Reads are probabilistically assigned, so raw read counts are not available from Cufflinks. Recovering raw fragment counts could be done by reverse-engineering the FPKM value, but Cufflinks doesn't do this for you. If you choose to do this, keep in mind that Cufflinks uses an "effective transcript length". Best, J. On Feb 8, 2012, at 11:06 PM, Li, Jilong (MU-Student) wrote:
Dear Jeremy,
Sorry, I didn't expressed my question clearly. I got the FPKM normalized values for each gene from cufflinks. And I want to get the original reads counts that were not normalized from cufflinks. Could you please tell me how to get those?
Thank you very much!
Victor
From: Jeremy Goecks [jeremy.goecks@emory.edu] Sent: Thursday, February 09, 2012 4:00 AM To: Li, Jilong (MU-Student) Cc: galaxy-user@lists.bx.psu.edu Subject: Re: [galaxy-user] How to get reads counts from cufflins?
Victor,
I got the normalized values (FPKM) from cufflinks. And I want to get relative reads counts. How can I do that?
It's not clear to me what you're looking for. FPKM is a normalized read count metric where the F stands for fragment, which is a single read (or half of a paired read).
Another question: how does cufflinks handle isoform genes while calculating the reads counts? Or what papers can help me understand this?
Expectation maximization is used to probabilistically assign reads to isoforms. See the Cufflinks documentation for details and paper links:
http://cufflinks.cbcb.umd.edu/
Best, J.