I'm using "pvclust" (in R) to determine the significance of clustering of 29 proteins (measured in 380 individuals). It runs smoothly, and I can get pretty pictures. Now I'm wondering, since there seem to be three larger clusters: can you set at which level pvclust should do the significance clustering? I'm not so much interested in the significance of clustering at the paired-protein level (which it almost does now), rather at a higher level.
Or should I revert to another package? Or a reverse approach, i.e. not agglomerative clustering, rather supervised clustering?