Dear all, I am using heatmap.2 to create a hierarchical clustering heatmap of some gene expression data. I am wondering how the order of the rows and columns affects the resulting dendogram/heatmap. I have seen that different order of rows gives different heatmaps. Is it valid to order the rows or columns in an order that seem logic to the study before I do the hierarchical clustering or should I rather randomize the order? As I understand, the hierarchical clustering will break the original order if it "violates" the clusters. Thank you! Best regards, Lina Hultin Rosenberg ________________________________ Lina Hultin Rosenberg Msc Molecular Biotechnology Evolutionary Biology Department Uppsala University Norbyv?gen 18 752 36 Uppsala Phone: +46-18-4716444 Email: lina.hultin.rosenberg at ebc.uu.se