hey,
we sent 3 Affected and 3 Control Mice Samples for Sequencing to compare Gene Expression Levels and we got back a bunch of files that were generated with cufflinks. So theoretically we would now compare the Gene expression Levels between Affected and Control to see where they vary and what might cause the phenotype.
The problem is that I dont know which files to use.
basically there are 3 excel Files.
isoform_exp.diff.with.fpkm_denovo
isoform_exp.diff.with.fpkm_known
dge_results
(Differential Gene Expression)
so the first 2 are the fpkm values(de novo and known) and the last one are basically raw read counts from what I can understand.
Theroetically it shouldnt matter which file I use because all we need is the ratio from control group to affected to find out which genes are less/more expressed in the affected group.
But when I try that with the different files I get huge differences that should not be there(if I understand it correctly)
For example Errc2.
In dge results it's a 1:1 ratio between control and affected, in fpkm_known it's 1:10 and in fpkm_denovo the gene isnt even listed.
also I would like to know how they get their FPKM values.
FPKM means something like Fragments per kilobase of exom per million reads.
But some Genes have like [S1:40, S2:34, S3:54] in raw reads, but in FPKM it looks like [S1:1.6, S2:0.00005, S3: 1.3]
I also dont understand how cufflinks calculates p-values.
There are some genes where FPKM is zero for all the samples, but they still got totally different p values how can that be?
we sent 3 Affected and 3 Control Mice Samples for Sequencing to compare Gene Expression Levels and we got back a bunch of files that were generated with cufflinks. So theoretically we would now compare the Gene expression Levels between Affected and Control to see where they vary and what might cause the phenotype.
The problem is that I dont know which files to use.
basically there are 3 excel Files.
isoform_exp.diff.with.fpkm_denovo
isoform_exp.diff.with.fpkm_known
dge_results
(Differential Gene Expression)
so the first 2 are the fpkm values(de novo and known) and the last one are basically raw read counts from what I can understand.
Theroetically it shouldnt matter which file I use because all we need is the ratio from control group to affected to find out which genes are less/more expressed in the affected group.
But when I try that with the different files I get huge differences that should not be there(if I understand it correctly)
For example Errc2.
In dge results it's a 1:1 ratio between control and affected, in fpkm_known it's 1:10 and in fpkm_denovo the gene isnt even listed.
also I would like to know how they get their FPKM values.
FPKM means something like Fragments per kilobase of exom per million reads.
But some Genes have like [S1:40, S2:34, S3:54] in raw reads, but in FPKM it looks like [S1:1.6, S2:0.00005, S3: 1.3]
I also dont understand how cufflinks calculates p-values.
There are some genes where FPKM is zero for all the samples, but they still got totally different p values how can that be?
Comment