cufflinks multiple FPKM values for same location

[abh]

Hi,

i have run cufflinks for my Bam files to look after the FPKM values and i got the output.

later i took the output transcripts.gtf and converted into bedgraph file to view in UCSC genome browser,but i got different fpkm values for same location

what should i do with them?should i sum up?is there any script or any tool that i can use it(just for saving the time)

here is the sample bedgraph file

track type=bedGraph name=s6 visibility=full color=255,0,0
chr1 11874 12227 0.0000000000
chr1 11874 12227 0.0286938296
chr1 11874 12227 0.1072298182
chr1 11874 14409 0.0000000000
chr1 11874 14409 0.0286938296
chr1 11874 14409 0.1072298182
chr1 12595 12721 0.0000000000
chr1 12613 12721 0.1072298182
chr1 12646 12697 0.0286938296

Thank you

[glados]

I get the same problem. I get several gene_ids with identical coordinates. It's so frustrating. How come cufflinks don't understand it is only one gene when they have identical coordinates? (using the latest cufflinks 2.1.1)

[SrCardgage]

It's not so much cufflink's problem as it is the records in the gtf file. Like me, you probably grabbed a publicly available version of the gtf file. This file has many TRANSCRIPTS for the same gene. These usually represent different isoforms of the gene. For some analyses, it is important to know the distinction.

My analysis did not need to know this, so I wrote a python script to choose only one transcript. Here is my script. It is VERY kludgy, but it works.

Please note that I had to change the file extension to ".txt" from ".py" in order to include it in this post.

The logic behind isolating down to a single transcript is near the top of the file, in the comments. If you have problems understanding this, please DO NOT pm me. Post to this thread and I will respond when I find the time (probably a few days).

And please, no comments on my coding conventions; positive or negative.