Prediction method for gestational diabetes mellitus

Abstract

The invention provides an application of a reagent for detecting a biomarker in preparation of a product for predicting gestational diabetes mellitus, the product for predicting gestational diabetes mellitus, and a method, equipment and system for predicting gestational diabetes mellitus. The system comprises an input module and an output module, wherein the input module is used for inputting the levels of biomarkers and basic clinical indicators in a biological sample derived from a subject; and the output module is used for indicating whether the subject has the risk of suffering from gestational diabetes mellitus or not. The system improves the accuracy of gestational diabetes prediction, is economical, feasible, rapid and accurate, and is suitable for clinical application.

Description

Technical Field

[0001] The present application relates to the field of bioinformatics, and specifically to the use of a reagent for detecting biomarkers in the preparation of a product for predicting gestational diabetes, a product for predicting gestational diabetes, a method, device and system for predicting gestational diabetes. Background Art

[0002] Gestational diabetes mellitus (GDM) is one of the common complications of pregnancy, which not only seriously threatens the health of pregnant women, but also brings health risks to offspring, such as preeclampsia, miscarriage, premature birth and fetal malformation. Therefore, identifying risk factors for predicting GDM and building an effective early prediction model for GDM have important clinical significance and value for reducing the incidence of GDM and alleviating the severity of the disease.

[0003] Candidate risk factors for GDM include traditional risk factors (such as maternal age and body mass index), biomarkers detected by laboratory tests (such as white blood cell count and hemoglobin), and genetic factors. Most prediction models in related technologies usually use traditional risk factors and biomarkers as candidate risk factors, and the accuracy of a few prediction models that take genetic factors into account is relatively low.

[0004] Therefore, there is an urgent need to provide a low-cost, highly accurate method for predicting early GDM that is suitable for large-scale clinical applications. Summary of the invention

[0005] The present invention aims to solve one of the technical problems in the related art to at least a certain extent. To this end, the embodiments of the present invention propose an application of a reagent for detecting a biomarker in the preparation of a product for predicting gestational diabetes, a product for predicting gestational diabetes, a method for predicting gestational diabetes, a device and a system. The system includes an input and output module, the input module is used to input the levels of biomarkers and basic clinical indicators in a biological sample from a subject; the output module is used to indicate whether the subject is at risk of gestational diabetes. On the basis of traditional risk factors, the system adds 5 basic clinical indicators that are highly correlated with GDM and GDM biomarker SNPs discovered based on the Chinese population, which improves the accuracy of gestational diabetes prediction, and is economically feasible, fast and accurate, and suitable for clinical application.

[0006] In a first aspect, an embodiment of the present application proposes the use of a reagent for detecting a biomarker in the preparation of a product for predicting gestational diabetes, wherein the biomarker includes rs3781637 and / or rs35261542.

[0007] In some embodiments, the reagent for detecting the biomarker is selected from: a probe that specifically recognizes rs3781637 and / or rs35261542; a primer that specifically recognizes rs3781637 and / or rs35261542; or an antibody or ligand that specifically binds to rs3781637 and / or rs35261542.

[0008] In some embodiments, when the base type of rs3781637 is C, it indicates that the subject has a risk of gestational diabetes; and / or when the base type of rs35261542 is A, it indicates that the subject has a risk of gestational diabetes.

[0009] In some embodiments, the product for predicting gestational diabetes mellitus further comprises a basic clinical indicator determination reagent, wherein the basic clinical indicator determination reagent is used to detect one or more of the following values: fasting blood glucose, γ-glutamyl transferase, prealbumin, urine glucose and uric acid.

[0010] A second aspect of an embodiment of the present application proposes a product for predicting gestational diabetes, the product comprising: a reagent for detecting a biomarker, the biomarker comprising rs3781637 and / or rs35261542.

[0011] In some embodiments, the agent is selected from:

[0012] A probe specifically recognizing rs3781637 and / or rs35261542;

[0013] Primers that specifically recognize rs3781637 and / or rs35261542; or

[0014] An antibody or ligand that specifically binds to rs3781637 and / or rs35261542.

[0015] In some embodiments, the product further comprises a basic clinical index determination reagent, wherein the basic clinical index determination reagent is used to detect one or more of the following values: fasting blood glucose, γ-glutamyl transferase, prealbumin, urine glucose and uric acid.

[0016] The third aspect of the embodiments of the present application proposes a method for predicting gestational diabetes, the method comprising: using biomarkers and optional basic clinical indicators as predictive variables, calculating a prediction result based on the levels of the biomarkers and optional basic clinical indicators in a biological sample from a subject; and indicating whether the subject is at risk of gestational diabetes based on the prediction result, wherein the biomarkers include rs3781637 and / or rs35261542, and the basic clinical indicators include one or more of the following: age, systolic blood pressure, diastolic blood pressure, pre-pregnancy weight, whether it is the first pregnancy, fasting blood glucose, γ-glutamyl transferase, prealbumin, urine sugar and uric acid.

