Application of AKR1C1 protein in preparation of osteoporosis diagnostic kit

Abstract

The invention belongs to the technical field of biomedicine, and discloses application of AKR1C1 protein in preparation of an osteoporosis diagnostic kit. The protein AKR1C1 related to occurrence of osteoporosis of a hypertensive patient is screened out through a machine learning algorithm, the protein is significantly down-regulated in the body of the hypertensive patient suffering from osteoporosis and is a key protein affecting occurrence of osteoporosis, and the expression quantity of the protein AKR1C1 is in negative correlation with the occurrence risk of osteoporosis. The AKR1C1 protein has remarkable sensitivity and specificity in the diagnosis of the osteoporosis, the AUC of the AKR1C1 protein is 0.794, the diagnosis accuracy of the AKR1C1 protein for detecting the osteoporosis reaches 76.25%, and the AKR1C1 protein has relatively high discrimination capability on the osteoporosis. The invention provides a marker for rapid and accurate diagnosis of osteoporosis of hypertension patients, and has important significance for clinical treatment of patients suffering from hypertension and osteoporosis at the same time.

Description

Technical Field

[0001] The present invention belongs to the field of biomedical technologies and relates to the use of AKR1C1 protein in the preparation of an osteoporosis diagnostic kit. Background Art

[0002] Osteoporosis (OP) is a systemic bone disease characterized by low bone mass, damage to the microstructure of bone tissue resulting in increased bone fragility, and a high susceptibility to fractures. According to its etiology, it can be divided into two major categories: primary and secondary. Primary OP includes postmenopausal OP, senile OP, and juvenile idiopathic OP; secondary OP refers to osteoporosis caused by diseases or drugs that affect bone metabolism or other clear etiologies, such as glucocorticoid-induced OP. With the aging of the population, the prevalence of osteoporosis is gradually increasing. The symptoms of osteoporosis are usually not obvious, and there may be no obvious symptoms in the early stage. However, as the disease progresses, symptoms such as pain, height loss, skeletal deformity, easy fractures, and dyspnea may occur.

[0003] Both the incidence of hypertension and osteoporosis are related to age, lifestyle (such as high-sodium diet, lack of exercise), genetic factors, etc. Moreover, patients with hypertension often suffer from metabolic syndrome, which is related to the occurrence of osteoporosis. In clinical trial protocols, for patients suffering from both hypertension and osteoporosis, the treatment needs of the two diseases need to be comprehensively evaluated, and a drug combination beneficial to both diseases should be selected. Evaluating whether a hypertensive patient has osteoporosis usually requires measuring the patient's bone density. However, this method has disadvantages such as high cost and inconvenience. Therefore, exploring a molecular marker that can quickly and accurately identify whether a hypertensive patient has osteoporosis is of great significance for the clinical treatment of patients suffering from both hypertension and osteoporosis. Summary of the Invention

[0004] To solve the above technical problems, the present invention provides the use of AKR1C1 protein in the preparation of an osteoporosis diagnostic kit. The AKR1C1 protein screened by the present invention is significantly down-regulated in the body of hypertensive patients with osteoporosis and is a key protein affecting the occurrence of osteoporosis. The expression level of AKR1C1 protein is negatively correlated with the risk of osteoporosis. For every 1-unit increase in the standardized AKR1C1 expression level, the corresponding OR value is 0.080. After experimental verification, AKR1C1 protein has significant sensitivity and specificity in the diagnosis of osteoporosis. The diagnostic accuracy rate of AKR1C1 protein in detecting osteoporosis reaches 76.25%, AUC = 0.7941, and it has a relatively high discriminative ability for osteoporosis.

[0005] To achieve the technical objectives of the present invention, on the one hand, the present invention provides the use of AKR1C1 protein in the preparation of an osteoporosis diagnostic kit or diagnostic device, and the diagnostic kit is for hypertensive patients.

[0006] Furthermore, the diagnostic kit or diagnostic device includes a reagent for detecting the expression level of AKR1C1 protein in a sample, and the sample includes a plasma sample.

