Is it okay to cluster rows and columns with a different linkage?

Question

[sjunlee89]

I have a question about the linkage methods of clustering.

In particular, if you look at the option of "complexheatmap" of the R package,

you can give different clustering linkage (e.g. ward, complete, single, etc.) for row and column.

I wonder if it is okay to give different linkage methods to each row and column when performing clustering analysis,

or how to explain the rationale for giving different linkage methods.

Thank you.