Autophagy biomarker related to depression, risk assessment model and application of autophagy biomarker

Abstract

The invention discloses an autophagy biomarker related to depression, a risk assessment model and application of the autophagy biomarker. A group of autophagy related biomarkers for auxiliary diagnosis of depression are screened out through multi-dimensional screening of large sample data and a machine learning algorithm, the autophagy related biomarkers comprise the following genes: GABARAPL2, RB1CC1, FOS and ULK1, and differential expression of the autophagy related biomarkers is proved in chronic stress induced depression-like mouse cortical brain tissues. The invention also provides a risk assessment model for auxiliary diagnosis of depression, a logistic regression model is constructed by taking the biomarker as a feature, a score formula and a risk boundary value of depression risk assessment are determined, and multi-aspect assessment is carried out on the prediction performance of the model. The prediction model has good prediction accuracy and calibration degree, can efficiently and conveniently realize early diagnosis of depression, and improves the accuracy of depression treatment.

Description

Technical Field

[0001] The present invention belongs to the technical field of bioinformatics, and specifically relates to an autophagy biomarker associated with depression, a risk assessment model and applications thereof. Background Art

[0002] Major depression disorder (MDD) is a common mental disorder characterized by high prevalence and high suicide rates. It is one of the five major causes of disability worldwide and is expected to top the global disease burden list by 2030. According to the World Health Organization (Global Mental Assessment published in 2019), the proportion of depression patients in the world reached 4.4% in 2015, or about 322 million people. At present, the most widely used diagnostic criteria for MDD worldwide are the Diagnostic and Statistical Manual of Mental Disorders (DSM5) and the International Classification of Diseases (ICD11). These diagnostic criteria rely on the subjective judgment of clinicians on symptomatology and always lack objective markers for disease diagnosis.

[0003] Autophagy is a process of cell self-degradation that can prevent the accumulation of toxic proteins and damaged organelles and protect cells from damage. It is an important mechanism for cell survival and homeostasis. Many studies have shown that autophagy genes are involved in the occurrence of various phenotypes and diseases such as neurological disorders. On the one hand, stable cell autophagy is essential for maintaining the function and activity of neuronal cells themselves and regulating synaptic homeostasis. At the same time, autophagy also affects the function and homeostasis of neurons by affecting the release of regulatory factors by glial cells. On the other hand, the "inflammatory hypothesis" and "neurotrophic hypothesis" are important mechanisms for the treatment of depression. Studies have reported that autophagy, as an emergency activated cell degradation pathway, has a very close bidirectional regulatory relationship with the inflammatory response and the production of inflammasomes. The autophagy signaling pathway can regulate the release of brain-derived neurotrophic factor (BDNF). A variety of antidepressants also have the effect of activating the autophagy pathway to regulate autophagy. More and more evidence shows that autophagy disorders participate in the occurrence and development of depression through multiple mechanisms and pathways.

[0004] Early diagnosis and appropriate treatment are important means to reduce the morbidity and mortality of depression. However, due to the complex causes of depression, great differences in symptoms and low identification rate of high-risk individuals, clinical judgment based on personal medical history lacks reliability. Therefore, an objective and reliable diagnostic method is urgently needed. Summary of the invention

[0005] Technical problem to be solved: In view of the problems in the prior art such as the lack of reliability and objectivity of clinical judgment based on personal medical history in the early screening and diagnosis of depression, the present invention provides an autophagy biomarker related to depression, a risk assessment model and its application, which can provide new methods and new ideas for the early screening and diagnosis of depression, and lay the foundation for early intervention treatment and timely correction of patients' dangerous behaviors to avoid the occurrence and further aggravation of depression.

[0006] Technical solution: In a first aspect, the present invention provides an autophagy biomarker associated with depression, wherein the marker comprises at least one of the following genes: GABARAPL2, RB1CC1, FOS and ULK1.

[0007] Preferably, the markers include the following genes: GABARAPL2, RB1CC1, FOS and ULK1.

[0008] The use of the above-mentioned autophagy biomarkers related to depression in the preparation of reagents or kits for diagnosing the occurrence of depression.

[0009] The above-mentioned autophagy biomarkers related to depression are used as targets for screening drugs for preventing or treating depression.