[0007] Furthermore, the expression level of the AKR1C1 protein is negatively correlated with the risk of osteoporosis occurrence.

[0008] On the other hand, the present invention claims protection for an osteoporosis auxiliary diagnosis system, including: an osteoporosis diagnosis model;

[0009] An osteoporosis diagnosis model is constructed with the expression level of AKR1C1 protein in a sample as the input variable: risk value = expression level of AKR1C1 protein;

[0010] The higher the risk value, the lower the probability of having osteoporosis.

[0011] Furthermore, the risk value is compared with a threshold value. If it is lower than the threshold value, it is determined to have osteoporosis. The threshold value is obtained from previous tests, that is, the difference degree of biomarkers between hypertensive patients with osteoporosis and hypertensive patients without osteoporosis is determined through tests and data analysis. When there is a difference or significant difference (the difference is statistically significant, such as p < 0.05, p < 0.01, p < 0.001, p < 0.0001) between the risk value and the threshold value, it is determined to have osteoporosis.

[0012] Specifically, in order to further evaluate the dose-response relationship between AKR1C1 protein and osteoporosis, a restricted cubic spline curve shows that when the expression level of AKR1C1 protein is less than 0.68, the risk of osteoporosis occurrence increases significantly. Therefore, the threshold value is set to 0.68.

[0013] Specifically, the present invention discovers through Logistic regression analysis that the expression level of AKR1C1 protein is negatively correlated with the risk of osteoporosis occurrence, that is, as the expression level of AKR1C1 protein increases, the risk of osteoporosis occurrence gradually decreases. Specifically, for every 1 unit increase in the standardized expression level of AKR1C1 protein, the corresponding OR value is 0.080, indicating that for every 1 unit increase in the expression level of AKR1C1 protein, the risk of osteoporosis occurrence is reduced to 8% of the original value, and AKR1C1 protein has a significant protective effect on osteoporosis.

[0014] Specifically, based on the analysis of the ROC curve, the present invention found that the area under the curve (AUC) of the AKR1C1 protein was 0.794, which was higher than the good standard of 0.70 in clinical diagnosis. At the same time, its diagnostic accuracy reached 76.25%, indicating that the AKR1C1 protein had a high discriminative ability in the diagnosis of osteoporosis.

[0015] Compared with the prior art, the technical solution provided by the present invention has at least the following beneficial effects or advantages:

[0016] (1) The present invention uses a machine learning method to screen out the key protein AKR1C1 related to the occurrence of osteoporosis in hypertensive patients. This protein is significantly down-regulated in the body of hypertensive patients with osteoporosis and is a key protein affecting the occurrence of osteoporosis, having extremely high diagnostic value for osteoporosis.

[0017] (2) The AKR1C1 protein of the present invention has high diagnostic value for osteoporosis in hypertensive patients. Through Logistic regression analysis, the present invention found that the expression level of the AKR1C1 protein was negatively correlated with the risk of osteoporosis, that is, as the expression level of the AKR1C1 protein increased, the risk of osteoporosis gradually decreased. Specifically, for every 1 unit increase in the standardized expression level of the AKR1C1 protein, the corresponding OR value was 0.080, indicating that for every 1 unit increase in the expression level of the AKR1C1 protein, the risk of osteoporosis decreased to 8% of the original value, and the AKR1C1 protein had a significant protective effect on osteoporosis.

[0018] (3) The AKR1C1 protein has high discriminative ability in the diagnosis of osteoporosis. Based on the analysis of the ROC curve, the present invention found that the area under the curve (AUC) of the AKR1C1 protein was 0.794, which was higher than the good standard of 0.70 in clinical diagnosis. At the same time, its diagnostic accuracy reached 76.25%. When the AKR1C1 protein of the present invention is used in the diagnosis of osteoporosis, rapid and accurate diagnosis of osteoporosis can be achieved through blood tests, avoiding the time and cost of traditional bone density tests. At the same time, it can guide clinicians to make personalized treatment decisions and avoid the risks and troubles brought by the further progression of osteoporosis to the health of patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention.

