search

hansenlab / bsseq

Devel repository for bsseq
35 stars 25 forks source link

DRM p-values to calculate FDR #37

Open BegonaMartinez opened 8 years ago

BegonaMartinez commented 8 years ago

Hi Kasper,

I have obtained a set of DMR between my two groups and I would like to estimate the FDR using the qvalue package in Bioconductor. I may have missed something during the process of analyses, but I am not able to recover the p-values associated to my DMRs. How could I get these p-values? My pipeline from the smoothing is writen below. Many thanks in advance.

Best, Begoña

File.data.smoothed <- BSmooth(File.data,mc.cores=7,verbose=TRUE) File.data.tstats <- BSmooth.tstat(File.data.smoothed,group1=c("A","B","C","D"), \ group2=c("E","F","G"),estimate.var="same",local.correct=TRUE, verbose=TRUE) dmr.list <- dmrFinder(File.data.tstats,qcutoff=c(.005,.995)) dmrs <- subset(dmr.list, n>=10 & abs(meanDiff) >= 0.3) write.table(x=dmrs, file="File.data.txt", sep="\t", row.names= FALSE, quote=FALSE)

kasperdanielhansen commented 8 years ago

we get p-pvalues corrected for multiple testing using permutation tests. You don't get a p-value out of this pipeline per se.

On Fri, May 27, 2016 at 12:34 PM, BegonaMartinez notifications@github.com wrote:

Hi Kasper,

I have obtained a set of DMR between my two groups and I would like to estimate the FDR using the qvalue package in Bioconductor. I may have missed something during the process of analyses, but I am not able to recover the p-values associated to my DMRs. How could I get these p-values? My pipeline from the smoothing is writen below. Many thanks in advance.

Best, Begoña

File.data.smoothed <- BSmooth(File.data,mc.cores=7,verbose=TRUE) File.data.tstats <- BSmooth.tstat(File.data.smoothed,group1=c("A","B","C","D"), \ group2=c("E","F","G"),estimate.var="same",local.correct=TRUE, verbose=TRUE) dmr.list <- dmrFinder(File.data.tstats,qcutoff=c(.005,.995)) dmrs <- subset(dmr.list, n>=10 & abs(meanDiff) >= 0.3) write.table(x=dmrs, file="File.data.txt", sep="\t", row.names= FALSE, quote=FALSE)

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/kasperdanielhansen/bsseq/issues/37, or mute the thread https://github.com/notifications/unsubscribe/AEuhnxLdS2RzpkqAx2YvcQkLUgbMbLojks5qFx0tgaJpZM4Ioqtb .

BegonaMartinez commented 8 years ago

thanks a lot for the answer. Is there a package to do the permutation tests?

kasperdanielhansen commented 8 years ago

We have some code in bsseq in the permutations.R file, but it is currently unexported and undocumented. Thats what we use, and we are working on finishing it up.

Best, Kasper

On Fri, May 27, 2016 at 3:05 PM, BegonaMartinez notifications@github.com wrote:

thanks a lot for the answer. Is there a package to do the permutation tests?

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/kasperdanielhansen/bsseq/issues/37#issuecomment-222228773, or mute the thread https://github.com/notifications/unsubscribe/AEuhn7ORom8FVT29q8D_-V_CmVSbeSOBks5qF0BigaJpZM4Ioqtb .

dktanwar commented 5 years ago

Hi,

I wonder if it's been implemented to obtain p-value?