[0010] In a second aspect, the present invention provides a method for screening the above-mentioned depression biomarkers associated with autophagy, the steps of which are as follows:

[0011] Step 1: Data collection and processing: Obtain peripheral blood gene microarray datasets of depression patients and healthy controls, construct original datasets, including discovery cohort datasets and independent validation datasets, and perform data preprocessing to obtain input datasets based on the discovery cohort datasets;

[0012] Step 2: Differential gene analysis: Perform differential gene analysis on the input data set. Based on the linear regression model, use the weighted least squares method to calculate the difference in gene expression between depressed patients and healthy controls, and perform the Bayesian method to correct the problem of multiple testing. The sites with adjusted P value < 0.05 and |log2 Fold change|> 1 are regarded as differential gene sites. The corresponding differential gene names are obtained through annotation to obtain the differential gene data set.

[0013] Step 3: Weighted gene co-expression network analysis: Based on the input data set, a co-expression network was constructed, and modules with differences less than 0.25 were merged. The Pierce correlation coefficient between each module and the depressive phenotype was calculated. Modules with a Pearson correlation coefficient greater than 0.8 were selected, and the gene significance = 0.2 and module membership = 0.5 were used as thresholds to screen out the genes most related to depression.

[0014] Step 4: Take the intersection of the genes obtained in steps 2 and 3 and perform clustering, dimensionality reduction analysis and screening to obtain autophagy-related biomarkers for diagnosing depression. The clustering, dimensionality reduction and screening method comprises the following steps:

[0015] (1) The genes in the intersection were analyzed using the cytoHubba plug-in of the cytoscape software to perform the maximum cluster centrality topological analysis algorithm to analyze the gene association, calculate the maximum number of clusters to which the node belongs to evaluate the centrality of the node, and select the top 30 genes as candidate biomarkers;

[0016] (2) randomly dividing the input data set of the candidate biomarkers into a training set and a test set, executing the LASSO regression algorithm and the support vector machine recursive feature elimination algorithm on the training set to establish depression diagnostic models, respectively, and applying the established models to the test set and an independent validation data set, evaluating the model results, fitting the optimal model, and obtaining the screened genes;

[0017] (3) The genes screened out by the two algorithms in step (2) are intersected and the genes related to the autophagy pathway are screened out to obtain a corresponding list of important genes as autophagy-related biomarkers for diagnosing depression.

[0018] In a third aspect, the present invention provides a risk assessment model for autophagy biomarkers associated with depression, wherein the risk assessment model uses the above-mentioned autophagy biomarkers for judgment, and includes a data acquisition unit, a risk index calculation unit, and a risk judgment unit. The data acquisition unit includes: obtaining the gene expression level data of the above-mentioned autophagy-related biomarkers for auxiliary diagnosis of depression in the biological sample of the subject; the risk index calculation unit includes: based on the gene expression level obtained by the data acquisition unit, calculating the risk score of the subject suffering from depression through the risk assessment model for autophagy-related biomarkers for auxiliary diagnosis of depression, and the risk score calculation formula is:

[0019]

[0020] Wherein X1, X2, X3, X4 are the gene expression level data of biomarkers GABARAPL2, RB1CC1, ULK1, FOS, β1, β2, β3, β4 are the weight coefficients corresponding to the biomarkers, β1 = 2.1128, β2 = -0.6776, β3 = 0.8163, β4 = -1.7358, intercept α = -7.8361;

[0021] The risk judgment unit includes: according to the calculated depression risk score, a risk score < a model definition threshold is judged as low risk; a risk score ≥ a model definition threshold is judged as high risk, and the model threshold is 0.656.

[0022] Preferably, the biological sample comprises a blood sample.

[0023] In a fourth aspect, the present invention provides a method for constructing a risk assessment model of autophagy biomarkers associated with depression as described above, the construction method comprising the following steps:

[0024] (1) using the gene expression matrix of all samples of the biomarker as the original data set, the original data set including the discovery cohort data set and the independent validation data set, and performing data preprocessing on the original data set to obtain an input data set based on the discovery cohort data set, randomly dividing the input data set into a training set and a test set, using the biomarker as a feature, and constructing a logistic regression model with whether or not depression is the outcome;

[0025] (2) The logistic regression model estimates the regression coefficient by fitting the model sample data of the training set, thereby establishing the relationship between the independent variable and the target variable, and determining the score formula for depression risk assessment, where P is the risk score, and the depression risk assessment score formula is:

[0026]

[0027] Among them, α is the intercept, β1, β2, β3, β4 are the weight coefficients of GABARAPL2, RB1CC1, ULK1, and FOS, β1 = 2.1128, β2 = -0.6776, β3 = 0.8163, β4 = -1.7358, intercept α = -7.8361;

[0028] The performance of the autophagy-related biomarker risk assessment model for auxiliary diagnosis of depression was evaluated. The test set and independent validation data set were input into the constructed risk assessment model. Discrimination and calibration were used as performance metrics. Discrimination included consistency index and AUC, and calibration included goodness of fit test. Combined with accuracy, the prediction performance of the model was verified.