[0020] Figure 1 It is a Venn diagram of three machine learning algorithms for jointly screening core proteins.

[0021] Figure 2 It is a verification result graph of the importance of Boruta variables to the importance of core proteins.

[0022] Figure 3 It is a restricted cubic spline curve between the expression level of AKR1C1 protein and the OR value of osteoporosis.

[0023] Figure 4 It is the ROC curve of AKR1C1 protein in diagnosing osteoporosis. Detailed implementation manners

[0024] The technical solutions of the present invention will be described in conjunction with embodiments. However, the present invention is not limited to the following embodiments. The experimental methods and detection methods described in the following embodiments are all conventional methods unless otherwise specified; the reagents and materials are all commercially available unless otherwise specified.

[0025] Embodiment 1

[0026] This embodiment provides a screening process of AKR1C1 protein as a key protein for osteoporosis in hypertensive patients.

[0027] A proteomics study was carried out, and key proteins affecting the occurrence of osteoporosis were screened out by using various statistical methods of machine learning.

[0028] 1. Sample data acquisition: The data in this embodiment is from participants who visited the Xinjiang Hypertension Center from January 2021 to July 2023 and were clearly diagnosed with hypertension. All selected participants completed bone density tests.

[0029] 2. Construction of the original data set: 282 participants with normal bone mass and 166 participants with osteoporosis were selected. Using the propensity score matching method, 1:1 matching was performed according to age, gender, BMI, and menopause status. Finally, 40 participants with normal bone mass and 40 participants with osteoporosis were included, and the blood samples of these study populations were subjected to proteomics detection;

[0030] 3. Processing of the original data set: Based on the results of proteomics detection of the osteoporosis group and the control group, through sample exclusion (excluding samples with less than 100 identified proteins), protein exclusion (proteins with missing values in 50% of the samples), missing value filling (multiple imputation of missing values based on the random forest method), and data transformation (z-score transformation), a total of 132 up-regulated proteins and 71 down-regulated proteins were screened out;

[0031] 4. Machine learning screening: The method of machine learning is used to screen out key proteins related to osteoporosis density. Three machine learning algorithms, namely the Random Forest model, XGBoost model, and Support Vector Machine (SVM) in R language, are used to screen and analyze osteoporosis-related proteins based on the importance of variable features.

[0032] (1) The Random Forest method mainly evaluates based on the importance of the Gini index: The importance of each feature is evaluated by calculating the reduction of the Gini index when splitting nodes; the importance is sorted.

[0033] (2) XGBoost mainly evaluates the importance of variables through weights: ① Train the XGBoost model: Use the training data to construct multiple decision trees, and each tree optimizes the objective function based on the gradient descent method. ② Count the number of splits of each feature: For each tree, count the number of times each feature is used as a split node. ③ Accumulate the number of splits of each feature: Accumulate the number of splits of each feature in all trees to obtain the weight of the feature. ④ Normalize the feature importance: Normalize the weights of all features to obtain the relative importance of each feature. Sort and select based on the importance of these features.

[0034] (3) Support Vector Machine (SVM) evaluates the importance of variables based on Recursive Feature Elimination (RFE): The importance of features is evaluated by recursively removing the features that contribute the least to the model. In each iteration, use the SVM model to sort the features and remove the least important feature.

[0035] Based on the importance of variable features of the three machine learning algorithms, the proteins with the greatest impact on osteoporosis are selected and sorted from high to low, and the importance of all selected proteins is greater than 0. The top 30 key proteins of the Random Forest model are screened out, the top 26 key proteins of the XGBoost model are screened out, and the top 30 key proteins of the SVM are screened out. Take the intersection of the screening results of the three machine learning algorithms. The Venn diagram is as Figure 1 shown, and a set of core proteins with potential important roles in the diagnosis of osteoporosis are obtained (Table 1). Further analysis found that the AKR1C1 protein was commonly identified as one of the key proteins by the three methods.

