Hi, you don’t need to subsample your metagenomes if their size is comparable, in case you have a big difference in size (around one order of magnitude), I would consider subsampling.
There are also recent threads on this topic, I’ll leave here the links in case you want to have a look (#283 and #320)