Documentation

Reports and AI interpretation

What the web report, PDF and data package contain for each analysis, and how the AI-written interpretation is produced.

Updated

Report formats

16S report

SectionContents
Read trackingReads at each stage, from input through primer removal, filtering, merging and chimera removal, and the number of ASVs
Taxonomic compositionPhylum, genus and species composition, as charts and tables
Alpha diversityThe metrics described in the amplicon chapter, with a rarefaction curve
Phylogenetic treeAn interactive tree of the ASVs
Functional predictionThe top predicted MetaCyc pathways, labelled as predicted
Quality assessmentThe 0 to 100 score, the verdict and each individual check
MethodsPrimers, truncation lengths and how they were chosen, DADA2 settings, the reference database and tool versions

Project reports add beta diversity (PCoA), PERMANOVA and ANCOM-BC2 differential abundance between the groups in your metadata, and the AI interpretation of the study. A single sample's report has no AI interpretation.

AI interpretation

Project reports include a short written interpretation of each section and an overall summary, produced after the analysis finishes. Reports on a single sample have none. It is written from the study's statistics: the diversity tests, PERMANOVA, differential abundance and the samples left out. It is generated by Google's Gemini API from summary results (taxonomic composition, diversity, quality metrics and study statistics). Sequencing reads are never sent.

A starting point, not a conclusion. The interpretation can be wrong. Check it against your data and the literature before you use it in a publication, and never use it for clinical decisions.

Data package contents