Study On Factors Of Intrauterine Infection Of Hepatitis B Virus Based On Cohort Study And Bioinformatics Mining

Objective:1.Through the pre-cohort study to understand the prevalence of intrauterine infection of hepatitis B virus?HBV?in pregnant women with positive HBsAg in Wuhan,and to identify the main influencing factors of intrauterine infection of HBV.2.We had constructed risk prediction models for HBV intrauterine infectio and evaluated the generalization ability of different models in the diagnosis of HBV intrauterine infection.3.We used the Gene Expression Omnibus?GEO?datasets to conduct the expression profiling array related to intrauterine infection of HBV and to identify genes related to intrauterine infection of HBV.Methods:1.In the first part,HBsAg positive pregnant women and their neonates were consecutively enrolled in this study from May 2012 to December 2016 at Wuhan Women and Children Medical and Healthcare Center,Wuhan,China.Epidemiological survey was carried out to investigate the influencing factors of HBV intrauterine infection.Neonatal femoral vein HBsAg or HBV DNA was detected as positive at birth as diagnostic criteria to intrauterine infection of HBV.To analyze the relationship between HBsAg positive maternal age,pregnancy,HBV serological exposure and intrauterine infection of HBV.2.The second part took the neonatal HBV intrauterine infection as the outcome,we used the non-conditional Logistic regression model,the decision tree model and the BP neural network model to construct the risk prediction model of the neonatal HBV intrauterine infection,compared the three models of the area under the ROC,and evaluates generalization ability of the three models under different prediction samples.3.The susceptibility genes of HBV intrauterine infection were screened by previous studies,and the candidate genes were selected by using the HBV expression spectrum chip GSE51489 and GSE83148 in the GEO database to change the 1.2 multiple?fold change,FC??logFC>1.2?and P<0.05.GO?Gene Ontology? bioconcentration analysis and KEGG?Kyoto Encyclopedia of Genes and Genomes?pathway were analyzed for the 12 candidate genes of GSE51489 and GSE83418.We had verified candiate genes by GSE65389.Result:1.HBV intrauterine infection rate was 7.06%?109/1544?.Risk of HBV infection in newborn babies born with HBeAg positive pregnant women is higher than that of HBeAg negative pregnant women?RR=3.98,95%CI:2.66-5.94?.The risk of HBV intrauterine infection in newborn infants with HBV DNA positive pregnant women is higher than HBVDNA negative?RR=2.96,95%CI:2.00-4.39?.The rate of intrauterine infection in newborn infants increased with HBV DNA load of pregnancy(?²_CMH=50.380,P<0.001).2.?1?Maternal HBeAg positive during pregnancy?OR=3.14,95%CI:1.96-5.01?,HBV DNA load in late pregnancy?OR=1.88,95%CI:1.19-2.98?and cesarean section?OR=0.41,95%CI:0.27-0.61?were risk factors of intrauterine infection of HBV.The area under ROC?AUC?of Logistic regression was 0.731?95%CI:0.66-0.80?.?2?The decision tree analysis showed that HBeA,HBV DNA and delivery mode entered the model.The factor that first entered the decision tree model was HBeAg??²=51.178,Adjusted P<0.001?during pregnancy and formed 2 subsets:negative and positive.The first positive group was centralized,the delivery mode entered the classification tree??²=20.626,P<0.001?,and there were two sub subsets:cesarean section and vaginal delivery.In the sub subgroup of cesarean section,HBVDNA entered the classification tree??²=7.621,P<0.01?at the third trimester,and AUC of decision tree was 0.667?95%CI:0.59-0.74?.?3?The results of BP neural network model showed that there were one hidden layers,three units in the hidden layer,and the top five of the predictive variables affecting the risk of intrauterine infection of HBV were HBeAg in pregnancy,HBV DNA in the late pregnancy,pregnancy,menstruation and maternal birth age,and the AUC of the model was 0.668?95%CI:0.58-0.72?.?4?We had compared AUC of the three models,there was no significant difference between them?P>0.05?.The generalization ability of the three models was evaluated using 30%,20%and 10%predictive sample size respectively.There was no significant difference between AUC of each model under different prediction samples?P>0.05?.The result indicated that the prediction model was more stable and could be used for the prediction of HBV intrauterine infection.3.GSE51489 and GSE83148 co-screened twelve differentially expressed genes:RGS1,PDE4B,MAFF,PMAIP1,JUND,PHLDA1,FOS,LAIR2,RGCC,CXCL8,SIK1 and CCL20.GO analysis and KEGG pathway analysis revealed that candidate genes were mainly concentrated in diseases and pathways related to immune function.PPI network analysis found that the relationship between candidate genes PPI was very close,and FOS,JUN and ATF3 were located at the center of the network.In the GSE65389,the genes screened in the PPI network protein and the early molecular epidemiology study were verified.It was found that TNFAIP3,CDKN1A,CD44,CXCL8,CCL20,CXCL13,PDCD1,LTA,MARS,TLR3,TLR4and SLC10A1 in the GSE65389.The expression of TNFAIP3,PDCD1,LTA and TLR3 was significant between difference groups?P<0.05?.Conclusions:1.The incidence of intrauterine infection of HBV in HBsAg positive pregnant women in Wuhan was 7.06%.HBeAg positive during pregnancy,HBV DNA positive in late pregnancy and vaginal delivery are risk factors for intrauterine infection of HBV.2.Three prediction models suggest that maternal HBeAg,HBV DNA positive in late pregnancy and delivery mode play a very important role in predicting the risk of intrauterine infection of HBV.The three models can be used to predict intrauterine infection of HBV.3.We had screened differentially expressed genes through bioinformatics mining and found that these genes were mainly enriched in diseases and pathways related to immune function.According to the genes related to intrauterine infection of HBV in the previous cohort,the expression difference between TNFAIP3,PDCD1,LTA and TLR3 in different groups was demonstrated in the target expression spectrum chip,which provided a new gene and research direction for the study of HBV infection and even the intrauterine infection of HBV.