[0036] 5. Verification of screening results: To verify the reliability of the screening results, we further verified the screened core proteins based on Boruta variable importance ( Figure 2 ), Figure 2 where the variable importance gradually increases from left to right, the red variables are used as rejected variables, the yellow variables are used as reference variables, and the green variables are used as important variables that can be included. Figure 2 The results show that the AKR1C1 protein is an important protein.

[0037] Table 1: Core proteins with potential roles in osteoporosis diagnosis

[0038] Protein Expression PDIA3 Down AKR1C1 Down CD58 Up HSPA1A Down IDH1 Down FLG2 Up KRT77 Up KRT5 Down

[0039] Example 2

[0040] This example provides the application and diagnostic value of AKR1C1 protein in an osteoporosis diagnosis model.

[0041] Through logistic regression analysis, it was found that the expression level of AKR1C1 protein was negatively correlated with the risk of osteoporosis, that is, as the expression level of AKR1C1 protein increased, the risk of osteoporosis gradually decreased. Specifically, for every 1-unit increase in the standardized expression level of AKR1C1 protein, the corresponding OR value was 0.080, indicating that for every 1-unit increase in the expression level of AKR1C1 protein, the risk of osteoporosis decreased to 8% of the original value. AKR1C1 protein has a significant protective effect on osteoporosis (Table 2).

[0042] To further evaluate the dose-response relationship between AKR1C1 protein and osteoporosis, the present invention constructed a restricted cubic spline curve ( Figure 3 ), and the restricted cubic spline curve showed that when the expression level of AKR1C1 protein was less than 0.68, the risk of osteoporosis increased significantly.

[0043] To further evaluate the diagnostic performance of AKR1C1 protein, we analyzed based on the ROC curve. The results showed that the area under the curve (AUC) of AKR1C1 protein was 0.794, higher than the good standard of 0.70 in clinical diagnosis, and its diagnostic accuracy reached 76.25% ( Figure 4 and Table 3). Therefore, the screened AKR1C1 protein has high discriminative ability in the diagnosis of osteoporosis, demonstrating its clinical application value as a potential diagnostic marker.

[0044] Table 2 Logistic regression of the expression level of AKR1C1 protein and the risk of osteoporosis

[0045] Protein OR 95% CI P AKR1C1 (per 1 unit increase) 0.080(0.030～0.210) <0.001

[0046] Table 3 Evaluation results of the diagnostic model

[0047] AUC threshold specificity sensitivity accuracy ppv npv 0.7941 0.9027467 0.6 0.925 0.7625 0.6981132 0.8888889

[0048] As described above, the basic principles, main features and advantages of the present invention have been preferably described. The above embodiments and the description are merely descriptions of the preferred embodiments of the present invention. The present invention is not limited by the above embodiments. Without departing from the spirit and scope of the present invention, various changes and improvements made by those of ordinary skill in the art to the technical solutions of the present invention shall fall within the protection scope determined by the present invention.

Claims

Use of AKR1C1 protein in the preparation of an osteoporosis diagnostic kit or diagnostic device.

2. The use according to claim 1, wherein The diagnostic kit is for hypertensive patients.

3. The use according to claim 1, wherein The diagnostic kit or diagnostic device comprises a reagent for detecting the expression level of AKR1C1 protein in a sample.

4. The use according to claim 3, characterized in that, The sample includes a plasma sample.

5. The use according to claim 1, wherein The expression level of the AKR1C1 protein is negatively correlated with the risk of osteoporosis.

6. An osteoporosis auxiliary diagnosis system, characterized in that, Comprising: An osteoporosis diagnostic model; An osteoporosis diagnostic model is constructed with the expression level of AKR1C1 protein in the sample as the input variable.

7. The osteoporosis auxiliary diagnosis system according to claim 6, wherein, The osteoporosis diagnostic model is: Risk value = Expression level of AKR1C1 protein; The higher the risk value, the lower the probability of having osteoporosis.

8. The osteoporosis auxiliary diagnosis system according to claim 7, wherein, The risk value is compared with a threshold value, and if it is lower than the threshold value, it is judged as having osteoporosis; The threshold value is 0.68.

Images