I followed the Mothur SOP up to calculation of the unifrac distances.
The SOP (https://www.mothur.org/wiki/MiSeq_SOP#Beta_diversity)
states to use
>mothur unifrac.unweighted(tree=stability.tre, count=stability.count_table, distance=lt, processors=2, random=F, subsample=2392)
Unfortunatly no stability.tre file was generated in the SOP. During the SOP two trees are generated one is stability.trim.contigs.good.unique.good.filter.unique.precluster.pick.pick.pick.phylip.tre as the phylogenetic tree relating sequences to eac hother and another one (stability.opti_mcc.thetayc.0.03.lt.ave.tre ) related the samples to each other. The command is only working when using the first tree.But since the nodes are not labelled with the OTU names, I'm still wondering if this is correct. In the SOP for both trees it is written, that they can be used, but unfortunately not when to use which.