[0029] Discrimination refers to the ability of the model to correctly distinguish between individuals with high and low risk of outcome, that is, the ability of the model to correctly classify whether the research event has occurred. A discrimination greater than 0.75 indicates high accuracy. Calibration refers to the consistency between the predicted risk and the actual risk of the prediction model, which is an important dimension to measure the accuracy of the model prediction. A goodness of fit test P greater than 0.05 indicates that the model calibration is good. The accuracy of the auxiliary diagnosis model for depression constructed by the present invention can reach up to 85%, AUC = 93.6%, and consistency index

[0030] =0.94, goodness of fit test P = 0.15.

[0031] (3) Determine the model threshold. The calculation method of the model threshold is to evaluate through sensitivity, specificity and Youden index. The Youden index measures the ability of the classifier to maximize the true negative rate while maintaining a high true positive rate. The maximum value of the Youden index corresponds to the maximum diagnostic critical point of the method. The model definition threshold is the maximum value of the Youden index. The calculation formula is:

[0032]

[0033] Youden index = sensitivity + specificity - 1;

[0034] In the formula, TP stands for true positive, which refers to the number of patients diagnosed by the gold standard who are judged as patients by the diagnostic model; FN stands for false negative, which refers to the number of patients diagnosed by the gold standard who are judged as non-patients by the diagnostic model; FP stands for false positive, which refers to the number of non-patients diagnosed by the gold standard who are judged as patients by the diagnostic model; TN stands for true negative, which refers to the number of non-patients diagnosed by the gold standard who are judged as non-patients by the diagnostic model.

[0035] In a fifth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon; when the computer program is executed by a processor, the method for constructing the risk assessment model of the autophagy biomarkers associated with depression is implemented.

[0036] In a sixth aspect, the present invention provides an information data processing terminal for implementing the method for constructing the above-mentioned risk assessment model of autophagy biomarkers associated with depression.

[0037] Beneficial effects:

[0038] 1. The present invention uses weighted gene co-expression network analysis, maximum cluster centrality topology algorithm, and LASSO and support vector machine recursive feature elimination machine learning algorithm to perform multi-level analysis on gene transcription level data of a large sample of depression patients and normal controls, which can process high-dimensional gene expression data and improve the effect of feature selection and model training;

[0039] 2. The present invention has developed a risk prediction model for auxiliary diagnosis of depression based on gene expression levels, and has conducted multiple repeated tests. The AUC of the model is stable above 0.75, and can reach up to 0.95, with high accuracy and stability, which improves the reliability of early diagnosis of depression;

[0040] 3. The biological sample used in the present invention is peripheral blood, which is easy to collect and obtain, so the technical solution of the present invention is easier to promote.

[0041] 4. The present invention has proved through animal experiments that the expression changes of the four markers in the cerebral cortex of depressed mice are consistent with the peripheral blood of depressed patients, which further confirms the accuracy and stability of the present invention and has reliable diagnostic value. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 This is a schematic diagram of screening autophagy-related biomarkers for auxiliary diagnosis of depression in Example 1 of the present invention;

[0043] Figure 2 This is a schematic diagram of a risk assessment model for auxiliary diagnosis of depression based on biomarker gene expression levels in Example 2 of the present invention;

[0044] Figure 3 The diagnostic efficacy of the risk assessment model for auxiliary diagnosis of depression in Example 2 is verified. (a) is the ROC curve of the auxiliary diagnosis model for depression, where AUC = 0.936; (b) is the goodness of fit test result of the model, P = 0.15; (c) is the consistency index = 0.94, and the accuracy rate is 85%;

[0045] Figure 4 This is a schematic diagram of the structure of a depression risk assessment system in Example 3 of the present invention;

[0046] Figure 5 The figures are the test results of the depression animal model in Example 4 of the present invention, wherein (a) is a sugar water preference test; (b) is an open field test; (c) is a tail suspension test; and (d) is a forced swimming test;

[0047] Figure 6 It is a result diagram of the expression of GABARAPL2, RB1CC1, ULK1, and FOS. (a) is the distribution diagram of the biomarker gene expression level in human peripheral blood in Example 2, and (b) is the result diagram of the expression test of GABARAPL2, RB1CC1, ULK1, and FOS in the cerebral cortex tissue of the depression model mice in Example 4. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. The following embodiments are used to illustrate the present invention, but cannot be used to limit the scope of the present invention.

