Hello,
Thank you for this wonderful tool!
I would like to ask about big clusters that contain both positively and negatively correlated features… How do we interpret these mixed blocks of associations? Do they have a biological meaning? Or is it merely a consequence of hierarchical clustering and adjusted p-values in block formation?
Thank you in advance!
The interpretation and biological meaning of hypothesis blocks is entirely context dependent, so it’s difficult to talk about generally without a specific example to refer to. But blocks that are mixtures of positive and negative correlations aren’t any more or less plausible than blocks with a single directionality. As a hypothetical example, say you’re comparing a dataset of microbial abundances to a dataset of metabolites. You might see a set of closely related bacteria (that correlate due to their similar metabolic functional potential) positively associated with some substrate metabolite they feed on and negatively associated with the waste product of that metabolism.
Thank you very much @andrewGhazi , your answer was clarifying for me. I am working exactly with microbial abundances and metabolites, so it was a really adequate example… Thank you again!