Cufflinks FPKM is different than Cuffdiff FPKM !!

[fuad193]

Why I get different FPKM values if I run a BAM file on cufflinks alone and on cuffdiff against another BAM file ? Aren't FPkM values consistant irrespective to whether it comes from cufflinks or cuffdiff ?

What should I do now, consider cuifflinks FPKM or cuffdiff FPKM ? I mainly want to compare samples in time series and on WT vs mutant. Any suggestions ?

Edited: MY main purpose is to find the half life of each Transcripts RNA, so I want to construct a plot of log(FPKM) vs time. Which FPKM should I use ?!

[NGSfan]

Why I get different FPKM values if I run a BAM file on cufflinks alone and on cuffdiff against another BAM file ? Aren't FPkM values consistant irrespective to whether it comes from cufflinks or cuffdiff ?

What should I do now, consider cuifflinks FPKM or cuffdiff FPKM ? I mainly want to compare samples in time series and on WT vs mutant. Any suggestions ?

Edited: MY main purpose is to find the half life of each Transcripts RNA, so I want to construct a plot of log(FPKM) vs time. Which FPKM should I use ?!

Please find my solution in this thread http://seqanswers.com/forums/showthread.php?t=26041. The problem is that the default settings between cufflinks and cuffdiff are different.



btw- there is a new version of cufflinks/cuffdiff 2.1.1 with some bug fixes, although I haven't tested that they have fixed the default setting of the multi-frag-hits (reported as unlimited but actually set to 1 in 2.0.2)