[0049] Example 1

[0050] like Figure 1 As shown, the present invention provides a group of autophagy-related biomarkers for auxiliary diagnosis of depression, which specifically include the following implementation steps:

[0051] 1. Data source: Data sets that meet the inclusion criteria in the GEO database were collected and sorted, and cases that did not meet the criteria were excluded (see Table 1 for exclusion criteria for irrelevant variables). The inclusion criteria were: (1) case-control design; (2) no other diseases besides depression; (3) no medication use.

[0052] Table 1 Exclusion of irrelevant variables

[0053]

[0054] The original data set was constructed, and a total of 6 data sets were included, including GSE201332, GSE217811, GSE38206, GSE76826, GSE39653 and GSE98793. The included data sets and sample information are shown in Table 2 below. The expression of the 6 data sets are all from human peripheral blood samples. Among them, GSE201332, GSE217811, GSE38206, GSE76826, and GSE39653 were used as discovery cohorts for biomarker screening data sets, including gene microarray data sets of 75 patients with depression and 70 healthy controls; GSE98793 was used as an independent validation data set, including gene microarray data sets of 64 healthy controls and 128 patients with depression.

[0055] Table 2 Datasets and sample information included in the invention

[0056]

[0057] 2. Preprocessing of the original data set: ① Use the arrayQulitMetrics package of R language to evaluate the quality of the data set, and use the Robust Multi-Array Average expression measure (RMA) to standardize the data to compensate for the impact of system deviation on the data; ② Discard sites with missing values ​​and probes mapped to multiple locations of the genome; ③ Use the empirical Bayes method to adjust the batch effect of the data to eliminate the batch effect caused by technical factors such as substrate chips and positions; ④ Filter out genes with expression levels similar to the background to obtain an input data set based on the discovery cohort data set. The classification of the data set in the present invention is as follows:

[0058] Table 3 Classification of datasets

[0059]

[0060] 3. Differential gene analysis: differential gene analysis was performed on the input data set. Based on the linear regression model, the weighted least squares method was used to calculate the difference in gene expression between depressed patients and healthy controls. The Bayesian method was used to correct the problem of multiple testing. The sites with adjusted P value < 0.05 and |log2 Fold change|> 1 were used as differential gene sites. A total of 2287 differential gene sites were obtained and the corresponding differential gene names were obtained through annotation to obtain the differential gene data set.

[0061] 4. Weighted gene co-expression network analysis: Based on the input data set, ① calculate the median difference of each gene in the data set, screen the top 500 genes and exclude missing values ​​and outliers; ② use R 2 =0.9 to construct a scale-free network, and β=6 was selected as the soft threshold according to scale independence and average connectivity; ③ Hierarchical clustering and dynamic shear tree were used to preliminarily construct network modules, and then modules with dissimilarity below 0.25 were merged to obtain 12 gene modules; ④ The Pierce correlation coefficient between each module and the depressive phenotype was calculated, and module genes with a Pearson correlation coefficient greater than 0.8 were selected: the correlation between turquoise module genes and MDD was =0.86, P<1e-200, and the correlation between brown module genes and MDD was =0.81, P≤4.5e-188; ⑤ According to gene significance (GS) =0.2 and module membership (MM) =0.5, a total of 2681 genes most related to depression were screened out from the two modules and the HUB genes of each module.

[0062] 5. Take the intersection of the genes obtained in steps 3 and 4 to obtain 302 genes that are highly correlated with depression.

[0063] 6. Maximum cluster centrality topology analysis algorithm analysis: The above 302 genes were analyzed using the cytoHubba plug-in of the cytoscape software to execute the maximum cluster centrality topology analysis algorithm, analyze the gene association, calculate the maximum number of clusters to which the node belongs to evaluate the centrality of the node, and select the top 30 genes as candidate biomarkers.

[0064] 7. Construction of machine learning model: ① The input data set of the candidate biomarkers was stratified and randomly sampled into a training set and a test set according to a ratio of 0.8:0.2; ② LASSO algorithm: First, the binomial LASSO regression algorithm was executed on the training set, and the constructed model was applied to the test set for verification, the ROC curve was drawn and the AUC was calculated for evaluation, and the iteration was repeated 1000 times to fit the best model and detect the hub genes significantly related to depression; ③ SVM-RFE algorithm: The support vector machine recursive feature elimination algorithm was executed on the training set at the same time to extract features, fit the model and evaluate the performance of the data, the constructed model was verified in the test set, and 10-fold cross validation was performed to screen out the optimal subset, fit the optimal model, screen the core genes, and improve the stability and accuracy of the model.

