Rational on how to chose between clustering methods ?

[PFRoux]

Dear Henry,

First of all, I would like to sincerely thank you for developing and maintaining this fantastic package.
I am using it in almost all my research projects for almost 10 years now :-)

But nevertheless, I am still struggling at picking the "best" clustering method(s) between all the options available to define gene modules (D-cluster, correlation, over expression, under expression ...).

Could you please comment on that ? And maybe provide ressources / references describing each of these methods when it comes to SOM ? I must admit that - despite reading all your work - I am especially confused regarding the "over expression" et "under expression" methods : over and under expression with respect to what / which condition ? And what is the difference between those and the "group over expression" method ?

I warmly thank you for this awesome package as well as for your help.

Sincerely,

Pef