Hello,

Iam Using affy and lima packages for differential analysis of gene for my data

i have 10 .CEL (Adult_A172.CEL, Adult_LN229.CEL, Adult_SF268.CEL, Adult_U87MG.CEL, Adult_U118MG.CEL, Pediatric_KNS42.CEL, Pediatric_Res186.CEL, Pediatric_Res259.CEL, Pediatric_SF188.CEL, Pediatric_UW479.CEL) files. this data i obtaied from the ArrayExpress website https://www.ebi.ac.uk/arrayexpress/experiments/E-TABM-579/ 

i have refered to one of  the example from bioconductor and the code is given below 

but i don't understand why its not working as I am creating the design matrix in similar way. once i fir the data using lmFit till top table its showing no error but i want to select p values less than 0.05 and after doing p value commad its showing error. and i want to crete a heatmap of those Deferentially expressed genes.

i have refered to one of  the example from bioconductor and the code is given below 

library(affy)
library(limma)

affy_data <- ReadAffy()
eset_rma <- rma(affy_data)
exprSet.nologs <- exprs(eset_rma)
head(exprSet.nologs)
# List the column (chip) names
colnames(exprSet.nologs)
#log transformation
exprSet = log(exprSet.nologs, 2)
head(exprSet)
#To print out our expression matrix
write.table(exprSet, file="MGMT_matrix.txt", quote=F, sep="\t")

XX<- as.matrix(read.table("MGMT_matrix.txt",header = TRUE,sep = "\t"))
head(XX)
minimalSet <- ExpressionSet(assayData = XX)
minimalSet
ex <- exprs(minimalSet)
head(ex)

groups<-as.factor(c(rep("Adult",5),rep("Pediatric",5)))
groups
design_new<-model.matrix(~0+groups)
colnames(design_new)=levels(groups)
design_new
fit<-lmFit(minimalSet, design_new)
head(fit)
fit
cont.matrix<-makeContrasts(Adult-Pediatric, levels=design_new)
cont.matrix
fit2<-contrasts.fit(fit, cont.matrix)
fit2
ebfit<-eBayes(fit2)
head(ebfit)
ebfit
tt_new <-topTable(ebfit, coef=1)
tt_new
select_new <- p.adjust(ebfit$p.value)<0.05 ####error here
select_new
gg <- minimalSet[select_new,]
gg
#tt_row <-nrow(topTable(ebfit, coef=1, lfc=1))
#tt_row

Another way i tried to do the same exercise usig the following code,

data_age <- ReadAffy()
eset <- rma(data_age)
pData(eset)

# matrix design
design.mat <- cbind(c(1,1,1,1,1,0,0,0,0,0),c(0,0,0,0,0,1,1,1,1,1))
colnames(design.mat)<- c("Adult","Pediatric")
design.mat

#design matrix: model.matrix
sample_10 <- factor(rep(c("Adult","Pediatric"),each=5))
design.mat<- model.matrix(~0+sample_10)
colnames(design.mat)<- levels(sample_10)
design.mat

#contrast matrix: makeContrasts()
contrast.mat <- makeContrasts(Diff = Adult-Pediatric,levels = design.mat)
contrast.mat

#fit limma
fit_data <- lmFit(eset, design.mat)
fit_data
fit_data2 <- contrasts.fit(fit_data,contrast.mat)
fit_data2
fit3 <- eBayes(fit_data2)
fit3
dim(fit3)
selected <- p.adjust(fit3$p.value[,2])<0.001 ### error
selected
deg <- topTable(fit3,coef = 'Diff',p.value=0.05, adjust.method='fdr',lfc = log2(1.5),number = nrow(eset))
dim(deg)[1]
#length(deg)
#fit4 <- exprs(fit3)

it again is not woring

There are actually no errors here from the limma package. You are just making some programming errors in R after you've run limma. It's not my responsibility to debug your code, but here is a hint. If a matrix X has only one column, then trying to access the second column by X[,2] will give an error. That is why fit3$p.value[,2] gives an error.

Note also that you never need to call p.adjust() yourself, because limma does this for you with topTable() etc. Anyway, I don't think you are running p.adjust() correctly -- you are asking it do perform Bonferroni adjustment, which I don't think is really what you want.