Reverse order of Cuffdiff samples?

Question

[biopat]

Hi all!

I have run cuffdiff using the --use-sample-sheet option and I separated my 16 samples into Knockout and Wildtype. In the text file the first 2 files are labeled Wildtype and for the rest the order is mixed. I was expecting that the comparison will be between Wildtype and Knockout in this order on the output sheet. However the Knockout is listed in the first column and the Wildtype in the second column. This also has an effect on my log2 fold change. I don't want to manually manipulate the excel file! Is there any way to tell cuffdiff to compare Wiltype to Knockout and not Knockout to Wildtype? (I want to continue using the sample sheet option)

Thank you!

[Satyajeet Khare]

Did you use a contrast file along with sample sheet? Can you try by swapping KO with WT and run cuffdiff again? What was your command?

[biopat]

I added the file. It actually seems to order it alphabetical when run this way, so what I did is call Wildtype "aWildtype" and the Knockout"bKnockout" it then used the Wildtype first and compared it to the KO. Thanks for the help though!

Answer 1

[Mike]

In the Cuffdiff command, the order of input sample matters, if you reverse your sample order, the sign of the log2 fold-change will flip.