Method and kit for constructing artificial intelligence model for pregnancy identification based on TPIS, ENOA and PPIB tear markers
Through the construction and kit of pregnancy identification artificial intelligence model based on TPIS, ENOA, and PPIB tear markers, the problem of single feature limitation in existing pregnancy identification technology was solved, and high accuracy and sensitivity were achieved using random forest method and random forest method and ELISA detection. pregnancy identification.
Patent Information
- Application Number
- CN202311418691.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-10-30
- Publication Date
- 2024-11-26
- Estimated Expiration
- 2043-10-30
AI Technical Summary
The existing pregnancy identification technology mainly relies on the level of human chorionic gonadotropin (hCG) in a single feature, limiting the development of the AI model for pregnancy identification, and it is urgent to develop new pregnancy identification characteristics to improve the accuracy and sensitivity of the technology.
The construction method of pregnancy identification artificial intelligence model based on TPIS, ENOA, and PPIB tear markers. By collecting and detecting tear samples from normal women and early pregnancy, using the random forest method to train the artificial intelligence model, and developing pregnancy identification reagents based on tear protein Cassette, using ELISA reagent to detect tear concentration data for identification.
The artificial intelligence model trained through tear concentration data of TPIS, ENOA, and PPIB achieves high accuracy and sensitivity pregnancy identification, providing a new pregnancy identification tool, and improving the performance of pregnancy identification technology.
Smart Images

Figure CN117524320B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of pregnancy identification, and in particular to a method for constructing an artificial intelligence model for pregnancy identification based on TPIS, ENOA and PPIB tear markers, and also to a kit for pregnancy identification. Background Art
[0002] The identification programs generated by artificial intelligence training have gradually replaced manually generated identification formulas or competent experience to become the mainstream tools for identification. The elements of artificial intelligence training include features, feature data, and machine learning or deep learning algorithms suitable for feature data types. At present, there is only one main feature for pregnancy identification, namely the level of human chorionic gonadotropin (hCG) in blood or urine. However, a single feature limits the development of artificial intelligence models for pregnancy identification. Therefore, it is urgent to develop new markers with pregnancy identification characteristics and provide a new pregnancy identification tool, which is conducive to the development of pregnancy identification technology. Summary of the invention
[0003] In view of this, one of the objects of the present invention is to provide a method for constructing an artificial intelligence model for pregnancy identification based on TPIS, ENOA, and PPIB tear markers; a second object of the present invention is to provide a pregnancy identification kit based on tear protein.
[0004] In order to achieve the above object, the present invention provides the following technical solutions:
[0005] 1. A method for constructing an artificial intelligence model for pregnancy identification based on TPIS, ENOA, and PPIB tear markers, including the following steps:
[0006] (1) Tear samples were collected from normal women and early pregnant women, and the tear concentrations of TPIS, ENOA, and PPIB were tested to obtain the data sets of TPIS, ENOA, and PPIB;
[0007] (2) The data set obtained in step (1) was input into the sklearn.ensemble.RandomForestClassifier program, and the pregnancy identification artificial intelligence model was trained by the random forest method. The training parameters were as follows: the number of neurons in the first hidden layer was 200, the number of neurons in the second hidden layer was 100, the activation function of the first hidden layer was relu, the activation function of the first hidden layer was tanh, epochs = 20, and batch size = 32; after 50 random test set tests, the AUC value of each identification exceeded 0.99 and passed the artificial intelligence model training.
[0008] Preferably, the Uniprot database receiving number of the TPIS is P60174, the Uniprot database receiving number of the ENOA is P06733, and the Uniprot database receiving number of the PPIB is P23284.
[0009] Preferably, in the present invention, there are significant differences in the tear concentrations of TPIS, ENOA and PPIB between non-pregnant women and women in early pregnancy, and the tear concentrations can be used as indicators for identifying pregnancy.
[0010] Preferably, the artificial intelligence model of the present invention refers to an artificial intelligence program capable of identifying whether a subject is pregnant based on the tear concentration data of the subject's TPIS, ENOA, and PPIB.
[0011] Preferably, the training of the present invention is based on machine learning and deep learning training.
