ANCOM Interpretation

Hi,
My experiment involves analyzing 16S rRNA data including control and experimental groups involving pre-treatment and post-treatment observations.
Is there any consensus about what percentile (69, 70, 80, or 90) should I consider for analyzing this data type? At 90% the differential abundance is FALSE for all OTU’s while at 60% it is TRUE (at W= 244-292).

Thank you
Sanjay