Hi Siouxsie, This came up in review - it was a consequence of rarefaction. We actually sampled more players, but after rarefying the dataset (so that all samples have the same number of observations) some of the SV team member samples were too depauperate to be included. This pretty much always happens, to some extent, with high-throughput amplicon datasets; some samples have too few sequences to include in downstream analyses. In this case, SV took the biggest hit. The resulting dataset was our attempt to balance high sequence quality and read depth with adequate n for meaningful analysis.