I run cuffcompare on 7 cufflinks' gtf outputs (each corresponding to a different sample). I would like to compare the FPKM values by gene across the samples, but in many cases a given transfrag is absent from one or more sample ('-' in the sample column of the *.tracking output file).
I thought of replacing the '-' with zeros, but then I cannot take the log. Does it make sense to replace the '-' with zeros and then add 1 to all the data?
Can I handle FPKMs as if they were expression values for genes?
Thanks in advance,
PFS
I thought of replacing the '-' with zeros, but then I cannot take the log. Does it make sense to replace the '-' with zeros and then add 1 to all the data?
Can I handle FPKMs as if they were expression values for genes?
Thanks in advance,
PFS