Cufflinks and cuffdiff FPKM values

[combiochem]

I have questions about the FPKM values in each program.
As I understand, the cufflinks or cuffdiff are counting the reads based on the gene structures, when the known annotations are provided. And FPKM value is a kind of absolute value for each transcript. In that case, can we expect the same FPKM value for the same sample, from each program, cuffdiff and cufflinks? or Is my understating not right?

I tried two programs with same mapping results and same a known annotation file, but the FPKM values are different each other (from transcripts.expr and isoforms.fpkm_tracking files) Is cuffdiff differently estimating the expression of the transcript (FPKM) because of other sample?

Thanks for any comments in advance.

[thinkRNA]

I have questions about the FPKM values in each program.
As I understand, the cufflinks or cuffdiff are counting the reads based on the gene structures, when the known annotations are provided. And FPKM value is a kind of absolute value for each transcript. In that case, can we expect the same FPKM value for the same sample, from each program, cuffdiff and cufflinks? or Is my understating not right?

I tried two programs with same mapping results and same a known annotation file, but the FPKM values are different each other (from transcripts.expr and isoforms.fpkm_tracking files) Is cuffdiff differently estimating the expression of the transcript (FPKM) because of other sample?

Thanks for any comments in advance.

I think cuffdiff uses a likelihood function to estimate FPKM which is not an absolute value (Read this thread: http://seqanswers.com/forums/showthread.php?t=3961)
Which other program did you compare it with? If this other program is not using the same methodology, you will not get the same answer.

[combiochem]

I think cuffdiff uses a likelihood function to estimate FPKM which is not an absolute value (Read this thread: http://seqanswers.com/forums/showthread.php?t=3961)
Which other program did you compare it with? If this other program is not using the same methodology, you will not get the same answer.

No, I'm not comparing the FPKM values from "other" program. I've compared the FPKM between cufflinks output and cuffdiff output. As far as I know, in cufflinks, when the known annotations are given, the FPKM for each transcript can be obtained (in one sample). Also in cuffdiff tracking file, the FPKM in each sample (q0 and q1) can be obtained. I've used exactly same annotations and mapping results (which have used for cufflinks) for cuffdiff running (therefore the read counts for each transcript should be same), but the FPKM outputs are different from cufflinks. I'm wondering whether the estimation is taking care of both samples in cuffdiff (which is different from cufflinks?) rather than each sample (q0 or q1).

[mrfox]

i actually went through the cufflinks-cuffcompare-cuffdiff pipeline and I also have the same problems. cufflinks and cuffdiff report different FPKM values for a same transcript.

[Thomas Doktor]

EDIT
My question was answered in another thread.

[lewewoo]

Where I can find it? Please post the link, thanks!

[mrfox]

Please download the latest Cufflinks from the link below. I think now cufflinks and cuffdiff return consistent values.

http://cufflinks.cbcb.umd.edu/manual.html

[reut]

same problem with the latest Cufflinks version

I am using cufflinks version 1.0.2 and have the same problem.
Any suggestions?

[Thomas Doktor]

This might explain it:

Cuffdiff and Cufflinks now accept new options controlling whether all hits are counted towards the FPKM denominator, or only those compatible with some transcript in the reference annotation. Counting only compatible hits avoids certain types of bias that arise when one sample contains far more hits that aren't compatible with any transcript than the other sample does. For example, if one sample contains vastly more mapped ribosomal RNA hits, FPKM values will appear lower in that sample, potentially leading to false positive differential expression calls. Cuffdiff by default now uses only compatible hits. Cufflinks still uses total hits by default, as using compatible hit accounting requires a reference GTF.

[arrchi]

Have anyone also seen that some isoforms exist in the files generated by Cufflinks but disappear in the files generated by Cuffcompare and Cuffdiff?

[lewewoo]

cufflinks or cuffcompare bugs?

1. New released 1.3.0, after Cuffcompare, FPKM column contains all 0, missing FPKM values even tracking files have them;

2. in all the versions of CuffDiff, if you compare different conditions against the same control samples, the FPKM in the same control samples in different comparing is different; for example,
CuffDiff I: condition 1 v.s. condition control;
CuffDiff II: condition 2 v.s condition control;

after CuffDiff, when FPKM numbers are tracked, the FPKM of Gene X in condition control in CuffDiff I is different from the FPKM of Gene X in condition control in CuffDiff II. GeneX roughly are 20-30% in total annotated genes and the rest are the same.
anybody has explanation or suggestions for this? Thanks!

[tangx_2010]

after CuffDiff, when FPKM numbers are tracked, the FPKM of Gene X in condition control in CuffDiff I is different from the FPKM of Gene X in condition control in CuffDiff II. GeneX roughly are 20-30% in total annotated genes and the rest are the same.
anybody has explanation or suggestions for this? Thanks!

Did you use the "-N" option? I found that FPKM are different when I used "-N" option. And FPKM is consistent between different comparison when I closed "-N" option.

[IBseq]

hi tangx 2010,
i have used cuffdiff on the galaxy platform...what exaclty is the N option?

how can i select/deselect this option?

thanks,ib