Hello all,
I have several subsets of a datasets from one RNA seq experiment and I
need to compare the reproducibility of these different subsets with
DESeq.
I have both FDR and P values for each gene in my subsets and I need to
transform P value or FDR to RANK coordination. Then I can plot the
reproducibility(y axis) according to this rank(x axis) which has
obtained from p value or FDR for each differentially expressed gene.
Any idea for transforming one of these parameters to obtain a rank?(I
should transform p value or fdr to this rank and use it as x axis of
my plot).
Thanks a lot in advance
Narges
Hi,
On 07/11/2012 11:00 AM, Fatemehsadat Seyednasrollah wrote:
> I have several subsets of a datasets from one RNA seq experiment and
> I need to compare the reproducibility of these different subsets
with
> DESeq. I have both FDR and P values for each gene in my subsets and
I
> need to transform P value or FDR to RANK coordination. Then I can
> plot the reproducibility(y axis) according to this rank(x axis)
which
> has obtained from p value or FDR for each differentially expressed
> gene. Any idea for transforming one of these parameters to obtain a
> rank?(I should transform p value or fdr to this rank and use it as x
> axis of my plot).
To convert a list of values to ranks in R, simply use the 'rank'
function.
However, I am rather sceptical that comapring ranks of p values is a
suitable method to assess reproducibility of an experiment.
Simon
