[gringer]

DESeq needs count data in the form of rectangular table. My question is whether is correct or possible to use the genes.count_tracking file generated with cuffdiff as the counts table that DESeq requires?

It is not correct, but it is possible to force DESeq to accept it. DESeq expects raw read counts, not normalised counts, as in the count_tracking files. As a warning, DESeq will complain if you try to feed it decimal values in the data table -- this is an indication that you're doing something you shouldn't.

If you're using cuffdiff to get the count data, why not use it to do the differential analysis as well?

[colaneri]

combining tophat with DESeq

It is not correct, but it is possible to force DESeq to accept it. DESeq expects raw read counts, not normalised counts, as in the count_tracking files. As a warning, DESeq will complain if you try to feed it decimal values in the data table -- this is an indication that you're doing something you shouldn't.

If you're using cuffdiff to get the count data, why not use it to do the differential analysis as well?

The reason is because I'm comparing the response of two genotype to two growing conditions, and I want to study the interactions of genotype x treatment

When using cuffdiff I just obtain a comparison of everything against everything. But most of that comparison are not usefull. DESeq can to two factor analysis.

The problem is I do not know how to create the entry table with good metadata about gene information, that is what I was thinking in use the gene.count tracking file. I like how tophat align the sequences also.

But now regarding what you told me of decimal numbers. I'm not talking to feed DESeq with RFPK values. It look to me that the gene.count traking files contains the raw number of reads mapped to each gene

[dpryan]

Aren't the counts in the gene.count_tracking file scaled (i.e. they can be decimal)? Normally, one would just use htseq-count to generate a count table for DESeq. That's the simplest way to go about things and then you can easily check if the raw counts do in fact match those in the gene.count_tracking file.