[0065] 8. Determine autophagy-related biomarkers for auxiliary diagnosis of depression: ① Take the intersection of the genes identified by the two algorithms ② and ③ in step 7 to obtain the 7 most important depression-related genes; ② Screen the 7 genes among 803 autophagy-related genes to obtain 4 autophagy-related biomarkers for diagnosing depression. The ENTREZ IDs of the 4 autophagy-related biomarkers GABARAPL2, RB1CC1, ULK1, and FOS for diagnosing depression are GABARAPL2: Entrez Gene ID: 11345; RB1CC1: Entrez Gene ID: 9821; ULK1: Entrez Gene ID: 8408; FOS: Entrez Gene ID: 2353.

[0066] Example 2

[0067] like Figure 2-3 As shown, based on the biomarkers screened in Example 1, a risk assessment model for auxiliary diagnosis of depression is constructed and the accuracy of the model is evaluated, which specifically includes the following implementation steps;

[0068] 1. The input data set of the four biomarkers (i.e., the input data set in Example 1) was divided into a training set and a test set by stratified random sampling at a ratio of 0.8:0.2, and the four biomarkers were used as features, and whether or not depression was used as the outcome, and a binary logistic regression model was constructed using Python;

[0069] 2. The logistic regression model estimates the regression coefficient by fitting the model sample data of the training set, thereby establishing the relationship between the independent variable and the target variable, and determining the score formula for depression risk assessment. The depression risk assessment score P formula is:

[0070]

[0071] Among them, β1, β2, β3, and β4 are the weight coefficients of GABARAPL2, RB1CC1, ULK1, and FOS, respectively, where β1 = 2.1128, β2 = -0.6776, β3 = 0.8163, β4 = -1.7358, and the intercept α = -7.8361.

[0072] 3. Model performance evaluation: The test set and the independent validation dataset GSE98793 were input into the constructed risk assessment model, and the discrimination (consistency index and AUC) and calibration (goodness of fit test) were used as performance metrics, combined with the accuracy, to verify the model prediction performance (results see Figure 3 ).

[0073] Discrimination refers to the ability of the model to correctly distinguish between individuals at high risk and individuals at low risk of outcomes, that is, the ability of the model to correctly classify whether the research event has occurred. A discrimination greater than 0.75 indicates high accuracy. Calibration refers to the degree of consistency between the predicted risk and the actual risk of the prediction model, which is an important dimension to measure the accuracy of model prediction. A goodness of fit test P greater than 0.05 indicates that the model has good calibration.

[0074] The accuracy of the auxiliary diagnosis model for depression constructed by the present invention can reach up to 85%, AUC = 93.6%, consistency index = 0.94, and goodness of fit test P = 0.15. It is significantly better than the single gene diagnosis efficiency: ULK1 AUC = 0.708, FOS AUC = 0.636, GABARAPL2 AUC = 0.588, RB1CC1 AUC = 0.587.

[0075] 4. Determine the model threshold. The calculation method of the model threshold is to evaluate through sensitivity, specificity and Youden index. Youden index measures the ability of the classifier to maximize the true negative rate while maintaining a high true positive rate. The maximum value of Youden index corresponds to the maximum diagnostic critical point of the method. The model definition threshold is the maximum value of Youden index. The calculation formula is:

[0076]

[0077] Youden index = sensitivity + specificity - 1;

[0078] TP stands for true positive, which refers to the number of patients diagnosed by the gold standard who were judged as patients by the diagnostic model; FN stands for false negative, which refers to the number of patients diagnosed by the gold standard who were judged as non-patients by the diagnostic model; FP stands for false positive, which refers to the number of non-patients diagnosed by the gold standard who were judged as patients by the diagnostic model; TN stands for true negative, which refers to the number of non-patients diagnosed by the gold standard who were judged as non-patients by the diagnostic model.

[0079] In the present invention, the threshold value (0.656) of the maximum Youden index (0.869) is selected as the model critical value.

[0080] Example 3

[0081] like Figure 4 As shown, the present invention provides a system (risk assessment model) for risk assessment of autophagy-related biomarkers for auxiliary diagnosis of depression, comprising:

[0082] A: a data acquisition unit, which acquires gene expression level data of autophagy-related biomarkers for auxiliary diagnosis of depression in a biological sample of a subject;

[0083] B: Risk index calculation unit, which can calculate the risk score of the subject suffering from depression based on the gene expression level obtained by the data acquisition unit through the risk assessment model of autophagy-related biomarkers for auxiliary diagnosis of depression. The risk score calculation formula is:

[0084]

[0085] Wherein X1, X2, X3, X4 are the gene expression level data of biomarkers GABARAPL2, RB1CC1, ULK1, FOS, β1, β2, β3, β4 are the weight coefficients corresponding to the biomarkers, β1 = 2.1128, β2 = -0.6776, β3 = 0.8163, β4 = -1.7358, intercept α = -7.8361;

[0086] C: Risk judgment unit. This unit uses the calculated depression risk score. If the risk score is less than the model threshold, it is judged as low risk; if the risk score is greater than or equal to the model threshold, it is judged as high risk. The model threshold is 0.656.

