st: Clustering variables instead of cases

[Alexey Shpenev <shpenevstatalist@gmail.com>]

I am performing some analyses on biomedical data. I have a set of patients with a lot of components in their blood measured and quantified. I am not interested in classifying those components.
I know there's a user-written command clv for such purposes, but I am not sure if anything better exists in stata or in some outsource packages. 
What I need is not only the results of clustering, but also some criteria on how to choose the best number of clusters and how to test those clusters for validity. I would also like to know if there are some measures that allow to test how likely each single variable is correctly classified
I have seldom used cluster or factor analysis before, so please forgive if I ask for something obvious.
Maybe you can just suggest some stata commands and I will read the help files.

Thanks a lot,
Alexey


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/