| BackgroundAlong with the emergence and development of Precision Medicine, the model of personalized medicine is established, the gene sequencing technology makes rapid development and the clinical big data and biological network are set up. Technology applications realized by interdisciplinary collaboration are improved in disease prediction, diagnosis and etiology analysis unceasingly. These entire make the personalized precision treatment of diseases and patients possible, and they are of great significance in improving the prevention of disease.Congenital heart disease is the most common congenital malformation, but also is the leading cause of infant death. The incidence rate in newborns reaches about 0.7% to 1%, seriously endangering the health of infants and young children. A variety of factors interacting in time and space are involved in the process of heart development, while genetic and environmental factors in the embryonic phase can lead to cardiac dysplasia. Because of the complexity of the genetic mechanisms of congenital heart disease, the causes leading to cardiac malformations remain unclear. The type of congenital heart disease is complex and diverse which has been a great problem for the treatment and prevention of congenital heart disease.Clinical studies have shown that serological indicators may reflect cardiac function. The clinical detection indicators are mainly high-sensitivity C-reactive protein, lipoprotein a, cardiac troponin, natriuretic peptide. The abnormal heart can cause the abnormal changes of serological markers levels. To establish the early diagnostic model through serum markers associated with congenital heart disease can provide a theoretical basis for early diagnosis and treatment of congenital heart disease.More and more studies suggest that the pathogenesis of congenital heart disease is related to certain transcription factors, whereas the relationship between congenital heart disease susceptibility genes and serological markers has yet been reported. With the rapid development of bioinformatics, Gene Ontology(GO) has become an important tool and method in the field of bioinformatics. GO database is a structured standard biological model, covering the genes cellular component, molecular function and biological process. GO plays a huge role in terms of gene function annotation. It can analyze gene or protein locations in the cell, molecular functions and the involved biological processes, thus simplifying the annotation of genes and their products as a standardized vocabulary.This study tried to adopt Meta analysis to screen the susceptibility genes and risk serum factors of congenital heart disease from domestic and foreign researches in recent years. Based on clinical serum marker levels, the early relational models of congenital heart disease markers were established by the use of logistic regression and support vector machine. GO analysis was used for functional annotation of the susceptibility genes and serological markers of congenital heart disease and the functional relationship between susceptibility genes and serum markers was established. The study is divided into following three parts.In this paper, clinical medicine, meta-analysis and GO analysis in bioinformatics and other multidisciplinary research methods were adopted in order to offer theoretical foundation in seeking the correlation between gene expressions and serum markers which were related to the causes of diseases. The relational model of serum markers established based on indicators such as serum markers may also provide technical support for the early prediction of other diseases.Part I Meta-analysis of susceptibility genes and serum risk factors in clinical research of congenital heart diseaseObjective This chapter aimed to screen the susceptibility genes and serum risk factors of congenital heart disease.Methods CNKI, VIP, WANFANG DATA and Pub Med databases are the main sources of literature. The Chinese search terms for susceptibility genes are “congenital heart disease”, “gene” and “mutation”, and the English search terms are “congenital heart disease”, “gene” and “mutations”. The publication period is from January 1, 2000 to June 31, 2014. The Chinese search terms for serum risk factors are “congenital heart disease”, “serum marker” and “early diagnosis”, and the English keywords are “congenital heart disease”, “serum markers” and “diagnosis”. The publication period is from January 1, 2006 to October 31, 2014. The full-length literatures included are Chinese or English. The effective data were extracted for Meta analysis using Revman5.1 software.Results 1. Nineteen literatures of congenital heart disease susceptibility genes were included in Meta-analysis, seven of which were related to NKX2.5, eleven for GATA4 gene and four for FOG2 gene. The pooled OR value for NKX2.5 gene was 2.02 and 95%CI was 1.42-2.86; GATA4 gene: pooled OR=2.05, 95%CI=1.50-2.88; FOG2 gene pooled OR=19.43, 95% CI = 4.52-83.63.2. Twenty literatures of congenital heart disease serum risk factors were included in Meta-analysis. The combined SMD for cardiac troponin was 0.33 and 95%CI was 0.12-0.55. The combined effect size SMD for high-sensitivity C-reactive protein was 1.84 and 95%CI was 1.36-2.32. The combined effect size SMD for BNP was 321.33 and 95% CI was 279.03-364.29.Conclusion As genetic factors, NKX2.5, GATA4 and FOG2 were the susceptibility genes for congenital heart disease, while serum c Tn I, hs CRP and BNP were risk factors for congenital heart disease.Part II Establishment of the relational model of the clinical marker group of congenital heart diseaseObjective This chapter tried to establish the early relational model of congenital heart disease and evaluate its practical application value.Methods 1.Eighty patients with congenital heart disease in the Department of