Category: QIIME2

RDP Classifier Updated Yet Again!

Author JQPosted on Categories classifying, QIIME2Leave a comment on RDP Classifier Updated Yet Again!

Time marches on, and Qiong Wang is in step with it. She has updated the RDP Classifer yet again, this time to version 2.15, and released it on SourceForge November 2026. This version is trained on  the RDP Bacteria and Archaea Taxonomy training set No. 21 which for the first time includes validated names of uncultured prokaryotes based on genome sequences from the SeqCode registry up to August 2026. The database now includes 23,850 species and 5,058 genera, increases of 25% and 30% over training set version 19 released in 2023. Seven percent of the sequences now come from SeqCode. The training data are available in both RDP and QIIME2 format.

For further details, see the release notes.

```

QIIME2 DADA2 Updated

Author JQPosted on Categories DADA2, QIIME2, RLeave a comment on QIIME2 DADA2 Updated

First appearing in the 2025.10 version of QIIME2, there are important changes to the DADA2 plugin. Two parameters available in the R version of DADA2 that greatly simplify correctly processing variable length amplicon data have been exposed. One is the capability of trimming overhangs using –p-trim-overhangs. The second, the –p-max-merge-mismatch parameter, allows mismatches in the overlap region when merging pairs. By default, DADA2 allows no mismatches in the overlap region, but with variable length amplicons when the overlap length can exceed 100 bp this is simply too strict.

There is also a new mandatory output, the base-transition-stats. Output as a qza file, they can be exported as large tsv file, but to make the stats more easily interpretable I have added the function plot_transition_stats() to QsRutils. The objective is to check how well DADA2 is modeling transition errors and correcting sequences before merging them. As an example:

QsRutils::plot_transition_stats(base-transition-stats.qza)

```