Vincent Carey 525-2265 <stvjc@channing.harvard.edu> writes: > > > > I am sorry to disturb with yet a simple question but I have tried my best and have now given up on solving it myself... :( > > > > I wonder if somebody knows how I easily can create mean values and their SD from subsets of columns (representing cases) over all rows individually (representing the genes) from an object - eset - (which is an Expression Set with 12625 genes, 7 samples). > > E.g the mean and SD from all rows individually in columns 4,5,6 and 7 of my data? > > > > Grateful for help! > > > > // Anna :o) > > mns <- apply(exprs(eset)[,4:7],2,mean) > sds <- apply(exprs(eset)[,4:7],2,function(x)sqrt(var(x))) Would mns <- rowMeans(exprs(eset)[,4:7]) work? If so, it is likely to be faster. (I'm not sure if exprs(eset) can be treated as a matrix.) Also, the function sd is equivalent to function(x)sqrt(var(x)) so the second could be written sds <- apply(exprs(eset)[,4:7],2,sd)

There are functions to compute row means or column means that are much faster than using apply(): > mns.col <- colMeans(exprs(eset)[,4:7]) > mns.row <- rowMeans(exprs(eset)[,4:7]) In S-Plus similar functions are available for variance (rowVars, colVars) and standard deviation (rowStdevs, colStdevs) but I didn't find them in R. Johannes Quoting Vincent Carey 525-2265 <stvjc@channing.harvard.edu>: > > > > I am sorry to disturb with yet a simple question but I have tried my > best and have now given up on solving it myself... :( > > > > I wonder if somebody knows how I easily can create mean values and > their SD from subsets of columns (representing cases) over all rows > individually (representing the genes) from an object - eset - (which is > an Expression Set with 12625 genes, 7 samples). > > E.g the mean and SD from all rows individually in columns 4,5,6 and 7 > of my data? > > > > Grateful for help! > > > > // Anna :o) > > mns <- apply(exprs(eset)[,4:7],2,mean) > sds <- apply(exprs(eset)[,4:7],2,function(x)sqrt(var(x))) > > _______________________________________________ > Bioconductor mailing list > Bioconductor@stat.math.ethz.ch > https://www.stat.math.ethz.ch/mailman/listinfo/bioconductor >