[0087] Example 4 Animal Model Verification

[0088] An animal model of depression was established and behavioral tests (sugar preference, open field, tail suspension, forced swimming) were used to evaluate the animal model of depression. Figure 5 As shown in the figure. The results in the figure show that the mouse's preference for sugar water decreased, indicating that it showed symptoms of pleasure. The immobility time of tail suspension and forced swimming decreased, showing a despair phenotype. The time spent in the central area of ​​the open field decreased, showing anxious behavior. In summary, the mouse depression-like model was successfully established.

[0089] The primer sequences of GABARAPL2, RB1CC1, ULK1, and FOS were designed and synthesized. The cortical brain tissues of depression mice (CUMS) and normal mice (Control) were taken, and the relative mRNA expression of the markers was detected by RT-qPCR. The primer sequences are as follows:

[0090] GABARAPL2-F: 5'-GGAACCTGTCCTCTGGGAATC-3' (SEQ ID No.1)

[0091] GABARAPL2-R: 5'-CAGAACTATGCATCAGCCCCT-3' (SEQ ID No.2)

[0092] RB1CC1-F: 5'-ACGTGGCAAAGAACTTAGGG-3' (SEQ ID No.3)

[0093] RB1CC1-R: 5'-ACGTGGCAAAGAACTTAGGG-3' (SEQ ID No.4)

[0094] ULK1-F: 5'-GTCGACACCGCGAGAAGC-3' (SEQ ID No.5)

[0095] ULK1-R: 5'-GAAGTCATACAGCGCCACGA-3' (SEQ ID No.6)

[0096] FOS-F: 5'-GGGAATGGTGAAGACCGTGTCA-3' (SEQ ID No.7)

[0097] FOS-R: 5'-GCAGCCATCTTATTCCGTTCCC-3' (SEQ ID No.8)

[0098] GAPDH-F: 5'-GGTTGTCTCCTGCGACTTCA-3' (SEQ ID No.9)

[0099] GAPDH-R: 5'-TGGTCCAGGGTTTCTTACTCC-3' (SEQ ID No.10)

[0100] The detection method is as follows:

[0101] 1. Total RNA extraction from mouse cortical brain tissue

[0102] Take 100 mg of mouse cortical brain tissue, let it stand at room temperature for half an hour, then add 1 mL of Trizol and vortex for 30 seconds, and let it stand at room temperature for 3 minutes; (2) Add 200 μL of chloroform, shake vigorously for 15 seconds, and let it stand at room temperature for 15 minutes; (3) Centrifuge at 12000g, 4℃ for 15 minutes; (4) Use a pipette to carefully aspirate the upper clear aqueous phase, transfer it to a 1.5 mL nuclease-free centrifuge tube, add 1.5 times the volume of anhydrous ethanol, and gently mix it upside down; (5) Isopropanol precipitation: add 500 ul of isopropanol, mix it upside down 10 times, let it stand at -20℃ for 20 minutes; centrifuge it at 4℃ for 12000g×10min. (6) Ethanol washing: discard the supernatant, gently add 1.5 mL of 75% ethanol, gently mix it upside down, and centrifuge it at 4℃ for 7500g×5min. (7) Discard the supernatant, dry the precipitate and then dissolve: remove the supernatant as much as possible, dry at room temperature for 5-10 minutes, and dissolve with 50ul of DEPC treated H2O. (8) Measure the RNA concentration using a One Drop spectrophotometer and proceed to the next step or store in a -80℃ refrigerator for later use.

[0103] 2. RNA Reverse Transcription

[0104] Reverse transcription kit was used using HiScript Q RT SuperMix for qPCR Kit (R123-01, Vazyme)

[0105] a) Genomic DNA removal: RNATotal RNA: 200 ng; 4×gDNAwiper Mix 4 μL; DNase / RNasefree ddH2O Up to 16 μL; Use a pipette to gently pipette and mix, 42°C, 2 min;

[0106] b) Reverse transcription reaction: 16 μL of the first step reaction solution; 4 μL of 5×HiScript III qRT SuperMix; use a pipette to gently pipette to mix, and then run the following program in the PCR instrument: 37°C for 15 min; 85°C for 5 sec; the obtained product can be directly used for qPCR reaction or stored in a -20°C refrigerator.

