survClust: survAnal: runs cox proportional hazard models against...
In PietaSchofield/plibb: Pieta Schofield's repository of former useful R functions

Description Usage Arguments Details Value Examples

The runCPH function will process three separate file one with sample ids one with survival data and one with expression values. In test mode if a list of sample ids are supplied but not the other two files it assumes the ids are from the TCGA SKCM data set and uses a cached version of the normalised expression values from the TCGA SKCM data set and a currated set of survival data.

1
2
3

survClust(survData = NULL, exprData = NULL, geneData = NULL,
  survTime = "time", cox = F, plotIt = F, survStatus = "status",
  mcores = 4, clustCol = "sampleCl", verbose = FALSE)

`survData`	survival data must contain a time and status
`exprData`	normalised gene expression data
`geneData`	names of the genes to be tested (must be in the rownames of exprData)
`percExpr`	filter for minimum percentage of samples required to be expression a gene to select genes for testing
`exprRange`	top and bottom quantiles of expression values to select samples for testing
`sep`	default separator for delimited survival and sample id files supplied
`exprSep`	default separator for expression file supplied
`outputDir`	location to write output files
`analysisName`	stub to add to output file names
`silent`	boolean return the full set of statistics as a data frame

It will also perform a limma differential gene expression test for the significance of the fold change between the top and bottom n percent of expression values

the function creates several csv output files with the Cox PH results, the Limma DGE results, a combined set of both results and a set of those genes for which the DGE result is significant at 0.05 FDR

analysisName="test3"
exprSep=","
res <- runCPH(sampleDataFile=sampleDataFile,expressionDataFile=expressionDataFile,
              survivalDataFile=survivalDataFile,
              analysisName=analysisName,exprSep=exprSep)