SECOM for cross-domain correlations

[SergioAlias]

Hello,

Thank you so much for this fantastic methodology, I use ANCOMBC a lot in my work. I was wondering if the correlation methodology included in the package (SECOM) would allow for cross-domain correlations (e.g. bacteria and fungi).

My understanding here is that I could simply merge a 16S and ITS feature table into one multi-domain feature table (considering they include the same samples) and then run SECOM. But I don't know if I would be breaking any methodology assumption with that. Also, I assume that I would need to filter results to keep only bacteria-fungi correlations and discard bacteria-bacteria and fungi-fungi. But again, I'm unsure I'm breaking any assumption.

Thank you in advance

Sergio