[0017] In some embodiments, based on the prediction variables, the prediction result is calculated using a gestational diabetes prediction model, wherein the gestational diabetes prediction model is a Bayesian generalized linear model, and the model formula is:

[0018] y=β 0 +β 1 x 1 +…+β i x i ,

[0019] Among them, β 0 =-17.993153, β 1 =0.003151, β 2 =0.022021, β 3 =0.007268, β 4 =0.104299, β 5 =0.017977, β 6 =0.913049, β 7 =0.517115, β 8 =0.005883, β 9 =0.025542, β 10 =0.005147, β 11 =1.473156, β 12 =-0.199924,

[0020] Among them, x 1 is the value of body mass index, x 2 is the value of systolic blood pressure, x 3 is the value of diastolic pressure, x 4 is the value of age, x 5 Is the value of whether it is the first pregnancy, x 6 is the value of fasting blood glucose, x 7 is the value of urine sugar, x 8 is the value of uric acid, x 9is the value of γ-glutamyl transferase, x 10 is the value of prealbumin, x 11 is the value of rs3781637, x 12 The value of rs35261542.

[0021] In some embodiments, based on the prediction result being greater than a cutoff value of -2.105, it indicates that the subject is at risk of having gestational diabetes; based on the prediction result being less than a cutoff value of -2.105, it indicates that the subject is not at risk of having gestational diabetes.

[0022] The fourth aspect of the embodiments of the present application proposes a device for predicting gestational diabetes, the device comprising: a calculation module for using biomarkers and optional basic clinical indicators as predictive variables, and calculating the prediction results based on the levels of the biomarkers and optional basic clinical indicators in the biological sample from the subject; and an indication module for indicating whether the subject is at risk of gestational diabetes based on the prediction results, wherein the biomarkers include rs3781637 and / or rs35261542, and the basic clinical indicators include one or more of the following: age, systolic blood pressure, diastolic blood pressure, pre-pregnancy weight, whether it is the first pregnancy, fasting blood sugar, γ-glutamyl transferase, prealbumin, urine sugar and uric acid. In some embodiments, the calculation module is also used to: calculate the prediction results based on the predictive variables using a gestational diabetes prediction model, wherein the gestational diabetes prediction model is

[0023] y=β 0 +β 1 x 1 +…+β i x i ,

[0024] Among them, β 0 =-17.993153, β 1 =0.003151, β 2 =0.022021, β 3 =0.007268, β 4 =0.104299, β 5 =0.017977, β 6 =0.913049, β 7 =0.517115, β 8 =0.005883, β 9 =0.025542, β 10 =0.005147, β 11 =1.473156, β 12 =-0.199924,

[0025] Among them, x 1 is the value of body mass index, x 2 is the value of systolic blood pressure, x 3 is the value of diastolic pressure, x 4 is the value of age, x 5 Is the value of whether it is the first pregnancy, x 6 is the value of fasting blood glucose, x 7 is the value of urine sugar, x 8 is the value of uric acid, x 9 is the value of γ-glutamyl transferase, x 10 is the value of prealbumin, x 11 is the value of rs3781637, x 12 The value of rs35261542.

[0026] In some embodiments, the indication module is specifically used to: indicate that the subject is at risk of gestational diabetes based on the prediction result being greater than a cutoff value of -2.105; and indicate that the subject is not at risk of gestational diabetes based on the prediction result being less than a cutoff value of -2.105.

[0027] The fifth aspect of the embodiments of the present application proposes a system for predicting gestational diabetes, the system comprising: a processor; an input module for inputting the levels of biomarkers and optional basic clinical indicators in a biological sample from a subject, wherein the biomarkers include rs3781637 and / or rs35261542, and the basic clinical indicators include one or more of the following: age, systolic blood pressure, diastolic blood pressure, pre-pregnancy weight, whether it is the first pregnancy, fasting blood glucose, γ-glutamyl transferase, prealbumin, urine glucose and uric acid; a computer-readable medium containing instructions, which, when executed by the processor, implement the method for predicting gestational diabetes in a subject as described in the embodiment of the third aspect; and an output module for indicating whether the subject is at risk of developing gestational diabetes.

[0028] The sixth aspect of the embodiments of the present application proposes a computer-readable storage medium, characterized in that computer program instructions are stored on the computer-readable storage medium, and when the computer program instructions are executed by a processor, the method for predicting gestational diabetes as described in the third aspect of the embodiments is implemented.

[0029] A seventh aspect of the embodiments of the present application proposes a computer program product, including a computer program, which, when executed by a processor, implements the method for predicting gestational diabetes as described in the third aspect of the embodiments.

[0030] An eighth aspect of the embodiments of the present application proposes a computer program, which includes computer program code. When the computer program code is run on a computer, the computer executes the method for predicting gestational diabetes as described in the third aspect of the embodiments.

[0031] The technical solution according to the embodiment of the present invention has the following beneficial technical effects:

[0032] (1) In addition to traditional risk factors, five basic clinical indicators highly correlated with GDM and GDM biomarker SNPs discovered based on the Chinese population were added to improve the accuracy of gestational diabetes prediction, with an average accuracy of over 76%.

[0033] (2) All information comes from routine physical examinations of pregnant women, which will not increase the medical burden of pregnant women. It is economically feasible, fast, accurate, and suitable for clinical application.

[0034] (3) The AUC of the disease prediction model of this application is calculated based on ten-fold cross validation and 100 repeated mean values, so it has high repeatability and is conducive to stable promotion and application. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 It is a flow chart of the method for constructing a disease prediction model in an embodiment of the present application.

[0036] Figure 2 It is a schematic diagram of the structure of a disease prediction device according to an embodiment of the present application.

[0037] Figure 3 It is a schematic diagram of the workflow of the disease prediction model of the embodiment of the present application.

[0038] Figure 4 It is a schematic diagram of the regression analysis results of the laboratory clinical variables and GDM in the embodiment of the present application.

[0039] Figure 5 It is a schematic diagram of the genome-wide association analysis quantile (QQ) of GDM in the embodiment of the present application.

