUC score among the Support Vector Machine models based on Radial Kernel Function. Attached Figure Description

[0057] Figure 1A shows the receiver operating characteristic (ROC) curve of the detection results of a serum RNA molecule for identifying endometriosis patients.

[0058] Figure 1B shows the receiver operating characteristic (ROC) curves of the detection results for identifying endometriosis patients using two types of serum RNA molecules.

[0059] Figure 1C shows the receiver operating characteristic (ROC) curves of the detection results for identifying endometriosis patients using three types of serum RNA molecules.

[0060] Figure 1D shows the receiver operating characteristic (ROC) curves of the detection results of four serum RNA molecules in identifying endometriosis patients.

[0061] Figure 1E shows the receiver operating characteristic (ROC) curves of the detection results for identifying endometriosis patients using five serum RNA molecules.

[0062] Figure 1F shows the receiver operating characteristic (ROC) curves of the detection results for identifying endometriosis patients using six serum RNA molecules.

[0063] Figure 1G shows the receiver operating characteristic (ROC) curves of the detection results for identifying endometriosis patients using seven serum RNA molecules.

[0064] Figure 1H shows the receiver operating characteristic (ROC) curves of the detection results for identifying endometriosis patients using eight serum RNA molecules.

[0065] Figure 1I shows the receiver operating characteristic (ROC) curves of the detection results of nine serum RNA molecules in identifying endometriosis patients.

[0066] Figure 1J shows the receiver operating characteristic (ROC) curves of the detection results of ten serum RNA molecules in identifying endometriosis patients.

[0067] Figure 2A shows the AUC scores of a Sigmoid kernel-based support vector machine model for identifying endometriosis patients using combinations of 1 to 10 RNA molecules.

[0068] Figure 2B shows the accuracy scores of a Sigmoid kernel function-based support vector machine model for identifying endometriosis patients using combinations of 1 to 10 RNA molecules.

[0069] Figure 2C shows the AUC scores of a support vector machine model based on radial basis kernel functions for identifying endometriosis patients using combinations of 1 to 10 RNA molecules.

[0070] Figure 2D shows the accuracy scores of a radial basis function-based support vector machine model for identifying endometriosis patients using combinations of 1 to 10 RNA molecules.

[0071] Figure 2E shows the AUC scores of a multinomial kernel-based support vector machine model for identifying endometriosis patients using combinations of 1 to 10 RNA molecules.

[0072] Figure 2F shows the accuracy scores of a multinomial kernel function-based support vector machine model for identifying endometriosis patients using combinations of 1 to 10 RNA molecules.

[0073] Figure 2G shows the AUC scores of a linear kernel function-based support vector machine model for identifying endometriosis patients using combinations of 1 to 10 RNA molecules.

[0074] Figure 2H shows the accuracy scores of a linear kernel function-based support vector machine model for identifying endometriosis patients using combinations of 1 to 10 RNA molecules.

[0075] Figure 2I shows the AUC scores of random forest models for identifying endometriosis patients using combinations of 1 to 10 RNA molecules.

[0076] Figure 2J shows the accuracy scores of random forest models for identifying endometriosis patients using combinations of 1 to 10 RNA molecules.

[0077] Figure 3A shows the AUC scores of a secondary discriminant analysis model for identifying endometriosis patients using combinations of 1 to 10 RNA molecules.

[0078] Figure 3B shows the accuracy scores of a secondary discriminant analysis model for identifying endometriosis patients using combinations of 1 to 10 RNA molecules.

[0079] Figure 3C shows the AUC scores of neural network models for identifying endometriosis patients using combinations of 1 to 10 RNA molecules.

[0080] Figure 3D shows the accuracy scores of neural network models for identifying endometriosis patients using combinations of 1 to 10 RNA molecules.

[0081] Figure 3E shows the AUC scores of a Naive Bayes model for identifying endometriosis patients using combinations of 1 to 10 RNA molecules.

[0082] Figure 3F shows the accuracy scores of a Naive Bayes model for identifying endometriosis patients using combinations of 1 to 10 RNA molecules.