Cardiac Surgery of our hospital(54 males and 76 females, who ranged in age from 7 days to 59 years old) from December 2009 to September 2014 and eighty outpatients with normal physical examination(38 males and 42 females aged from 0.3 to 51 years) were enrolled as the research subjects. Serum BNP level was detected using enzyme-linked immunosorbent assay(ELISA); immunoturbidimetric assays was utilized to detect serum hs-CRP; the immune luminescence method was used to detect serum c Tn I level; serum level of Lp(a) was examined using enzyme-linked immunosorbent assay(ELISA) double antibody sandwich method. The test methods were carried out in strict accordance with the kit instructions. Each sample was performed parallel testing for twice and the average value of them was considered as the final test results. Based on the clinical trial results, this study compared and complemented the serological results screened by Meta.2. The diagnostic result of congenital heart disease(patient = 1, healthy person = 0) was regarded as dependent variable. The serological markers in deciding congenital heart disease were then screened using forward method for stepwise logistic regression analysis. The modeling indicators were finally obtained through stepwise operations. The relational model of congenital heart disease serum markers was established based on support vector machine(SVM). One hundred and twenty cases(60 cases from patients with congenital heart disease and 60 cases from healthy controls) were selected as the test set sample and then input into the support vector machine(SVM) for training.3. The level of variable of the logistic regression model was considered as the test variable and the result of pathological diagnosis was regarded as state variable to draw the ROC curve. According to the area under the curve of(AUC) of ROC, its application value in congenital heart disease was evaluated. The remaining 20 patients with congenital heart disease and 20 controls were selected as the test set sample, then they were input into the trained support vector machine network, so the corresponding discriminated result(1 or 0) was obtained. Compared with the target, the discriminated accuracy can be obtained and the application value of this model can thus be estimated.Results The levels of c Tn I, hs-CRP, BNP and Lp(a) in patients with congenital heart disease were dramatically higher than healthy subjects and the difference was statistically significant.2. The logistic regression analysis showed that single variables Lp(a), c Tn I and BNP were associated with the occurrence of congenital heart disease. The multivariates c Tn I, BNP and Lp(a) were not significantly correlated with the incidence of congenital heart disease, the joint variables: Lp(a) and c Tn I, Lp(a) and BNP as well as BNP and c Tn I were also related to the occurrence of congenital heart disease, and their joint detection area under the curve(AUC) of ROC were 0.994, 0.981 and 0.999, their diagnostic accuracy was 93.4%、87.1%、97.2%, indicating a high value of application. The diagnostic accuracy of support vector machine diagnosis model was 85%.Conclusion Compared with the normal subjects, there existed significant abnormities in the serum levels of c Tn I, hs-CRP, BNP and Lp(a) of patients with congenital heart disease. The Lp(a) and c Tn I, Lp(a) and BNP as well as BNP and c Tn I joint detection model based on logistic regression and the relational model of congenital heart disease markers based on support vector machine had high auxiliary diagnosis value, The diagnostic accuracy of logistic regression diagnosis model was higher, which thus can provide technical support for the auxiliary diagnosis of congenital heart disease.Part III Bioinformatics functional analysis of serum markers Lp(a), BNP and congenital heart disease susceptibility genesObjective This chapter was planned to investigate the functional relationships between congenital heart disease susceptibility genes and relevant serum markers.Methods Gene Ontology(GO) analysis was performed for functional annotation on susceptibility genes and serum markers of congenital heart disease using bioinformatics methods to look for the same functions between susceptibility genes and serum markers and build the functional relational graph between serums and genes.2. Real-time fluorescence quantification PCR was adopted to detect the susceptibility gene expression levels in the congenital heart disease patients and the healthy controls. When compared with the serum marker levels, the possible relationship between susceptibility genes and serum markers was analyzed.Results 1.GO analysis suggested that there were functional associations between congenital heart disease susceptibility gene NKX2.5, GATA4 and FOG2 and serum markers Lp(a) and BNP. And there were no associations found temporarilybetween NKX2.5, GATA4 and FOG2 and serum markers c Tn I and hs-CRP.2. Compared to the healthy subjects, the NKX2.5, GATA4 and FOG2 gene relative expression levels 2-ΔΔCt of patients with congenital heart disease were 0.59 ± 0.18, 0.47 ± 0.14 and 0.33 ± 0.99, relatively. The expression levels of the patients group was significantly lower than those of the control group, while the serum Lp(a) and BNP levels of the patients group were significantly higher than those of the control group.Conclusion There existed a link between congenital heart disease and susceptibility genes NKX2.5, GATA4 and FOG2 and serum markers Lp(a) and BNP in GO functional annotation, in which the relationship between Lp(a) was mainly in aspects of Lipoprotein transport across the membrane and blood circulation and that between BNP was mainly gene expression and metabolic process. The abnormal expression of NKX2.5, GATA4 and FOG2 may be one of the reasons for the exceptional levels of Lp(a) and BNP. |
PDF Full Text Request