[0040] Figure 6 It is a Manhattan diagram of the genome-wide association analysis of GDM in the embodiment of the present application.

[0041] Figure 7 It is a partial enlarged schematic diagram of the CDKAL1 gene region of the embodiment of the present application.

[0042] Figure 8 It is a partial enlarged schematic diagram of the MTNR1B gene region of the embodiment of the present application.

[0043] Fig. 9It is a schematic diagram of the AUC value distribution results of the disease prediction model in the embodiment of the present application.

[0044] Fig.10 It is a schematic diagram of the ROC curve of the fitting model when the optimal parameters are repeated for the 100th time in the disease prediction model of the embodiment of the present application. DETAILED DESCRIPTION

[0045] Embodiments of the present invention are described in detail below, and examples of the embodiments are shown in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and intended to be used to explain the present invention, but should not be understood as limiting the present invention.

[0046] Gestational diabetes is a common complication of pregnancy, which not only seriously threatens the health of pregnant women, but also brings health risks to offspring, such as preeclampsia, miscarriage, premature birth and fetal malformations.

[0047] Traditional risk factors for gestational diabetes include basic clinical information of the pregnant woman, such as age, systolic blood pressure, diastolic blood pressure, pre-pregnancy weight, body mass index, overweight, and whether it is the first pregnancy.

[0048] The basic clinical indicators of laboratory tests for gestational diabetes include indicators of metabolism, liver function, urine routine, kidney function, blood indicators, protein, hormones, infection, electrolytes, etc. In some embodiments, the basic clinical indicators include 93 types, specifically urine cast examination, urine epithelial cell count, urine conductivity, crystals in urine, percentage of lysed red blood cells in urine, renal tubular epithelial cells in urine, percentage of unlysed red blood cells in urine, urinary tract infection, urine protein, bacterial count in urine, urine bilirubin, urine occult blood, urine ketone bodies, urine leukocyte esterase, urine nitrite, urine red blood cell count, urine clarity, urine glucose, urine pH, urine specific gravity, urobilinogen, urine white blood cell count, urine fiber, urine Yeast-like cells, ferritin, albumin, albumin / globulin ratio, globulin, total protein, glycated hemoglobin, vitamin D, oral glucose tolerance test-2 hours, oral glucose tolerance test-1 hour, oral glucose tolerance test-fasting, blood glucose, total bile acid, alanine aminotransferase, alkaline phosphatase, aspartate aminotransferase, direct bilirubin, prealbumin, total bilirubin, gamma-glutamyl transferase, cystatin C, blood urea nitrogen, serum creatinine, uric acid, herpes simplex virus type 2-specific IgG antibody, herpes simplex virus type 2 Specific IgM antibody, cytomegalovirus IgG antibody, cytomegalovirus IgM antibody, Toxoplasma IgG antibody, Toxoplasma IgM antibody, Rubella IgG antibody, Rubella IgM antibody, C-reactive protein, hepatitis B core antibody, hepatitis B envelope antibody, hepatitis B envelope antigen, hepatitis B surface antibody, hepatitis C virus antibody, beta-human chorionic gonadotropin, free thyroxine, free triiodothyronine, thyroid stimulating hormone, basophil count, basophil percentage, eosinophil count, eosinophil percentage Cell percentage, lymphocyte count, lymphocyte percentage, monocyte count, monocyte percentage, neutrophil count, neutrophil percentage, platelet hematocrit, platelet-large cell ratio, mean platelet volume, platelet volume distribution width, red blood cell distribution width-standard deviation, hematocrit, hemoglobin, mean corpuscular hemoglobin, mean corpuscular hemoglobin concentration, mean corpuscular volume, platelet count, red blood cell count, red blood cell distribution width-coefficient of variation, white blood cell count, calcium, chloride, potassium, sodium.

[0049] The occurrence of disease is usually determined by the interaction of environmental and genetic factors. With the development of sequencing technology, the severity of the adverse effects of genetic risk factors for disease on a person's overall health can be measured to a certain extent, such as through polygenic risk scores (PRS), which use genomic information to assess a person's chance of having or developing a specific disease. However, on the one hand, genetic factors cannot completely determine the disease; on the other hand, genomic information is very complex and contains a large number of genetic variants. In practical applications, it is necessary to screen genes and genetic variants that have a significant impact on the occurrence and development of specific diseases from complex genomic information.

[0050] Single nucleotide polymorphism (SNP) mainly refers to DNA sequence polymorphism caused by the variation of a single nucleotide at the genome level, and is the most common type of heritable variation in humans. SNPs are widely present in the human genome, with an average of 1 in every 300 base pairs, and the total number is estimated to be 3 million or even more. Therefore, it is necessary to screen out sites that are significantly associated with specific diseases from tens of thousands of SNPs. Genome-wide association study (GWAS) analysis is an analysis of the association between specific diseases and SNPs at the whole genome level of a population, which can identify existing sequence variations at the whole genome level through biostatistical methods.

[0051] Prediction methods for gestational diabetes in related technologies usually use traditional risk factors and common biomarkers as candidate risk factors, pay less attention to genetic factors and uncommon biomarkers, and there are differences in GDM genetic loci among different populations. Due to the incompleteness of the risk factors included in GDM prediction, the diversity of data and the uncertainty of the selection method, the prediction effect is difficult to guarantee and cannot be applied to large-scale and widespread applications.