[0083] Figure 3G shows the AUC scores of logistic regression models for identifying endometriosis patients using combinations of 1 to 10 RNA molecules.

[0084] Figure 3H shows the accuracy scores of logistic regression models for identifying endometriosis patients using combinations of 1 to 10 RNA molecules.

[0085] Figure 3I shows the AUC scores of linear discriminant analysis models for identifying endometriosis patients using combinations of 1 to 10 RNA molecules.

[0086] Figure 3J shows the accuracy scores of linear discriminant analysis models for identifying endometriosis patients using combinations of 1 to 10 RNA molecules.

[0087] Figure 4A shows the AUC scores of the K-nearest neighbor algorithm model for identifying endometriosis patients using combinations of 1 to 10 RNA molecules.

[0088] Figure 4B shows the accuracy scores of the K-nearest neighbor algorithm model for identifying endometriosis patients using combinations of 1 to 10 RNA molecules.

[0089] Figure 4C shows the AUC scores of decision tree models for identifying endometriosis patients using combinations of 1 to 10 RNA molecules.

[0090] Figure 4D shows the accuracy scores of decision tree models for identifying endometriosis patients using combinations of 1 to 10 RNA molecules.

[0091] Figure 4E shows the AUC scores of an adaptive enhancement model for identifying endometriosis patients using combinations of 1 to 10 RNA molecules.

[0092] Figure 4F shows the accuracy scores of an adaptive enhancement model for identifying endometriosis patients using combinations of 1 to 10 RNA molecules.

[0093] Figure 4G shows the AUC scores of guided aggregation algorithm models for identifying endometriosis patients using combinations of 1 to 10 RNA molecules.

[0094] Figure 4H shows the accuracy scores of guided aggregation algorithm models for identifying endometriosis patients using combinations of 1 to 10 RNA molecules.

[0095] Figure 5A shows the highest AUC score curves for identifying endometriosis patients using combinations of 1 to 10 RNA molecules across 14 machine learning models.

[0096] Figure 5B shows the highest accuracy score curves among 14 machine models for identifying endometriosis patients using combinations of 1 to 10 RNA molecules. Detailed Implementation

[0097] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present application.

[0098] It should be noted that the terms "comprising" and "having" and any variations thereof in the specification, claims and accompanying drawings of this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or units that are not explicitly listed or that are inherent to such process, method, product or device.

[0099] In this invention, "polynucleotide," "oligonucleotide," "nucleic acid," or "nucleic acid sequence" includes, but is not limited to, mRNA, cDNA, cccDNA, genomic DNA, and synthetic DNA and RNA sequences containing the natural nucleoside bases adenine, guanine, cytosine, thymine, and uracil, and also includes sequences having one or more modified or unmodified nucleosides. The term "nucleic acid" or "nucleic acid sequence" includes oligonucleotides and polynucleotides. None of these terms used herein are limited in length or synthetic origin.

[0100] In this invention, a "sample" can be a tissue or body fluid sample or an isolated sample. As used herein, the term "tissue" refers to any biological material consisting of a single cell, multiple cells, a cluster of cells, or an entire organ. The term tissue as used herein includes one or more cells that may be normal or abnormal.

[0101] In this invention, "PCR reaction" is a molecular biology technique used to amplify specific DNA fragments. It can be regarded as a special DNA replication outside of living organisms. The most significant feature of PCR is that it can greatly increase a trace amount of DNA.

[0102] The invention will be further analyzed below with reference to specific examples.

[0103] Example 1

[0104] The collected clinical samples were mainly from female patients who underwent surgical treatment (see Table 1). In situ and ectopic tissue samples were collected from patients with endometriosis, and endometrial tissue samples were collected from five age-matched non-endometriosis patients (control group). High-throughput sequencing was used to analyze five pairs of ectopic endometrial tissue and situ endometrial ESC cells, and ten significantly differentially expressed RNA molecules were screened based on their expression levels.

[0105] (See Table 2).

