[Preview] QIIME 2 2026.1 development changelog

Developer Discussion Developer Announcements
,
1

QIIME 2 2026.1 has been released!
Please see the official changelog for more details.

Original content of post

:exclamation: Important :exclamation:

The following is an early developer preview of the changes expected in 2026.1

This post is a live-document which will be updated throughout our development cycle. Any links will in this topic will be broken until the release is officially published. When we are ready for release, we’ll copy this changelog and create a new post in the Announcements category.

Developer Project Board:

QIIME 2 2026.1 Project on Github

:police_car_light: Important Announcements :police_car_light:

Distribution Updates :package:

We will be renaming the amplicon distribution to qiime2, since this is the historical collection of packages that our user base is familiar with in the context of the qiime2 namespace. Additionally, we will be trimming some plugins from the moshpit and pathogenome distributions such that most plugins with actions downstream of a feature table will be contained within the qiime2 distribution.

Environment File Updates :file_folder:

MACOS Environment Files :green_apple:

Starting sometime in the upcoming development cycle (leading up to the 2026.4 release), we will begin providing two sets of environment files for our Mac users: osx-64 and arm-64. This is intended to provide the beginning of a runway towards deprecating support for osx-64 (spurred on by Github Actions no longer providing intel runners for macos after August of 2027).

What this means for 2026.4: If you are using a Mac with a silicon chip, you will be able to utilize environment files specifically built for arm-64 and will no longer need to provide the CONDA_SUBDIR when creating a new environment :partying_face: If you are using a Mac with an Intel chip, you will use the environment files that are specified for macos-intel (but this won't change anything about your current workflow).

Distribution-Specific Environment Files :bookmark_tabs:

We will no longer be building macos environment files for the moshpit and pathogenome distributions after 2026.1, as there are a significant number of plugins/actions that require too much compute power to run successfully on a Mac. If you are currently using either of these distributions on a Mac and are planning to update your installed version in 2026.4, our recommendation is to either utilize your institution's HPC or to install via Docker.

:bangbang: Breaking Changes :bangbang:

  • q2-feature-table
    • As was announced in the 2025.10 changelog, @Oddant1 replaced the old summarize visualizer with the summarize_plus pipeline. The old summarize visualizer has been renamed to _summarize and the old summarize_plus pipeline has been renamed to summarize. This strictly increases the functionality of summarize, and adds several new additional outputs that any workflows using summarize will need to account for. :snake:

:sparkles: Here are the Highlights of the Release! :sparkles:

Framework (i.e., rachis) Updates :grapes:

  • As promised in the recent announcement, the qiime2 framework has been renamed to rachis. It can be found at rachis-org/rachis on Github. @ebolyen wrote a backwards compatibility shim so that imports of qiime2 still function as expected.
  • @Oddant1 fixed a bug in the cache causing a permission error to be raised during garbage collection if garbage collection lacked permission to remove a file. :wastebasket:
  • @Oddant1 made Cache.get_keys and Cache.get_pools return sorted lists instead of sets. :package:
  • @Oddant1 added the CaptureHolder view type, allowing actions to capture the value set on random parameters in provenance. :magnifying_glass_tilted_left:
  • @Oddant1 made it so all provenance replay functionality works natively with Results objects, meaning that using the replay methods in the Python API will now be easier because Artifacts can be loaded and handed into them instead of fiddling with zip files directly. :zipper_mouth_face:
  • @Oddant1 and @colinvwood fixed a bug preventing the use of pronenance replay on provenance containing the use of elements of ResultCollections as Metadataparameters. :beetle:

Interface Updates :tv:

  • q2cli
    • @Macabe222 added a RachisWarning warning that will always be visible in the command line interface, regardless of whether --verbose is set. :warning:
    • @Oddant1 made it so qiime tools cache status will show cache items in alphabetical order where it was previously random. :game_die:

Plugin Updates :electric_plug:

  • q2-boots
    • @felicia added support for protein sequences to mafft, mafft-add, and mask. :poultry_leg:
    • @felicia exposed the underlying alignment strategies in the mafft tool to QIIME 2. Where previously we only used the default mafft algorithm. :straight_ruler:
  • q2-boots
    • @Oddant1 removed the memory bottleneck in the medoid calculation used in averaging distance matrices such that memory scales linearly now (where previously it scaled quadratically). This allows for larger values of n than had previously been possible, and as a result medoid is now the default averaging method for distance matrices. :1234:
    • @colinvwood fixed a :bug: in the demux-* actions caused by an interaction between batch_size and cut that was causing cuts to get erroneously applied multiple times when two or more batches were provided (via the combination of forward_cut, reverse_cut and batch_size). Thanks again for reporting this @ppflrs :raising_hands:t3:
  • q2-feature-table
    • @Macabe222 added a warning in the merge-taxa action to alert users when taxonomies of differing depths are merged. :evergreen_tree:
    • @Oddant1 replaced the old summarize visualizer with the summarize_plus pipeline. The old summarize method has been renamed to _summarize and the old summarize_plus pipeline has been renamed to summarize. This strictly increases the functionality of summarize
  • q2-types
    • @cherman2 and @gregcaporaso fixed a bug that caused paired-end read data and demux artifacts to fail record count validation (i.e. forward and reverse files for the same sample should have the same number of records) if the data contained "overlapping ids". (example: Validation would fail because sample 5 and sample 51 have a different number of records). :bug:
    • @Macabe222 added warnings that will appear when importing taxonomies that a) have a depth of one or b) have a trailing semicolon in one or more taxonomic classification strings. :warning:
    • @Vinzent_Risch added partition actions for GenomeData[Genes], GenomeData[Proteins], and GenomeData[Loci]. :dna:
    • @misialq added a new format: AnnotationToContigs and two new types: FeatureMap[TaxonomyToContigs] & FeatureMap[FunctionToContigs] to enable storing the mapping between annotations (like taxonomy) and contigs with the annotation.
1 Like