plot_orthotree will seem to randomly drop some species (that are indeed available) when mc.cores>1, which parallelises certain steps with parallel::mclapply.
I'm not sure why this happens, but it does perhaps bring up some concerns about parallel::mclapply in general.