[0106] Table 1: Basic Information of Clinical Samples

[0107] Table 2: Basic Information about RNA Molecules

[0108] Example 2

[0109] Quantitative PCR was used to detect differentially expressed RNA molecules:

[0110] Step 1: Human serum samples were collected from 163 normal women and 164 patients with endometriosis. Clinical characteristics of endometriosis patients and the normal control group are shown in Table 3. Step 2: Total RNA was extracted from human serum using Trizol reagent. Step 3: cDNA was obtained through reverse transcription of RNA using the riboSCRIPT™ Reverse Transcription Kit. Step 4: Primers designed with specific EM-free RNA were added to the PCR system, and quantitative PCR was performed using a real-time PCR instrument. The PCR reagent was the bulger-loop miRNA qRT-PCR Stater Kit (Ribobio, China) designed for RNA molecular sequences. The reverse transcription reaction system in Step 3 included 2 μL of 15x RT buffer, 2 μL of 10 mM mixture of each dNTP, 0.5 μL of RNase Inhibitor, and 1.5 μL of gene-specific reverse primer mixture. The reaction steps were: incubation at 42°C for 60 minutes, incubation at 70°C for 10 minutes, and storage at 4°C. Step 4 involves quantitative PCR using a real-time PCR instrument. The reaction mixture consists of 5 μl of SYBR Green, 2 pmol each of forward and reverse primers, 1 μl of cDNA, and DEPC water to a final volume of 10 μl. The reaction conditions are: 95℃ for 5 minutes for one cycle, followed by 45 cycles of 95℃ for 10 seconds and 60℃ for 30 seconds.

[0111] Table 3: Clinical characteristics of patients with endometriosis and normal controls

[0112] BMI: Body Mass Index; AFS: American Fertility Association Laparoscopic Diagnosis of Endometriosis Scoring System.

[0113] Example 3

[0114] To assess the correlation between the expression levels of 10 differentially expressed RNA molecules in serum and endometriosis, the datasets used were the quantitative PCR results from 163 normal women and 164 endometriosis patients. A generalized linear model, logistic regression, was used to model and plot the receiver operating characteristic (ROC) curve. The ROC curve is a tool used to evaluate the performance of binary classification models. It demonstrates the classification ability (i.e., diagnostic ability) of RNA molecules by plotting the relationship between the true positive rate (TPR) and the false positive rate (FPR) at different thresholds. Generally, the closer the curve is to the upper left corner and the larger the area under the curve, the better the model's classification performance, i.e., the better its diagnostic ability.

[0115] The results showed that using a generalized linear logistic regression model, different combinations of 10 RNA molecules were used for diagnostic testing and evaluation. Receiver operating characteristic (ROC) curves (Figures 1A-1J) were plotted. It can be seen that classification performance was achieved using 1 to 10 RNA molecules as diagnostic indicators, and the classification ability improved when multiple RNA molecules were used as combined diagnostic indicators. Figure 1 illustrates that the expression levels of the 10 differentially expressed RNA molecules in serum were all associated with endometriosis.

[0116] Example 4

[0117] The dataset was obtained from the quantitative PCR test results of 163 normal women and 164 patients with endometriosis. 20% of the test result dataset was divided into a test set and the remaining 80% was used as a training set. Fourteen different machine learning models were compiled. For learning algorithms that require multiple iterations, a learning rate adjustment callback was set to optimize the convergence speed and final performance of the model. Early stopping was set to prevent overfitting. The 14 machine learning models are: Support Vector Machine with Sigmoid Kernel, Support Vector Machine with Radial Kernel, Support Vector Machine with Polynomial Kernel, Support Vector Machine with Linear Kernel, Random Forest, Quadratic Discriminant Analysis (QDA), Neural Network, Naive Bayes, Logistic Regression, Linear Discriminant Analysis (LDA), K-Nearest Neighbors (KNN), Decision Tree, Boost, and Bootstrap aggregating (bagging). The final output includes model evaluation results, specifically the area under the curve (AUC) and accuracy score (the proportion of correctly predicted cases) of different RNA molecules used to identify endometriosis patients, evaluated across 14 machine learning models for both single and combined detection. Higher AUC and accuracy scores indicate higher model detection accuracy.

