the gwa by default uses benjami hochberg fdr adjusted procedure, but it is nice to have user input fdr adjustment to the first thing that works; so by default you can input holm, and if that is too conservative it will return the next best conservative list, ending at "none" , or stopping. here we grab only the Normal samples so this should error out; thus getting no DE. this shows a custom adjust
library(arkas) library(TxDbLite) library(arkasData) #outputPath <- system.file("extdata", "", package="arkasData") #samples <- list.files(outputPath, pattern="^[ns][124]$") #covs <- data.frame(outputDir=samples, # row.names=samples) #NN <- mergeKallisto(covariates=covs, # outputPath=outputPath) # #show(NN) #NN <- annotateFeatures(NN) # loads TxDbLite libs outputPath<-system.file("extdata/annotatedKexp","",package="arkasData") load(paste0(outputPath,"NS.RData")) NN<-NS NN$norm <- as.factor(substr(colnames(NN), 1, 1)) NN$subject <- as.factor(substr(colnames(NN), 2, 2)) NN_design <- with(as(colData(NN), "data.frame"), model.matrix(~ norm+ subject)) gwa<-geneWiseAnalysis(NN,design=NN_design,how="cpm",species="Homo.sapiens",adjustBy="BY",fitOnly=FALSE) rwa<-repeatWiseAnalysis(NN,design=NN_design,how="cpm",species="Homo.sapiens",adjustBy="holm")
Repeat elements usually have low signal, so testing safe exit would be useful here.
the MDS data set had 0 DE returned from rwa. The mds data will return 0 DE for repeat level analyses, and is a great example of flexible FDR searching. Here users can input any adjust method, and it will run until it finds the selected fdr procedure or the next conservative one.
outputPath <- system.file("extdata", "annotatedKexp", package="arkasData") load(paste0(outputPath,"/mds.RData")) design<-metadata(mds)$design rwa<-repeatWiseAnalysis(mds,design=design,how="cpm",species="Homo.sapiens",adjustBy="holm",p=0.01)
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.