[0052] The first aspect of the embodiment of the present application proposes the use of a reagent for detecting a biomarker in the preparation of a product for predicting gestational diabetes, wherein the biomarker includes rs3781637 and / or rs35261542. Among them, the rs3781637 and rs35261542 are SNPs that have an important impact on gestational diabetes identified for the first time in the present application, and can be used as biomarkers to predict early GDM.

[0053] In some embodiments, the reagent for detecting the biomarker is selected from: a probe that specifically recognizes rs3781637 and / or rs35261542; a primer that specifically recognizes rs3781637 and / or rs35261542; or an antibody or ligand that specifically binds to rs3781637 and / or rs35261542.

[0054] In some embodiments, when the base type of rs3781637 is C, it indicates that the subject has a risk of gestational diabetes; and / or when the base type of rs35261542 is A, it indicates that the subject has a risk of gestational diabetes.

[0055] In some embodiments, the product for predicting gestational diabetes mellitus further comprises a basic clinical indicator determination reagent, wherein the basic clinical indicator determination reagent is used to detect one or more of the following values: fasting blood glucose, γ-glutamyl transferase, prealbumin, urine glucose and uric acid.

[0056] A second aspect of an embodiment of the present application proposes a product for predicting gestational diabetes, the product comprising: a reagent for detecting a biomarker, the biomarker comprising rs3781637 and / or rs35261542.

[0057] In some embodiments, the agent is selected from:

[0058] A probe specifically recognizing rs3781637 and / or rs35261542;

[0059] Primers that specifically recognize rs3781637 and / or rs35261542; or

[0060] An antibody or ligand that specifically binds to rs3781637 and / or rs35261542.

[0061] In some embodiments, the product further comprises a basic clinical index determination reagent, wherein the basic clinical index determination reagent is used to detect one or more of the following values: fasting blood glucose, γ-glutamyl transferase, prealbumin, urine glucose and uric acid.

[0062] The present invention provides a third aspect, a method for predicting gestational diabetes, the method comprising: using biomarkers and optional basic clinical indicators as predictive variables, calculating a prediction result based on the levels of the biomarkers and optional basic clinical indicators in a biological sample from a subject; and indicating whether the subject is at risk of gestational diabetes based on the prediction result, wherein the biomarkers include rs3781637 and / or rs35261542, and the basic clinical indicators include one or more of the following: age, systolic blood pressure, diastolic blood pressure, pre-pregnancy weight, whether it is the first pregnancy, fasting blood glucose, γ-glutamyl transferase, prealbumin, urine sugar and uric acid.

[0063] On the basis of traditional risk factors, the method not only incorporates 5 basic clinical indicators that are highly correlated with GDM screened out from 93 laboratory test indicators through statistical analysis, but also adds GDM-significantly associated genetic loci discovered based on the Chinese population as biomarkers for the prediction of gestational diabetes. Based on a large sample size of gestational diabetes data, incorporating reliable genetic risk factors into the disease risk prediction model helps to improve the prediction performance, with an average prediction accuracy of more than 76%, which helps hospitals screen high-risk groups for GDM in early pregnancy, and promptly intervene in their early pregnancy lifestyle to reduce the occurrence of diabetes in mid- and late pregnancy. All information for the prediction method provided in this application comes from routine physical examinations of pregnant women, which will not increase the additional medical burden on pregnant women, is economically feasible, and is fast and accurate.

[0064] In some embodiments, the prediction result is calculated based on the prediction variables using a gestational diabetes prediction model, wherein the gestational diabetes prediction model is

[0065] y=β 0 +β 1 x 1 +…+β i x i ,

[0066] Among them, β 0 =-17.993153, β 1 =0.003151, β 2 =0.022021, β 3 =0.007268, β 4 =0.104299, β 5 =0.017977, β 6 =0.913049, β 7 =0.517115, β 8 =0.005883, β 9 =0.025542, β 10 =0.005147, β11 =1.473156, β 12 =-0.199924,

[0067] Among them, x 1 is the value of body mass index, x 2 is the value of systolic blood pressure, x 3 is the value of diastolic pressure, x 4 is the value of age, x 5 Is the value of whether it is the first pregnancy, x 6 is the value of fasting blood glucose, x 7 is the value of urine sugar, x 8 is the value of uric acid, x 9 is the value of γ-glutamyl transferase, x 10 is the value of prealbumin, x 11 is the value of rs3781637, x 12 The value of rs35261542.

[0068] In some embodiments, based on the prediction result being greater than a cutoff value of -2.105, it indicates that the subject is at risk of having gestational diabetes; based on the prediction result being less than a cutoff value of -2.105, it indicates that the subject is not at risk of having gestational diabetes.

[0069] It should be noted that the gestational diabetes prediction model of the present application is a Bayesian generalized linear model, and its parameter formula is:

[0070] GDM (1 for yes, 0 for no)~BMI+SBP+DBP+Age+Number of pregnant+GLU+GLU U+UA+γGT+PA+rs3781637+rs35261542.

[0071] It is understandable that the parameters such as body mass index, systolic blood pressure, diastolic blood pressure, age, whether it is the first pregnancy, fasting blood sugar, urine sugar, uric acid, γ-glutamyl transferase, prealbumin, etc. can be measured directly. The value of rs3781637 and the value of rs35261542 need to be obtained by sequencing, genotype filling, etc. Specifically, the value can be a coded value of 0, 1, or 2. 0 indicates homozygosity with the REF allele, 1 indicates heterozygosity with the REF allele, and 2 indicates homozygosity with the ALT allele; it can also be a genotype dosage (continuous value between 0-2). Genotype dosage (Dosage) represents the expected non-reference sequence base dosage of a sample at a site, and its value is always between 0 and 2. The genotype dosage calculation formula is: Dosage = Pr (Het|data) + 2*Pr (Alt|data). Pr(Het|data) refers to the posterior probability that the genotype of a sample at this site is a heterozygous mutation, and Pr(Alt|data) refers to the posterior probability that the genotype of a sample at this site is a homozygous mutation. i The above formula can be used to obtain the predicted results of the risk of gestational diabetes mellitus.