[0118] The results show:

[0119] Figures 2, 3, and 4 illustrate the AUC and Accuracy scores of 1 to 10 RNA molecules used alone and in combination for the diagnosis of endometriosis across 14 machine learning models. As can be seen from Figures 2, 3, and 4, the AUC and Accuracy scores are high for both single and combined diagnoses using 1 to 10 RNA molecules, indicating high accuracy across the 14 machine learning models. In particular, the combined use of multiple RNA molecules significantly improves the model's detection accuracy.

[0120] 2. The highest AUC and Accuracy scores of different models and different RNA molecule combinations were plotted (Figure 5). As can be seen from Figure 5, the AUC and Accuracy scores of Support Vector Machine (SVM with Radial Kernal), Neural Network, and Logistic Regression based on Radial Kernel Function are high and stable, indicating that the models have good reliability and accuracy.

[0121] The inventors selected a support vector machine (SVM) model based on radial basis kernel functions to diagnose all combinations of RNA molecules. The results are shown in Table 4. Finally, the group of RNA molecules with the highest AUC score in the support vector machine model based on radial basis kernel functions was selected, as shown in Table 5.

[0122] Table 4: AUC and Accuracy scores of all RNA molecules in the support vector model based on radial basis kernel functions.

[0123] Table 5: Composition of RNA molecules with the highest AUC scores in the support vector machine model based on radial basis kernel function.

Claims

1. A reagent kit, characterized in that, The kit contains a detection reagent for detecting the expression levels of one or more RNA molecules.

2. The reagent kit according to claim 1, characterized in that, The nucleotide sequence of the RNA molecule is shown in any one of SEQ ID NO.1 to SEQ ID NO.

10.

3. The kit according to claim 1 or 2, characterized in that, The kit is used to diagnose endometriosis, and the test reagent is used to detect the expression levels of one or more RNA molecules in serum.

4. The reagent kit according to claim 3, characterized in that, The detection reagent is used to detect the expression levels of 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 RNA molecules in serum; preferably, the detection reagent is used to detect the expression levels of 3, 4, 5, or 6 RNA molecules in serum; more preferably, the detection reagent is used to detect the expression levels of 4 or 5 RNA molecules in serum.

5. The reagent kit according to claim 4, characterized in that, The detection reagents include reverse transcription reagents and PCR reaction reagents.

6. The reagent kit according to claim 5, characterized in that, The reverse transcription reagents include 5x RT buffer, RT Mix, RNase Inhibitor, and specific reverse transcription primers.

7. The use of the kit according to any one of claims 1 to 6 in the diagnosis of endometriosis.

8. A method for diagnosing endometriosis, characterized in that, Includes the following steps: (1) Collect human serum samples; (2) Detect the expression levels of one or more RNA molecules as described in claim 1 in serum samples; (3) Input the expression level of RNA molecules in step (2) into the trained machine learning model, output the model evaluation results, and determine whether the patient has endometriosis.

9. The method for diagnosing endometriosis according to claim 8, characterized in that, The machine learning models mentioned are: Support Vector Machine (SVM) with Sigmoid Kernal, Support Vector Machine with Radial Kernal, Support Vector Machine with Polynomial Kernal, Support Vector Machine with Linear Kernal, Random Forest, Quadratic Discriminant Analysis (QDA), Neural Network, Naive Bayes, Logistic Regression, Linear Discriminant Analysis (LDA), K-Nearest Neighbors (KNN), Decision Tree, Boost, or Bootstrap aggregating (bagging).

10. The method for diagnosing endometriosis according to claim 9, characterized in that, The machine learning model is a support vector machine (SVM) with radial basis function (RBF) kernel, a neural network, or a logistic regression; preferably, the machine learning model is a support vector machine model with RBF kernel.

Images