I understand that K-means clustering is used very often for gene expression analysis and usually dissimilarity is measured by euclidean distance but are there any particular applications of in which euclidean distance may not be the most appropriate tool in clustering? Is there any other way dissimilarity in gene expression be measured?