[0072] In some embodiments, the method first classifies the candidate risk factors for gestational diabetes of the subject into qualitative basic clinical variables, quantitative basic clinical variables, laboratory clinical variables, and genetic variables. Then, based on the classification results as the qualitative basic clinical variables, a chi-square test is performed to determine the disease predictor variable; based on the classification results as the quantitative basic clinical variables, a two-sample t-test is performed to determine the disease predictor variable; based on the classification results as the laboratory clinical variables, a logistic regression analysis is performed to determine the disease predictor variable; and based on the classification results as the genetic variables, a genome-wide association analysis of single nucleotide polymorphism sites is performed to determine the disease predictor variable. Wherein, the disease predictor variable is selected from the group consisting of age, systolic blood pressure, diastolic blood pressure, pre-pregnancy weight, first pregnancy, fasting blood sugar, γ-glutamyl transferase, prealbumin, urine sugar, uric acid, SNP1 (rs3781637) genotype and SNP2 (rs35261542) genotype. Then, the data of the disease prediction variables and the disease phenotype data of the sample are combined as an initial data set; the initial data set is randomly divided into a training set and a test set; and the disease prediction model is obtained in the training set through a Bayesian generalized linear model.

[0073] AUC (Area under the Curve of ROC) refers to the area under the ROC curve, which is usually used to evaluate the training effect of the binary classification model. During verification, the closer the AUC is to 1.0, the higher the authenticity of the detection method; when AUC is equal to 0.5, the authenticity is the lowest. The AUC of the gestational diabetes prediction model used in the embodiments of the present application can reach an average of more than 76%, and up to 85%, indicating that the model has high authenticity and good predictive ability.

[0074] The fourth aspect of the embodiments of the present application proposes a device for predicting gestational diabetes, the device comprising: a calculation module for using biomarkers and optional basic clinical indicators as predictive variables, and calculating the prediction results based on the levels of the biomarkers and optional basic clinical indicators in the biological sample from the subject; and an indication module for indicating whether the subject is at risk of gestational diabetes based on the prediction results, wherein the biomarkers include rs3781637 and / or rs35261542, and the basic clinical indicators include one or more of the following: age, systolic blood pressure, diastolic blood pressure, pre-pregnancy weight, whether it is the first pregnancy, fasting blood sugar, γ-glutamyl transferase, prealbumin, urine sugar and uric acid. In some embodiments, the calculation module is also used to: use a gestational diabetes prediction model based on the predictive variables

[0075] Calculate the prediction result, wherein the gestational diabetes prediction model is y=β 0 +β 1 x 1 +…+β i x i ,

[0076] Among them, β 0 =-17.993153, β 1 =0.003151, β 2 =0.022021, β 3 =0.007268, β 4 =0.104299, β 5 =0.017977, β 6 =0.913049, β 7 =0.517115, β 8 =0.005883, β 9 =0.025542, β 10 =0.005147, β 11 =1.473156, β 12 =-0.199924,

[0077] Among them, x 1 is the value of body mass index, x2 is the value of systolic blood pressure, x 3 is the value of diastolic pressure, x 4 is the value of age, x 5 Is the value of whether it is the first pregnancy, x 6 is the value of fasting blood glucose, x 7 is the value of urine sugar, x 8 is the value of uric acid, x 9 is the value of γ-glutamyl transferase, x 10 is the value of prealbumin, x 11 is the value of rs3781637, x 12 The value of rs35261542.

[0078] In some embodiments, the indication module is specifically used to: indicate that the subject is at risk of gestational diabetes based on the prediction result being greater than a cutoff value of -2.105; and indicate that the subject is not at risk of gestational diabetes based on the prediction result being less than a cutoff value of -2.105.

[0079] The fifth aspect of the embodiments of the present application proposes a system for predicting gestational diabetes, the system comprising: a processor; an input module for inputting the levels of biomarkers and optional basic clinical indicators in a biological sample from a subject, wherein the biomarkers include rs3781637 and / or rs35261542, and the basic clinical indicators include one or more of the following: age, systolic blood pressure, diastolic blood pressure, pre-pregnancy weight, whether it is the first pregnancy, fasting blood glucose, γ-glutamyl transferase, prealbumin, urine glucose and uric acid; a computer-readable medium containing instructions, which, when executed by the processor, implement the method for predicting gestational diabetes in a subject as described in the embodiment of the third aspect; and an output module for indicating whether the subject is at risk of developing gestational diabetes.

[0080] The sixth aspect of the embodiments of the present application proposes a computer-readable storage medium, characterized in that computer program instructions are stored on the computer-readable storage medium, and when the computer program instructions are executed by a processor, the method for predicting gestational diabetes as described in the third aspect of the embodiments is implemented.

[0081] A seventh aspect of the embodiments of the present application proposes a computer program product, including a computer program, which, when executed by a processor, implements the method for predicting gestational diabetes as described in the third aspect of the embodiments.

