Hi,
I have problem on DESeq, it generates different results using same dataset (without replicate), and almost same code.
Here is code for the 1st test:
it generates 225 differentially expressed genes with adjust pvalue < 0.05.
But when i change the code for the 2nd test:
it just generates 84 differentially expressed genes.
The 2nd test just estimate the two samples related to the comparison, but the 1st test estimate all the samples.
I do not know why they are different, and I want to know which one is more reliable.
Thanks a lot.
By the way, before import dataset into DESeq, I usually to remove the genes without expression in each conditon.
Here is a example.
cond1 cond2 cond3 treat1 treat2 treat3
gene1 1 2 3 2 2 0
gene2 0 0 1 0 0 0
gene3 0 0 0 0 0 0 (removed)
For the dataset before, there are 225 DE genes identified at 1st test.
After removing none-expressed genes, 220 DE genes were identified with adjust pvalue < 0.05.
I want to know which method is correct.
Thanks again
I have problem on DESeq, it generates different results using same dataset (without replicate), and almost same code.
Here is code for the 1st test:
Code:
conds <- factor( c( "c0","c6","c12","c24","c48","c72","c96","c120","t0","t6","t12","t24","t48","t72","t96","t120") ) cds<-newCountDataSet(countsTable, conds) cds <- estimateSizeFactors( cds ) sizeFactors( cds ) cds <- estimateDispersions( cds, method="blind", sharingMode="fit-only" ) test0 <- nbinomTest( cds, "c24", "t24" ) write.table( test0, sep="\t", file="test0.txt" )
But when i change the code for the 2nd test:
Code:
conds <- factor( c( "c0","c6","c12","c24","c48","c72","c96","c120","t0","t6","t12","t24","t48","t72","t96","t120") ) cds<-newCountDataSet(countsTable, conds) cds <- estimateSizeFactors( cds ) sizeFactors( cds ) cds2 <- cds[ ,c( "c24", "t24" ) ] cds2 <- estimateDispersions( cds2, method="blind", sharingMode="fit-only" ) test1 <- nbinomTest( cds2, "c24", "t24" ) write.table( test1, sep="\t", file="test1.txt" )
The 2nd test just estimate the two samples related to the comparison, but the 1st test estimate all the samples.
I do not know why they are different, and I want to know which one is more reliable.
Thanks a lot.
By the way, before import dataset into DESeq, I usually to remove the genes without expression in each conditon.
Here is a example.
cond1 cond2 cond3 treat1 treat2 treat3
gene1 1 2 3 2 2 0
gene2 0 0 1 0 0 0
gene3 0 0 0 0 0 0 (removed)
For the dataset before, there are 225 DE genes identified at 1st test.
After removing none-expressed genes, 220 DE genes were identified with adjust pvalue < 0.05.
I want to know which method is correct.
Thanks again
Comment