Combining bac120 and ar53 de novo trees for weighted UniFrac across both domains

I have run de_novo_wf separately on my bacterial and archaeal MAGs, as required, since the bac120 and ar53 marker sets are disjoint and use independent reference trees.

For a downstream analysis such as weighted UniFrac, however, I would need a single phylogenetic tree spanning all genomes.

Would it be defensible to graft/stitch the bacterial and archaeal de novo trees together at an arbitrary root or outgroup branch? Or is there a preferred universal single-copy marker set that can be used to build a single phylogeny containing both bacterial and archaeal MAGs directly?

What approach would be recommended for generating a combined tree suitable for phylogeny-based beta-diversity analyses such as weighted UniFrac?