[0082] An eighth aspect of the embodiments of the present application proposes a computer program, which includes computer program code. When the computer program code is run on a computer, the computer executes the method for predicting gestational diabetes as described in the third aspect of the embodiments.

[0083] Example 1

[0084] This example collects hospital electronic medical record information and prenatal genomic data of 21,813 pregnant women who were archived in Wuhan Children's Hospital from 2015 to 2020, effectively combines the clinical information and genetic information of the subjects, and constructs an economically feasible, fast and accurate disease prediction model.

[0085] 1.1 Identification of risk factors

[0086] (1) Data on basic clinical factors, biomarkers, and genetic factors were collected from a total of 21,813 pregnant women who were registered in Wuhan Children's Hospital from 2015 to 2020. The first 15 gestational weeks were defined as the early pregnancy. The specific information on candidate disease risk factors was collected as shown in Table 1.

[0087] Table 1 Candidate disease risk factors

[0088]

[0089] (2) The GDM diagnosis results of the above pregnant women were collected. The blood glucose levels in the oral glucose tolerance test (OGTT) at 24 to 28 weeks of pregnancy met any one or more of the following criteria: fasting blood glucose ≥5.1mmol / L, 1h blood glucose ≥10.0mmol / L, and 2h blood glucose ≥8.5mmol / L as the criteria for determining GDM. Among the 21,813 samples, 2,502 were diagnosed as GDM cases.

[0090] (3) Classifying the candidate disease risk factors into qualitative basic clinical variables, quantitative basic clinical variables, laboratory clinical variables, and genetic variables.

[0091] (4) Analyze the corresponding data set of the candidate disease risk factors according to the classification results, and determine the significant basic clinical variables as disease prediction variables, wherein the chi-square test is performed to determine the disease prediction variables based on the qualitative basic clinical variables based on the classification results, and the two-sample t-test is performed to determine the disease prediction variables based on the quantitative basic clinical variables based on the classification results, and the threshold value is 1E-20.

[0092] (5) Based on the classification results as the laboratory clinical variables, that is, the biomarkers detected by the laboratory, a single marker is used as a predictor variable, age is used as a covariate, and GDM is used as a response variable. Logistic regression analysis is performed to determine the significant laboratory clinical variables as the disease prediction variables, wherein the average value is taken for multiple continuous variables, and for multiple binary variables, once a positive result occurs, it is defined as positive. The logistic regression threshold of the biochemical index is 1E-20.

[0093] (6) Based on the classification result, the genetic variables, i.e., the genotype data of the whole genome, are preprocessed on the genomic data. The preprocessing includes alignment and filtering. First, the sample genomic data is aligned to the hg38 human genome reference genome using BWA software, and then samples with a sequencing depth lower than 0.05X and an alignment rate lower than 90% are deleted.

[0094] (7) samtools software was used to remove PCR duplications in the aligned gene data, and GATK software was used to complete the realignment and base quality correction of the indel region.

[0095] (8) STITCH imputation software was used to fill in the genotype data after correction using CHS and CHN provided by the 1000 Genomes Project as reference genomes. Basevar software was then used to perform variation detection to obtain single nucleotide polymorphism sites.

[0096] (9) Genome-wide association analysis of single nucleotide polymorphism sites was performed using PLINK software. The number of samples that underwent OGTT testing and had non-invasive prenatal testing genomic data that could be used for analysis was 14,744, and the significance threshold was 5e-08.

[0097] (10) The significant single nucleotide polymorphism sites in the whole genome association analysis were used to perform conditional and joint association analysis in GCTA software to exclude linkage disequilibrium and ultimately determine the significant and independent SNPs as the disease predictive variables.

[0098] This example ultimately identifies reliable risk factors that can be used to build a disease prediction model.

[0099] The results of the candidate risk factor analysis classified as basic clinical variables are shown in Table 2. The results showed that female age, systolic blood pressure, diastolic blood pressure, pre-pregnancy weight, body mass index, overweight, and whether it was the first pregnancy were considered to be basic clinical risk factors significantly associated with GDM. Since pre-pregnancy weight, body mass index, and overweight risk factors are highly correlated, in order to avoid collinearity in the model and pre-pregnancy weight is the most easily obtained risk factor, female age, systolic blood pressure, diastolic blood pressure, pre-pregnancy weight, and whether it was the first pregnancy were included in the disease predictor variables.

[0100] Table 2 Correlation analysis between basic clinical information and GDM

[0101]

[0102] The results of the analysis of candidate risk factors for biomarkers classified as laboratory clinical variables are as follows Figure 4The results showed that the five basic clinical indicators most significantly correlated with GDM were fasting blood glucose, γ-glutamyl transferase, prealbumin, urine glucose, and uric acid. The above were included in the disease prediction variables.

[0103] The results of genome-wide association analysis of single nucleotide polymorphism sites classified as genetic variables are shown in Figure 5-8 shown. Figure 5 QQ plot of genome-wide association analysis, the observed values ​​are in -log 10 3, which is consistent with the expected uniform distribution. 10 3 and then significantly deviate from the expected value, thus indicating the presence of phenotypic variation caused by mutation. The population data and gene data of this embodiment can be used for SNP association analysis.

[0104] Figure 6 This is a Manhattan plot of the genome-wide association analysis, in which the CDKAL1 gene on chromosome 6 and the MTNR1B gene on chromosome 11 are significantly associated with GDM. Figure 7 and Figure 8 The results of the association analysis were used to obtain the P value and recombination rate of each SNP in the region. Among them, the significant SNP in the CDKAL1 gene was rs35261542, and the significant SNP in the MTNR1B gene was rs3781637.

