By default Cufflinks calculates FPKM by using all aligned hits in the denominator while Cuffdiff calculates FPKM by using only reads aligned to a reference transcript.
From reading through other posts I understand why Cuffdiff does this to prevent bias between samples. But why are the methods not consistent?
From reading through other posts I understand why Cuffdiff does this to prevent bias between samples. But why are the methods not consistent?