[0107] 3.qPCR

[0108] use Universal SYBR qPCR Master Mix (Q511, Vazyme) Kit:

[0109] a) Reaction system:

[0110] 2×AceQ qPCR SYBR Green Master Mix 10μL;

[0111] Forward Primer (10μM)0.4μL;

[0112] Reverse Primer (10μM)0.4μL;

[0113] Template DNA 2 μL;

[0114] DNase / RNase free ddH2O Up to 20μL.

[0115] b) Reaction conditions:

[0116] Pre-denaturation: 95℃5min

[0117] Cycle reaction (45 times): 95℃10s; 66℃30s

[0118] Melting curve: 95℃15s; 60℃1min; 95℃15s.

[0119] 4. Data Processing

[0120] The qPCR data were obtained, and the relative content of the target gene mRNA was calculated using GAPDH as the internal reference. The depression group was then compared with the normal control group to calculate ΔΔCT. Finally, 2 -ΔΔCT Standardized analysis was performed using the method.

[0121] The results show Figure 6 As shown in (b), compared with normal mice, the expression of GABARAPL2, RB1CC1, ULK1, and FOS in the cortical tissue of depression model mice was significantly increased, which was consistent with the expression level changes of the four biomarker genes in the peripheral blood of depression patients in the discovery dataset ( Figure 6 (a)) are consistent. In the figure, MDD represents depressed patients, CUMS represents depressed mice, and control represents normal controls.

[0122] In summary, the present invention provides an autophagy-related biomarker and risk assessment model for auxiliary diagnosis of depression constructed at the gene expression level through multidimensional screening of large sample data and machine learning algorithms, and constructs a technical solution based on autophagy-related differential gene expression levels as biomarkers for diagnosis of depression. This solution has high accuracy and stability, and has important application value for early diagnosis and treatment of depression.

[0123] The above adjustment parameters of the present invention are merely examples for explaining the present invention, and are not intended to limit the implementation methods of the present invention. For those skilled in the art, other different forms of changes and modifications can be made based on the above description. It is impossible to list all implementation methods here. Any obvious changes or modifications derived from the technical solution of the present invention are still within the scope of protection of the present invention.

Claims

1. An autophagy biomarker associated with depression, characterized in that: The markers include the following genes: At least one of GABARAPL2, RB1CC1, FOS and ULK1.

2. The autophagy biomarker according to claim 1, characterized in that The markers include the following genes: GABARAPL2, RB1CC1, FOS and ULK1.

3. Use of the autophagy biomarker associated with depression according to claim 1 in the preparation of a reagent or kit for diagnosing the occurrence of depression.

4. Use of the depression-related autophagy biomarker of claim 1 as a target for screening drugs for preventing or treating depression.

5. A method for screening biomarkers of depression associated with autophagy according to claim 1, characterized in that: Here are the steps: Step 1: Data collection and processing: Obtain peripheral blood gene microarray datasets of depression patients and healthy controls, construct original datasets, which include discovery cohort datasets and independent validation datasets, and perform data preprocessing on the original datasets to obtain input datasets based on the discovery cohort datasets; Step 2: Differential gene analysis: Perform differential gene analysis on the input data set. Based on the linear regression model, use the weighted least squares method to calculate the difference in gene expression between depressed patients and healthy controls, and perform the Bayesian method to correct the problem of multiple testing. The sites with adjusted P value < 0.05 and |log2 Fold change|> 1 are regarded as differential gene sites. The corresponding differential gene names are obtained through annotation to obtain the differential gene data set. Step 3: Weighted gene co-expression network analysis: Based on the input data set, a co-expression network was constructed, and modules with differences less than 0.25 were merged. The Pierce correlation coefficient between each module and the depressive phenotype was calculated. Modules with a Pearson correlation coefficient greater than 0.8 were selected, and the gene significance = 0.2 and module membership = 0.5 were used as thresholds to screen out the genes most related to depression. Step 4: Take the intersection of the genes obtained in steps 2 and 3 and perform clustering, dimensionality reduction analysis and screening to obtain autophagy-related biomarkers for diagnosing depression. The clustering, dimensionality reduction and screening method comprises the following steps: (1) The genes in the intersection were analyzed using the cytoHubba plug-in of the cytoscape software to perform the maximum cluster centrality topological analysis algorithm to analyze the gene association, calculate the maximum number of clusters to which the node belongs to evaluate the centrality of the node, and select the top 30 genes as candidate biomarkers; (2) randomly dividing the input data set of the candidate biomarkers into a training set and a test set, executing the LASSO regression algorithm and the support vector machine recursive feature elimination algorithm on the training set to establish depression diagnostic models, respectively, and applying the established models to the test set and an independent validation data set, evaluating the model results, fitting the optimal model, and obtaining the screened genes; (3) Take the intersection of the genes screened by the two algorithms in step (2) and screen out the genes related to the autophagy pathway, and obtain The corresponding important gene list was obtained as autophagy-related biomarkers for diagnosing depression.

