| Gene expression microarray techonology is widely used in post genome studies. Experimental techniques of gene expression microarray were well developed,but the statistical methods and bioinformatics analysis on gene expression microarray data has lagged behind.In this study,two statistical methods(ANOVA and bayesian methods) were used to find differentially expressed genes in porcine placenta of mid-later gestation, and some bioinformatics methods such as gene annotation,clustering and the maping of pathway were also used to understand the gene expression patterns.The main results are as follows:1.By variance analysis,a total of 90 and 319 differentially expressed transcripts were detected between E75 and L75 and between E90 and L90(P＜0.05;FDR＜0.2; FC＞2 or FC＜0.5).Among the 90 genes,35 genes were up-regulated at E75,55 genes were down-regulated at E75.Among the 319 genes,135 genes were up-regulated at E90, 184 genes were down-regulated at E90.2.By Bayesian analysis,a total of 271 and 538 differentially expressed genes were detected between E75 and L75 and between E90 and L90(p＜0.05;FC＞2 or FC＜0.5). Among the 271 genes,108 genes were up-regulated at E75,163 genes were down-regulated at E75.Among the 538 genes,256 genes were up-regulated at E90,282 genes were down-regulated at E90.3.Results of the two methods(variance methods and bayesian methods) were compared.We found 61 common genes were identified as differential expression between E75 and L75 by both methods.247 common genes were identified as differential expression between E90 and L90 by both methods.4.Functional annotations were pursued for differentially expressed genes selected. The TC accession numbers were first updated from TIGR 5.0 to TIGR 11.0 and the corresponding Human Gene IDs were pulled out so that the DAVID analysis software could be interrogated.The data were then analyzed using DAVID 2.0 and 2.1 beta.Gene Ontology(GO) biological process classification of 472 differentially expressed genes between E75 and L75 indicated that encoding proteins of these genes were associated with cellular physiological process(58.6%),cell communication(22.5%),signal transduction(20%),development(16.7%) and organismal physiological process(16.7%). In E90 and L90,biological process classification of 657 differentially expressed genes indicated that encoding proteins of these genes were associated with cell communication (21.6%),signal transduction(19.2%),development(16.1%),organismal physiological process(14.1%),and biopolymer modification(10.2%).5.Hierarchical cluster analysis of differentially expressed genes was conducted using the Gene Cluster 3.0 and treeview 1.6 software(Stanford University,2002).To gain insight into transcriptome-scale similarities among all four placenta types,all the differentially expressed genes were used to do systematic cluster analysis.The result showed that L75 and L90 were initially clustered together because their expression profiles were most similar.E75 and E90 were clustered to form another class.The result also showed the genes were classified to 6 functional parts.6.Real-time quantitative PCR was used to verify the differential expression of 9 genes detected by the Affymetrix GeneChip.We selected nine genes(ALDH1A1,DIO3, DIRAS3,PLAGL1,PON2,DCN,ASCL2,WIF1,SLC38A4) to confirm microarray data by Q-PCR.Expression patterns of 8 genes were in consistent with the microarray data.In these genes,SLC38A4 showed no significant change on the microarray by ANOVA method.It also was confirmed by Q-PCR.One gene,DCN,showing differentially expressed on the microarray,did not show significant differential expression in Q-PCR result.The results of the study showed that the variance analysis of differentially expressed genes have a certain credibility.The new Bayesian method for the study of the chip provides a new way of analyzing microarray data.This work laid a good foundation for functional research of genes in further study.
|
PDF Full Text Request