diff --git a/9-microarray-examples/GSE70529.R b/9-microarray-examples/GSE70529.R
new file mode 100644
index 0000000..a85dc02
--- /dev/null
+++ b/9-microarray-examples/GSE70529.R
@@ -0,0 +1,50 @@
+library(GEOquery)
+library(limma) 
+GSE70529 <- getGEO('GSE70529', destdir=".",getGPL = F)
+exprSet=exprs(GSE70529[[1]])
+library("annotate")
+GSE70529[[1]]
+
+platformDB='hugene10sttranscriptcluster.db'
+library(platformDB, character.only=TRUE)
+probeset <- featureNames(GSE70529[[1]])
+#EGID <- as.numeric(lookUp(probeset, platformDB, "ENTREZID"))
+SYMBOL <-  lookUp(probeset, platformDB, "SYMBOL")
+a=cbind(SYMBOL,exprSet)
+## remove the duplicated probeset 
+rmDupID <-function(a=matrix(c(1,1:5,2,2:6,2,3:7),ncol=6)){
+  exprSet=a[,-1]
+  rowMeans=apply(exprSet,1,function(x) mean(as.numeric(x),na.rm=T))
+  a=a[order(rowMeans,decreasing=T),]
+  exprSet=a[!duplicated(a[,1]),]
+  #
+  exprSet=exprSet[!is.na(exprSet[,1]),]
+  rownames(exprSet)=exprSet[,1]
+  exprSet=exprSet[,-1]
+  return(exprSet)
+}
+exprSet=rmDupID(a)
+rn=rownames(exprSet)
+exprSet=apply(exprSet,2,as.numeric)
+rownames(exprSet)=rn
+
+#exprSet=log(exprSet)
+#boxplot(exprSet,las=2)
+
+pdata=pData(GSE70529[[1]])
+## check the metadat and find the correct group information
+# we must tell R that group should be interpreted as a factor !
+individuals=factor(unlist(lapply(pdata$title,function(x) strsplit(as.character(x),"-")[[1]][1])))
+treatment=unlist(lapply(pdata$title,function(x) strsplit(as.character(x),"-")[[1]][2]))
+treatment=factor(sub('2','',treatment))
+
+## if only take the treatment into accont 
+design=model.matrix(~treatment)
+fit=lmFit(exprSet,design)
+fit=eBayes(fit) 
+#vennDiagram(fit,include=c("up","down"))
+write.csv(topTable(fit,coef='treatmentB',n=Inf,adjust='BH'),"treatmentBvsA.DEG.csv")
+write.csv(topTable(fit,coef='treatmentC',n=Inf,adjust='BH'),"treatmentCvsA.DEG.csv")
+write.csv(topTable(fit,coef='treatmentD',n=Inf,adjust='BH'),"treatmentDvsA.DEG.csv")
+## and we don't care about the difference between individuals
+