Getting Data Into R

The last block of Day 2 is a short demo of the bridge between QIIME2 and R: how to export the artifacts you've just produced into flat TSV / CSV files that R can ingest directly. The full export reference (every metric and every command you'll likely want for publication-quality figures) lives on the Publication Plots in R (Self-Study) page, work through that on your own afterward.

Why Export?

QIIME2 artifacts (.qza, .qzv) are zip archives with provenance and metadata wrapped around the underlying data. R can't read them directly, so we use qiime tools export to unwrap one and emit the inner table as a TSV.

Demo, Export an α-Diversity Vector

Pick one alpha metric you computed in Phase 4 of the core metrics step. Export Shannon as an example:

qiime tools export \
  --input-path core_metrics/shannon_vector.qza \
  --output-path output_for_R

mv output_for_R/alpha-diversity.tsv output_for_R/shannon_vector.tsv

The rename matters: every alpha export writes to alpha-diversity.tsv by default, so you have to rename right after export or each subsequent metric will overwrite the last.

Demo, Export a β-Diversity Distance Matrix

qiime tools export \
  --input-path core_metrics/bray_curtis_distance_matrix.qza \
  --output-path output_for_R

mv output_for_R/distance-matrix.tsv output_for_R/bray_curtis_distance_matrix.tsv

Same rename pattern.

Pulling Into R

Once you have the TSVs, load them in R as a normal tab-delimited file. For an alpha vector:

shannon <- read.table(
  "output_for_R/shannon_vector.tsv",
  header = TRUE, sep = "\t", check.names = FALSE
)

For a distance matrix:

bc <- as.matrix(read.table(
  "output_for_R/bray_curtis_distance_matrix.tsv",
  header = TRUE, row.names = 1, sep = "\t", check.names = FALSE
))

Join them to your metadata on the sample-ID column and you're set up for ggplot2, vegan, phyloseq, or whatever downstream package you prefer.

Going Further

The Publication Plots in R (Self-Study) page contains export commands for:

• All four α-diversity vectors (Shannon, Faith's PD, observed features, evenness)
• All four β-diversity distance matrices (Bray-Curtis, Jaccard, weighted & unweighted UniFrac)
• The taxonomy + feature table joined into a single tabulated visualization
• ANCOM-BC2 differential abundance results
• Random Forest feature importances

It also walks you through copying the workshop's visualizations.Rmd and R.Rproj into your working directory, knitting the report, and producing the figures from the lecture slides.

---

This completes Day 2. Continue to Day 3, Cow Dataset Pipeline.