[R] constrast in edge R

Noha Osman nmo_138 at usc.edu
Wed Jan 14 20:41:21 CET 2015


Hey everyone,


I have data frame called subtest as following:
RNA.LATER.MEN2B_S5 RNA.LATER.ROSA_S4 RNA.MEN2B.1_S2 RNA.MEN2B.2_S3 RNA.ROSA_S1
1 13707 13866 12193 12671 10178
2 0 0 0 0 1
3 7165 5002 1256 1341 2087
6 8537 16679 9042 9620 19168
10 19438 25234 15563 16419 16582
16 3 3 11 3 5
I would like to analysis the MEN samples to ROSA samples I did script like that .
> group=c("LMS5","LRS4","MS2","MS3","RS1")
> y=DGEList(counts=data.matrix(subtest),group=group ,genes=genes)
> indices=which(rowSums(cpm(y)>1)<3)
> y=y[-indices,]
> y=calcNormFactors(y)
> design=model.matrix(~0+group,data=subtest)
> cont.matrix= makeContrasts(groupLMS5-groupMS2-groupMS3-(groupLRS4-groupRS1),levels=design)
> fit <- glmFit(y, design,dispersion=y$trended.dispersion)
> lrt.all <- glmLRT(fit, contrast=cont.matrix)
> topTags(lrt.all)
summary(dexp<-decideTestsDGE(lrt.all,p=0.05,adjust="BH"))
-1 1
0 36
1 14248
I got one gene as down regulated and 14248 as upregulated so I think I did something wrong.please Iam a new user in that package and I want to get a appropriate number of up and down regulated genes to downstream analysis

	[[alternative HTML version deleted]]



More information about the R-help mailing list