Tweet
Hi,
I am using cuffdiff and Deseq/edgeR to compare Gene expression levels between mice kidneys between a control and an affected group.
Both outputs are comparable, the difference is that cufflinks also finds gene isoforms.
So for example if you compare a random gene you will probably get a 1:1 ratio with the raw read data. But if you look it up in cuffdiff, there are sometimes up to 5 isoforms with very expression ratios. But if you add them together you will get a 1:1 ratio again.
There are some cases where the isoform seems to "switch".
So in the control group you have average 5 fpkm for isoform A and 0 for isoform B,
but in the affected group it's the other way around: isoform B has average 5 fpkm and isoform A has nothing.
Most of the times when the isoform varies, its with very low FPKM
example:
From Gene X,
isoform A has 40 FPKM in control and 38 in affected
isoform B has 0 FPKM in control and 0.5 in affected
isoform C has 5 in control and 2 in affected
if you add these values, you get the same ratio for Gene X you would get in the raw reads file.
So my question now is if that means something. Can different isoforms affect the kidneys, or is what I see nothing out of the ordinary and the variations are natural/statistical Errors
I am using cuffdiff and Deseq/edgeR to compare Gene expression levels between mice kidneys between a control and an affected group.
Both outputs are comparable, the difference is that cufflinks also finds gene isoforms.
So for example if you compare a random gene you will probably get a 1:1 ratio with the raw read data. But if you look it up in cuffdiff, there are sometimes up to 5 isoforms with very expression ratios. But if you add them together you will get a 1:1 ratio again.
There are some cases where the isoform seems to "switch".
So in the control group you have average 5 fpkm for isoform A and 0 for isoform B,
but in the affected group it's the other way around: isoform B has average 5 fpkm and isoform A has nothing.
Most of the times when the isoform varies, its with very low FPKM
example:
From Gene X,
isoform A has 40 FPKM in control and 38 in affected
isoform B has 0 FPKM in control and 0.5 in affected
isoform C has 5 in control and 2 in affected
if you add these values, you get the same ratio for Gene X you would get in the raw reads file.
So my question now is if that means something. Can different isoforms affect the kidneys, or is what I see nothing out of the ordinary and the variations are natural/statistical Errors