Fig. 1

A scatter plot of the winner prevalence and winner diversity for different genera in the American Gut data. The winner prevalence is defined as the fraction of samples in which a winner OTU (an OTU occupying more than 90% of the group’s abundance) is observed. The winner diversity is defined as the Shannon diversity of the winner distribution (i.e., the frequency as which each OTU occurs as the lottery winner among all samples in which a lottery winner is observed). The winner diversity was normalized by the maximum diversity (log₂ of the number of winners, representing the Shannon diversity if all winners are observed at the same frequency) and hence range from 0 to 1. A low diversity suggests that the same OTU occurs as the lottery winner in all samples, while a high diversity suggests a more even frequency of different OTUs as lottery winners