6. A risk assessment model for autophagy biomarkers associated with depression, characterized in that: The risk assessment model uses the autophagy biomarker as claimed in claim 2 for judgment, including a data acquisition unit, a risk index calculation unit and a risk judgment unit, wherein the data acquisition unit includes: obtaining the gene expression level data of the autophagy-related biomarker for auxiliary diagnosis of depression as claimed in claim 2 in the biological sample of the subject; the risk index calculation unit includes: Based on the gene expression levels obtained by the data acquisition unit, the risk score of the subject suffering from depression is calculated by the risk assessment model of autophagy-related biomarkers for auxiliary diagnosis of depression. The risk score calculation formula is: Wherein X1, X2, X3, X4 are the gene expression level data of biomarkers GABARAPL2, RB1CC1, ULK1, FOS, β1, β2, β3, β4 are the weight coefficients corresponding to the biomarkers, β1 = 2.1128, β2 = -0.6776, β3 = 0.8163, β4 = -1.7358, intercept α = -7.8361; The risk judgment unit includes: according to the calculated depression risk score, a risk score < a model definition threshold is judged as low risk; a risk score ≥ a model definition threshold is judged as high risk, and the model threshold is 0.

656.

7. A risk assessment model for autophagy biomarkers associated with depression according to claim 6, characterized in that: The biological sample includes a blood sample.

8. A method for constructing a risk assessment model of autophagy biomarkers associated with depression as claimed in claim 6, characterized in that: The construction method comprises the following steps: (1) using the gene expression matrix of all samples of the biomarker as the original data set, the original data set including the discovery cohort data set and the independent validation data set, and performing data preprocessing on the original data set to obtain an input data set based on the discovery cohort data set, randomly dividing the input data set into a training set and a test set, using the biomarker as a feature, and constructing a logistic regression model with whether or not depression is the outcome; (2) The logistic regression model estimates the regression coefficient by fitting the model sample data of the training set, thereby establishing the relationship between the independent variable and the target variable, and determining the score formula for depression risk assessment, where P is the risk score, and the depression risk assessment score formula is: Among them, α is the intercept, β1, β2, β3, β4 are the weight coefficients of GABARAPL2, RB1CC1, ULK1, and FOS respectively. β1=2.1128, β2=-0.6776, β3=0.8163, β4=-1.7358, intercept α=-7.8361; (3) To evaluate the performance of the autophagy-related biomarker risk assessment model for auxiliary diagnosis of depression, the test set and the independent validation dataset were input into the constructed risk assessment model, and discrimination and calibration were used as performance metrics. The discrimination included consistency index and AUC, and the calibration included goodness of fit test. Combined with the accuracy, the prediction performance of the model was verified; (4) Determine the model threshold. The calculation method of the model threshold is to evaluate it through sensitivity, specificity and Youden index. The Youden index measures the ability of the classifier to maximize the true negative rate while maintaining a high true positive rate. The maximum value of the Youden index corresponds to the maximum diagnostic critical point of the method. The model definition threshold is the maximum value of the Youden index. The calculation formula is: Youden index = sensitivity + specificity - 1; In the formula, TP stands for true positive, which refers to the number of patients diagnosed by the gold standard who are judged as patients by the diagnostic model; FN stands for false negative, which refers to the number of patients diagnosed by the gold standard who are judged as non-patients by the diagnostic model; FP stands for false positive, which refers to the number of non-patients diagnosed by the gold standard who are judged as patients by the diagnostic model; TN stands for true negative, which refers to the number of non-patients diagnosed by the gold standard who are judged as non-patients by the diagnostic model.

9. A computer-readable storage medium, characterized in that: The computer readable storage medium stores a computer program; When the computer program is executed by a processor, the method for constructing a risk assessment model of autophagy biomarkers associated with depression as described in claim 8 is implemented.

10. An information data processing terminal, characterized in that: A method for constructing a risk assessment model for autophagy biomarkers associated with depression as described in claim 8.

Images