[0012] Preferably, in the present invention, the principle of detecting the tear concentration of TPIS, ENOA and PPIB of the subject is affinity labeling, spectroscopy, chromatography or mass spectrometry.
[0013] Preferably, the principle of detecting the tear concentrations of TPIS, ENOA and PPIB of the subject is enzyme-linked immunosorbent assay.
[0014] 2. A pregnancy identification kit based on tear proteins, including detection reagents for tear proteins TPIS, ENOA and PPIB.
[0015] Preferably, in the present invention, the detection reagent for the tear proteins TPIS, ENOA and PPIB is an ELISA reagent.
[0016] The beneficial effects of the present invention are as follows: the present invention discloses a method for constructing an artificial intelligence model for pregnancy identification based on TPIS, ENOA, and PPIB tear markers, and through research, it is found that the tear concentrations of TPIS, ENOA, and PPIB have the properties of being pregnancy identification features, and the training of the artificial intelligence model for pregnancy identification is completed through the tear concentration data of TPIS, ENOA, and PPIB. The present invention provides a new method for preparing a pregnancy identification tool, which is conducive to improving the accuracy and sensitivity of pregnancy identification technology. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] In order to make the purpose, technical solution and beneficial effects of the present invention clearer, the present invention provides the following drawings for illustration:
[0018] Figure 1 This is the tear concentration atlas of non-pregnant and early pregnant women with TPIS.
[0019] Figure 2 AUC values for identifying pregnancy for TPIS.
[0020] Figure 3 This is the tear concentration spectrum of ENOA in non-pregnant and early pregnant women.
[0021] Figure 4 The AUC values for ENOA identification of pregnancy.
[0022] Figure 5 This is the tear concentration spectrum of PPIB in non-pregnant and early pregnant women.
[0023] Figure 6 The AUC value for PPIB in identifying pregnancy.
[0024] Figure 7 Schematic diagram of the process of constructing an artificial intelligence model for pregnancy identification based on TPIS, ENOA, and PPIB tear markers DETAILED DESCRIPTION
[0025] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments so that those skilled in the art can better understand the present invention and implement it, but the embodiments are not intended to limit the present invention.
[0026] Example 1: TPIS, ENOA, PPIB detection of tear protein concentration in non-pregnant and early pregnant women
[0027] 182 normal females (Group C) and 176 volunteers in early pregnancy (Group ES) were recruited. The inclusion criteria were: no history of pregnancy, no history of chronic diseases, no history of medication within 6 months, no medical abnormalities during pregnancy, no history of ophthalmic diseases, and no abnormalities of key organs as shown by physical examination. Tear samples were obtained and the tear concentrations of TPIS, ENOA, and PPIB were detected by ELISA. The tear concentration spectrum of TPIS is shown in Figure 2. Figure 1 As shown in Figure 2, the AUC value for identifying pregnancy is 0.9854. Figure 2 The tear concentration spectrum of ENOA is shown in Figure 3 As shown in Figure 2, the AUC value for identifying pregnancy is 0.9586. Figure 4 The tear concentration spectrum of PPIB is shown in Figure 5 As shown in Figure 2, the AUC value for identifying pregnancy is 0.8657. Figure 6 As shown. The ELISA detection method for tear total protein extraction and specified protein tear concentration in this embodiment can be found in Benitez-Del-Castillo JM, Soria J, Acera A, AM, Rodríguez S, Suárez T. Quantification of a panel for dry-eye protein biomarkers in tears: A comparative pilot study using standard ELISA and customized microarrays. Mol Vis. 2021 May 8; 27: 243-261. PMID: 34012227; PMCID: PMC8116255. The data are shown in Table 1.
