Determining replication for discrimination among microbial communities in environmental samples using community-level physiological profiles

A statistical approach was employed to assess microbial community variability within a single 4-l water sample and among 10 independent 4-l water samples from a single location using community-level physiological profiles. Power calculations demonstrated that duplicate analyses could distinguish between two different locations at two times during the year. Variability associated with replicates from a single container or from different containers was nearly the same for the two sites examined. Duplicate assays (one from each of two independent samples) were sufficient to resolve the between-site differences at alpha=0.05 with power exceeding 0.95 at both times of the year. Similar methods are recommended to determine the appropriate number of replicates for environments of interest.