Hei, I have been trying my hands on coinertia analysis (CIA) for comparing two data sets containing different genesets for same samples. The actual RV coefficient obtained is a measure of global similarity between the datasets and has a range of [0, 1] . Is there a built in (randomization) test for checking the significance of the RV coefficient. I couldn't find one in the made package, perhaps ade4 (not bioconductor supported) has something. If not, I would like to know which pf the following approaches would be valid: 1) Given two datasets A(mxn) and B(pxn), where m, p are the rows (genes) and n the cols. (samples). 2a) Hold A constant, and randomly scramble the elements of B (r times) and then judge the tail prob. of the obtained RV coefficient against the CIA obtained on the randomized sets of B. or alternatively 2b) Hold A constant, and take r different gene sets from the same data set of the same size as B, and then judge the tail prob. of the obtained RV coefficient against the CIA obtained on the different sets of B. I haven't actually tried this at the moment, but should there be much difference between alts. (2a) and (2b)? Any help or pointers would be appreciated. Narinder PS!! the actual code for CIA is as follows: coin <- cia(Xscaled.A, Xscaled.B) c.1 <- coin$coinertia$RV;