Cuffdiff reports FPKMs outside the range of replicates that go into it

[rflrob]

Hi all,
I'm having an issue where cuffdiff is reporting an FPKM value for the sample is greater than any of the FPKMs reported for the 3 replicates. I'm pretty sure I'm running everything with the same bias estimation parameters. Is this a sign of something fishy, or is there an easy explanation I can't think of?

Thanks!
Peter

[Wolfgang Huber]

Dear Peter

I have no substantial answer to your problem, but I suppose that the people who might have one would appreciate if you could post a 'reproducible example' of your problem. A reproducible example is a sample of code and data that another user can run and get the same results as you do. You should also include the version number of any software you are using. cuffdiff in particular has been noted for its version-dependent behaviour.

best wishes
Wolfgang