[0028] Table 1. Dataset for constructing artificial intelligence model for pregnancy identification based on TPIS, ENOA, and PPIB tear markers
[0029]
[0030]
[0031]
[0032]
[0033]
[0034]
[0035]
[0036]
[0037] Example 2: Construction of an artificial intelligence model for pregnancy identification based on TPIS, ENOA, and PPIB tear markers
[0038] The data set measured in Example 1 (Table 1) was input into the sklearn.ensemble.RandomForestClassifier program, and the pregnancy identification artificial intelligence model was trained by the random forest method. The schematic diagram and hyperparameters are shown in Figure 7The results of 50 random test set tests are shown in Table 2, and the AUC value of each identification exceeds 0.99. For the method of using the sklearn.ensemble.RandomForestClassifier training program in this embodiment, see Kim J, Lee M, Lee MK, Wang SM, Kim NY, Kang DW, Um YH, Na HR, Woo YS, Lee CU, Bahk WM, Kim D, Lim HK. Development of Random Forest Algorithm Based Prediction Model of Alzheimer's Disease Using Neurodegeneration Pattern. Psychiatry Investig. 2021 Jan; 18(1): 69-79. doi: 10.30773 / pi.2020.0304. Epub 2021 Jan 25. PMID: 33561931; PMCID: PMC7897872. For the definitions of the terms AUC, sensitivity, specificity, etc. in this embodiment, see AM.Measures of Diagnostic Accuracy:Basic Definitions.EJIFCC.2009Jan 20;19(4):203-11.PMID:27683318;PMCID:PMC4975285 and Hess AS,Shardell M,Johnson JK,Thom KA,Strassle P,Netzer G,Harris AD.Methods and recommendations for evaluating and reporting a new diagnostic test.Eur J Clin Microbiol Infect Dis.2012Sep;31(9):2111-6.doi:10.1007 / s10096-012-1602-1.Epub 2012Mar 29.PMID:22476385;PMCID:PMC3661219.
[0039] Table 2. Test results of 50 randomly assigned test sets during training
[0040]
[0041]
[0042] The above-described embodiments are only preferred embodiments for fully illustrating the present invention, and the protection scope of the present invention is not limited thereto. Equivalent substitutions or changes made by those skilled in the art based on the present invention are within the protection scope of the present invention. The protection scope of the present invention shall be subject to the claims.
Claims
1. A method for constructing an artificial intelligence model for pregnancy identification based on TPIS, ENOA, and PPIB tear markers, characterized in that: The steps include: (1) Tear samples were collected from normal women and early pregnant women, and the tear concentrations of TPIS, ENOA, and PPIB were tested to obtain the data sets of TPIS, ENOA, and PPIB; (2) The data set obtained in step (1) was input into the sklearn.ensemble.RandomForestClassifier program, and the pregnancy identification artificial intelligence model was trained by the random forest method. The training parameters were as follows: the number of neurons in the first hidden layer was 200, the number of neurons in the second hidden layer was 100, the activation function of the first hidden layer was relu, the activation function of the first hidden layer was tanh, epochs = 20, and batch size = 32; after 50 random test set tests, the AUC value of each identification exceeded 0.99 and passed the artificial intelligence model training.
2. The method for constructing an artificial intelligence model for pregnancy identification according to claim 1, characterized in that: The Uniprot database receiving number of the TPIS is P60174, the Uniprot database receiving number of ENOA is P06733, and the Uniprot database receiving number of PPIB is P23284.
3. The method for constructing an artificial intelligence model for pregnancy identification according to claim 1, characterized in that: There are significant differences in the tear concentrations of TPIS, ENOA, and PPIB between non-pregnant women and women in early pregnancy, and the tear concentrations can be used as indicators for identifying pregnancy.
4. The method for constructing an artificial intelligence model for pregnancy identification according to claim 1, characterized in that: The artificial intelligence model refers to an artificial intelligence program that can identify whether the subject is pregnant based on the subject's TPIS, ENOA, and PPIB tear concentration data.
5. The method for constructing an artificial intelligence model for pregnancy identification according to claim 1, characterized in that: The training is based on machine learning and deep learning training.
6. The method for constructing an artificial intelligence model for pregnancy identification according to claim 1, characterized in that: The principle of detecting the tear concentration of TPIS, ENOA and PPIB of the subject is affinity labeling, spectrum, chromatography or mass spectrometry.
7. The method for constructing an artificial intelligence model for pregnancy identification according to claim 1, characterized in that: The principle of detecting the tear concentrations of TPIS, ENOA and PPIB of the subject is enzyme-linked immunosorbent assay.
8. The method for constructing an artificial intelligence model for pregnancy identification according to claim 1, characterized in that: The detection reagents of the tear proteins TPIS, ENOA and PPIB are ELISA reagents.
