I suspect they could do better than random chance in guessing the gender of the sample source. Their tables show pretty big differences in the categories Actin

I suspect they could do better than random chance in guessing the gender of the sample source. Their tables show pretty big differences in the categories Actinobacteria (73% more in men) and Verrucomicrobia (88% more in women). Maybe their sample is so small even these very different averages are just attributable to random chance, but it seems they've made a conscious choice not to reveal the size of their database.