Hi everyone,

I am analyzing the expression profile of about 12,000 genes, cluster analysis using  fuzzy C-means by using Mfuzz package in R. However, whether I use a C value of 100 or 50, for each cluster, I only got very few genes with a membership value higher than 0.7, say only 20 or 40 genes. There are still far too much genes are not in the cluster.  Has anyone ever met this kind of question before? What should I do to get more genes with high membership value in each cluster?

Thanks in advance.

Hi Emily,

How many clusters do you expect? Did you do PCA or other clustering methods to see how many clusters you may expect? I mean why did you take c = 100 or 50?

Ben