Grouping non replicates in Cuffdiff

[jtipton]

I'm new to bioinformatics and I have what seems like a weird task. I have 7 samples of a fungus with 2 replicates each. I want to compare 1 of these samples vs all 6 of the rest. However I'm not sure how to set this up.

Would I basically do cuffdiff -L group1, group2 sample1r1,sample1r2 sample2r1,sample2r2,sample3r1,sample3r2... and put them all in the same group and treat them all like replicates? This doesn't seem right, but I haven't found a better way to group them.

[jparsons]

it depends on what you mean by "vs all 6 of the rest".

If you want to treat them like replicates, then yes, you do it like you suggested.

If you want to compare vs each of the other 6 individually, you want to treat each of the 7 as its own separate group, label each of them individually, and then filter your results to include only group 1 vs 6-7 (there will be data for each pairwise comparison)

You'd want to do something like -l group1,group2(through7) sample1r1,1r2 sample2r1,2r2(through7).