[BioC] Insignificant q-values in minfi

[kaushal Raj Chaudhary [guest]]

Hello List,

I am analyzing 450 k methylatio data using minfi package and getting large (~1) q-values.  I want to remove the probes with lower variance to improve q-value (so that there will be less test(multiple test corection ).  How can I do this in beta values object prior to using dmpfinder function?

Thanks for help.
code used:

gRatioSet.quantile <- preprocessQuantile(RGset, fixOutliers = TRUE,
                                         removeBadSamples = TRUE, badSampleCutoff = 10.5,
                                         quantileNormalize = TRUE, stratified = TRUE,
                                         mergeManifest = FALSE, sex = NULL)
colnames(gRatioSet.quantile)


beta<-getBeta(gRatioSet.quantile)
nrow(beta)

Beta.noSNPs=rmSNPandCH(beta, dist = 2, mafcut = 0.05)    ### removing SNP probes 
nrow(Beta.noSNPs)

dmp <- dmpFinder(Beta.noSNPs, pheno=pData(gRatioSet.quantile)$Sample_Group, type="categorical")   ### methylated sites
head(dmp,n=100)

 -- output of sessionInfo(): 

> head(dmp,n=100)
                 intercept        f         pval      qval
cg19314777      0.85646150 23.31203 3.057187e-05 0.9417651
cg15306012      0.60817234 23.23892 3.125969e-05 0.9417651
cg01224520      0.81419046 23.19462 3.168447e-05 0.9417651
cg08515264      0.81986249 23.18189 3.180761e-05 0.9417651
cg25110423      0.79139248 22.89596 3.471317e-05 0.9417651
cg11642412      0.19004546 22.71170 3.673422e-05 0.9417651
cg26331096      0.70306567 22.46844 3.959597e-05 0.9417651

--
Sent via the guest posting facility at bioconductor.org.