[0105] The results of conditional and joint association analysis are shown in Table 3 Independent SNPs rs35261542 (chr6:20675561) and rs3781637 (chr11:92980604) were obtained as disease predictive variables.

[0106] Table 3 Genetic variables included in disease predictors

[0107]

[0108] 1.2 Construction and verification of disease prediction model

[0109] (1) The significant variables obtained in 1.1, namely, female age, systolic blood pressure, diastolic blood pressure, pre-pregnancy weight, first pregnancy, fasting blood glucose, γ-glutamyl transferase, prealbumin, urine glucose, uric acid, SNP1 rs35261542 and SNP2 rs3781637, were used as disease predictive variables.

[0110] (2) Constructing a matrix by merging the data of the disease prediction variables and the disease phenotype data of the sample (GDM diagnosis results).

[0111] (3) The na.omit function in R software was used to remove all sample information with missing values ​​in the matrix and obtain an initial data set derived from 2677 samples.

[0112] (4) The sample function in R software was used to randomly divide the initial data set into a training set (90%) and a test set (10%), wherein the training set included 2409 samples and the test set included 268 samples.

[0113] (5) The Bayesian generalized linear model (bayesglm) method (method = "bayesglm") in the train function of the caret package in R software was used to estimate the model coefficients, and the initial disease prediction model was obtained by training the model through ten-fold cross validation. The parameter formula was input as GDM (yes 1, no 0) ~ BMI + SBP + DBP + Age + Number of pregnant + GLU + GLU U + UA + γGT + PA + rs3781637 + rs35261542. The caret package finalModel selected the fitting model with the best parameters and extracted the model coefficient matrix coefficients column as the coefficient matrix B.

[0114] (6) The values ​​corresponding to the disease prediction variable information in the training set are taken as matrix A, matrix A and matrix B are input into the initial disease prediction model, and the gestational diabetes prediction value y of each sample is calculated to obtain matrix C. The calculation formula is as follows:

[0115] y=β 0 +β 1 x 1 +…+β i x i

[0116] Among them, the matrix B of coefficient βi is shown in Table 4, β 0 =-17.993153, β 1 =0.003151, β 2 =0.022021, β 3 =0.007268, β 4 =0.104299, β 5 =0.017977, β 6 =0.913049, β 7 =0.517115, β 8 =0.005883, β 9 =0.025542, β 10 =0.005147, β 11 =1.473156, β 12=-0.199924;

[0117] Correspondingly, the prediction factor Xi matrix A is shown in Table 5, x 1 is the value of body mass index, x 2 is the value of systolic blood pressure, x 3 is the value of diastolic pressure, x 4 is the value of age, x 5 Is the value of whether it is the first pregnancy, x 6 is the value of fasting blood glucose, x 7 is the value of urine sugar, x 8 is the value of uric acid, x 9 is the value of γ-glutamyl transferase, x 10 is the value of prealbumin, x 11 is the value of rs3781637, x 12 The value of rs35261542.

[0118] Table 4 Matrix B

[0119] intercept -17.993153 Body Mass Index 0.003151 Systolic blood pressure 0.022021 Diastolic blood pressure 0.007268 age 0.104299 Is it the first pregnancy? 0.017977 Fasting blood sugar 0.913049 Urine sugar 0.517115 Uric acid 0.005883 γ-glutamyl transferase 0.025542 Prealbumin 0.005147 rs3781637 1.473156 rs35261542 -0.199924

[0120] Table 5 Matrix A

[0121]

[0122] (7) The disease prediction results obtained in (6) (Table 6) were compared with the true GDM information in the training set using the auc function in the pROC package.

[0123] Table 6 Matrix C

[0124]

[0125] (8) Repeat the above steps (4)-(7) 100 times and calculate the average AUC and ROC curve of the fitted model to evaluate the model quality.

[0126] The average AUC results are as follows Fig. 9 As shown, the ROC curve of the fitting model of the best parameters for the 100th repetition in this embodiment is as follows Fig.10 As shown in the figure, the AUC is about 0.7652 and the cutoff value is -2.105. The continuous prediction value provided by the disease prediction model can be used to indicate the risk of disease. That is, when the prediction value is greater than -2.105, the subject is at risk of gestational diabetes; when the prediction value is less than -2.105, the subject is not at risk of gestational diabetes.

[0127] The disease prediction model obtained in this embodiment can be used to provide early prediction results of GDM. The prediction accuracy of GDM can reach 85%, and the average accuracy of 100 repetitions exceeds 76% (median = 0.7715, Q3 = 0.7972). The prediction model provided by this application has high authenticity, good prediction ability, economic feasibility, rapidity and accuracy.

[0128] In the present invention, the terms "one embodiment", "some embodiments", "examples", "specific examples", or "some examples" etc. mean that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the described specific features, structures, materials or characteristics may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0129] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations of the present invention. A person skilled in the art may change, modify, replace and vary the above embodiments within the scope of the present invention.

Claims

1. Use of a reagent for detecting a biomarker in the preparation of a product for predicting gestational diabetes, characterized in that: The biomarkers include rs3781637 and / or rs35261542.

2. The use according to claim 1, characterized in that: The reagent for detecting biomarkers is selected from: Probes that specifically recognize rs3781637 and / or rs35261542; Primers that specifically recognize rs3781637 and / or rs35261542; or An antibody or ligand that specifically binds to rs3781637 and / or rs35261542.

