Thank you for developing the Q2 plugin for picrust2!
Could you add a section to Q2-picrust2 library on warnings/quality control? There's a quality control section in picrust1 documentation. Could you give some recommendations on QC using picrust2, especially for environmental samples?
I'm working on a manuscript for PICRUSt2, which will give better insight into how different maximum NSTI cut-offs affect PICRUSt2 performance across different environments. Predictions for environmental data tend to be less reliable since they are poorly represented by available reference genomes. Testing for large-scale shifts in functional potential (e.g. using ordination approaches) is especially useful in these contexts. In contrast, identifying individual biomarkers in these environments based on PICRUSt2 predictions alone requires corroborating evidence to be convincing. I'll be sure to point to the manuscript on the PICRUSt google-group when it is available as a pre-print and I plan to add additional pages to the wiki over time with more details.