When I am running cuffnorm (from cufflinks) on some RNA-Seq BAM files with the GENCODE v19 annotation, I do get out several outputs, most notably genes.fpkm_table
and isoforms.fpkm_table
.
When I look into these files I do obtain something like
tracking_id q1_0 q2_0
ENSG00000000003.10 4.46257 2.81516
ENSG00000000005.5 0 0
ENSG00000000419.8 58.9657 73.9058
ENSG00000000457.9 9.86299 5.55451
ENSG00000000460.12 20.4105 9.2218
ENSG00000000938.8 0.01044 0.0226546
....
for the genes files and something like
tracking_id q1_0 q2_0
ENSG00000000003.10 0.139293 0.00827424
ENSG00000000005.5 0 0
ENSG00000000419.8 0.43631 0.520737
ENSG00000000457.9 0.440151 0.236529
ENSG00000000460.12 0.0721438 0.0884936
ENSG00000000938.8 0.000730916 0.00195449
...
ENST00000000233.5 179.357 91.6825
ENST00000000412.3 58.6248 75.4492
ENST00000000442.6 23.4939 13.6916
ENST00000001008.4 238.631 382.162
ENST00000001146.2 0.0702716 0.0796778
...
for the isoforms.
So we do see that the genes are incorporated into the isoforms FPKM file. Here, however, the values are very close 0 and in particular different from the respective values of the genes FPKM file.
Why at all these gene FPKM values are appearing in the isoform FPKM file as well?
Why are they basically 0 and hence different from those values of the gene FPKM file?
Did I do something wrong and, if so, how do I fix it?
Many thanks for your help in advance!
Can you make it clear ? What is your question exactly ? Are you trying to compare the values between genes vs transcripts from cuffdiff ? Can you take the values of one gene and values of its transcripts and ask the question clearly ? Instead of copy pasting some random lines and ask users to understand ?
Many thanks for your reply. I hope my changes are making my issues a bit clearer.
Usually in the isoform file you will have several isoforms for each gene, and some of them might have very low read counts.
How many isoforms you have for ENSG00000000003? And how many entry for it in the gene file? Please, paste the names and the counts for both files.