3. The use according to claim 1, characterized in that: When the base type of rs3781637 is C, it indicates that the subject has a risk of having gestational diabetes; and / or when the base type of rs35261542 is A, it indicates that the subject has a risk of having gestational diabetes.

4. The use according to any one of claims 1 to 3, characterized in that The product for predicting gestational diabetes also includes a basic clinical indicator determination reagent, which is used to detect one or more of the following values: fasting blood glucose, γ-glutamyl transferase, prealbumin, urine sugar and uric acid.

5. A product for predicting gestational diabetes, characterized in that: The product comprises: a reagent for detecting a biomarker, wherein the biomarker comprises rs3781637 and / or rs35261542, Optionally, the reagent is selected from: A probe specifically recognizing rs3781637 and / or rs35261542; Primers that specifically recognize rs3781637 and / or rs35261542; or An antibody or ligand that specifically binds to rs3781637 and / or rs35261542, Optionally, the product further comprises a basic clinical index determination reagent, which is used to detect one or more of the following values: fasting blood glucose, γ-glutamyl transferase, prealbumin, urine glucose and uric acid.

6. A method for predicting gestational diabetes, characterized in that: The method comprises: Using the biomarkers and optional basic clinical indicators as predictive variables, and calculating a prediction result based on the levels of the biomarkers and optional basic clinical indicators in a biological sample from a subject; and According to the prediction result, indicating whether the subject has a risk of gestational diabetes, The biomarkers include rs3781637 and / or rs35261542, and the basic clinical indicators include one or more of the following: age, systolic blood pressure, diastolic blood pressure, pre-pregnancy weight, whether it is the first pregnancy, fasting blood glucose, γ-glutamyl transferase, prealbumin, urine sugar and uric acid.

7. The method according to claim 6, characterized in that Based on the predicted variables, the predicted result is calculated using a gestational diabetes prediction model, wherein the gestational diabetes prediction model is a Bayesian generalized linear model, and the model formula is: y=β0+β1x1+…+β i x i , Where, β0=-17.993153, β1=0.003151, β2=0.022021, β3=0.007268, β4=0.104299, β5=0.017977, β6=0.913049, β7=0.517115, β8=0.005883, β9=0.025542, β 10 =0.005147, β 11 =1.473156, β 12 =-0.199924, Among them, x1 is the value of body mass index, x2 is the value of systolic blood pressure, x3 is the value of diastolic blood pressure, x4 is the value of age, x5 is the value of whether it is the first pregnancy, x6 is the value of fasting blood sugar, x7 is the value of urine sugar, x8 is the value of uric acid, x9 is the value of γ-glutamyl transferase, x 10 is the value of prealbumin, x 11 is the value of rs3781637, x 12 For the value of rs35261542, optionally, based on the prediction result being greater than the cutoff value -2.105, it indicates that the subject is at risk of gestational diabetes; based on the prediction result being less than the cutoff value -2.105, it indicates that the subject is not at risk of gestational diabetes.

8. A device for predicting gestational diabetes, characterized in that: The device comprises: A calculation module, for calculating a prediction result based on the levels of the biomarkers and optional basic clinical indicators in a biological sample derived from a subject, using the biomarkers and optional basic clinical indicators as predictive variables; and an indication module, for indicating whether the subject has a risk of gestational diabetes according to the prediction result, The biomarkers include rs3781637 and / or rs35261542, and the basic clinical indicators include one or more of the following: age, systolic blood pressure, diastolic blood pressure, pre-pregnancy weight, whether it is the first pregnancy, fasting blood glucose, γ-glutamyl transferase, prealbumin, urine sugar and uric acid.

9. The device according to claim 8, characterized in that The calculation module is further used to calculate the prediction result based on the prediction variable using a gestational diabetes prediction model, wherein the gestational diabetes prediction model is a Bayesian generalized linear model, and the model formula is: y=β0+β1x1+…+β i x i , Where, β0=-17.993153, β1=0.003151, β2=0.022021, β3=0.007268, β4=0.104299, β5=0.017977, β6=0.913049, β7=0.517115, β8=0.005883, β9=0.025542, β 10 =0.005147, β 11 =1.473156, β 12 =-0.199924, Among them, x1 is the value of body mass index, x2 is the value of systolic blood pressure, x3 is the value of diastolic blood pressure, x4 is the value of age, x5 is the value of whether it is the first pregnancy, x6 is the value of fasting blood sugar, x7 is the value of urine sugar, x8 is the value of uric acid, x9 is the value of γ-glutamyl transferase, x 10 is the value of prealbumin, x 11 is the value of rs3781637, x 12 is the value of rs35261542, Optionally, the indication module is specifically used to: indicate that the subject is at risk of gestational diabetes based on the prediction result being greater than a cutoff value of -2.105; and indicate that the subject is not at risk of gestational diabetes based on the prediction result being less than a cutoff value of -2.

105.

10. A gestational diabetes prediction system, characterized in that: The system comprises: processor; An input module for inputting the levels of biomarkers and optional basic clinical indicators in a biological sample from a subject, wherein the biomarkers include rs3781637 and / or rs35261542, and the basic clinical indicators include one or more of the following: age, systolic blood pressure, diastolic blood pressure, pre-pregnancy weight, whether it is the first pregnancy, fasting blood sugar, γ-glutamyl transferase, prealbumin, urine sugar and uric acid; A computer-readable medium comprising instructions which, when executed by the processor, implement the method of predicting gestational diabetes in a subject as claimed in claim 6 or 7; and An output module is used to indicate whether the subject has a risk of gestational diabetes.

Images