dnncha · dnncha · Jul 23, 2026 · Jul 23, 2026 · Jul 23, 2026 · Jul 23, 2026
diff --git a/.github/workflows/release.yml b/.github/workflows/release.yml
@@ -71,14 +71,19 @@ jobs:
       - uses: actions/setup-python@v6
         with:
           python-version: "3.11"
+      - name: Set up QEMU for Linux aarch64 wheel tests
+        uses: docker/setup-qemu-action@v4
+        with:
+          platforms: arm64
       - name: Build and test repaired Linux wheels
         uses: pypa/cibuildwheel@v4.1.0
         env:
           CIBW_PLATFORM: linux
+          CIBW_ARCHS_LINUX: "x86_64 aarch64"
         with:
           output-dir: dist-linux
       - name: Verify repaired Linux wheels
-        run: python scripts/check_python_wheel.py --wheel-only --out-dir dist-linux
+        run: python scripts/check_python_wheel.py --wheel-only --out-dir dist-linux --require-repaired-linux-architectures x86_64 aarch64
       - uses: actions/upload-artifact@v7
         with:
           name: dotmatch-linux-repaired-wheels
@@ -108,6 +113,10 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - uses: actions/checkout@v7
+      - name: Set up QEMU for Linux arm64 container tests
+        uses: docker/setup-qemu-action@v4
+        with:
+          platforms: arm64
       - uses: docker/setup-buildx-action@v4
       - name: Build local smoke-test image
         run: docker build -t dotmatch:ci .
@@ -118,6 +127,14 @@ jobs:
           docker run --rm dotmatch:ci dist ACGT AGGT | grep '^1$'
           docker run --rm dotmatch:ci leq 1 ACGT AGGT | grep '^true$'
           docker image inspect dotmatch:ci --format '{{ index .Config.Labels "org.opencontainers.image.version" }}' | grep "^${VERSION}$"
+      - name: Build Linux arm64 smoke-test image
+        run: docker buildx build --platform linux/arm64 --load -t dotmatch:ci-arm64 .
+      - name: Smoke test Linux arm64 container
+        run: |
+          VERSION=$(python -c "import tomllib; print(tomllib.load(open('pyproject.toml', 'rb'))['project']['version'])")
+          docker run --rm --platform linux/arm64 dotmatch:ci-arm64 --version | grep "^dotmatch ${VERSION}$"
+          docker run --rm --platform linux/arm64 dotmatch:ci-arm64 dist ACGT AGGT | grep '^1$'
+          docker run --rm --platform linux/arm64 dotmatch:ci-arm64 leq 1 ACGT AGGT | grep '^true$'
       - uses: docker/metadata-action@v6
         id: meta
         with:
@@ -143,8 +160,18 @@ jobs:
         with:
           context: .
           push: ${{ startsWith(github.ref, 'refs/tags/') }}
+          platforms: linux/amd64,linux/arm64
           tags: ${{ steps.meta.outputs.tags }}
           labels: ${{ steps.meta.outputs.labels }}
+      - name: Verify published multi-architecture manifest
+        if: startsWith(github.ref, 'refs/tags/')
+        env:
+          VERSION: ${{ github.ref_name }}
+        run: |
+          docker buildx imagetools inspect "ghcr.io/dnncha/dotmatch:${VERSION}" --raw > ghcr-manifest.json
+          python scripts/check_oci_manifest.py ghcr-manifest.json \
+            --require-platform linux/amd64 \
+            --require-platform linux/arm64
 
   pypi-sdist:
     name: Publish PyPI sdist, macOS wheel, and repaired Linux wheels

diff --git a/.zenodo.json b/.zenodo.json
@@ -1,7 +1,7 @@
 {
   "title": "DotMatch: deterministic known-target short-DNA assignment for sequencing workflows",
   "upload_type": "software",
-  "version": "0.2.2",
+  "version": "0.3.0",
   "conceptdoi": "10.5281/zenodo.20541628",
   "creators": [
     {
@@ -25,7 +25,7 @@
   ],
   "related_identifiers": [
     {
-      "identifier": "10.5281/zenodo.20541629",
+      "identifier": "10.5281/zenodo.21511337",
       "relation": "isNewVersionOf",
       "scheme": "doi",
       "resource_type": "software"

diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -2,6 +2,25 @@
 
 All notable user-facing changes are tracked here. Public statements in release notes must stay aligned with `docs/scientific-claims.md`.
 
+## 0.3.0 - 2026-07-23
+
+### Added
+
+- Added `dotmatch feature matrix` for deterministic cell-by-feature matrices
+  from pre-extracted observation tables. The command retains unique,
+  ambiguous, unmatched, and invalid observations in explicit output artifacts.
+- Added strict paired R1/R2 support to `dotmatch pair-count`, including
+  canonical read-name and record-count checks before counting.
+- Added a Galaxy XML wrapper and Planemo fixtures for CRISPR guide counting.
+  The wrapper is a local integration asset and does not indicate IUC
+  acceptance.
+
+### Changed
+
+- Configured tagged releases to build repaired Linux `x86_64` and `aarch64`
+  wheels and publish multi-architecture `linux/amd64` and `linux/arm64`
+  container manifests after their checks pass.
+
 ## 0.2.2 - 2026-07-23
 
 ### Changed

diff --git a/CITATION.cff b/CITATION.cff
@@ -8,8 +8,7 @@ authors:
     orcid: "https://orcid.org/0009-0003-5012-7229"
 repository-code: "https://github.com/dnncha/dotmatch"
 license: Apache-2.0
-version: "0.2.2"
-doi: 10.5281/zenodo.21511337
+version: "0.3.0"
 abstract: "DotMatch is a deterministic known-target short-DNA assignment engine for CRISPR guide counting, barcode demultiplexing, and fixed-target FASTQ workflows."
 keywords:
   - bioinformatics

diff --git a/DESCRIPTION b/DESCRIPTION
@@ -1,6 +1,6 @@
 Package: dotmatch
 Title: R Interface to DotMatch (Deterministic Short-DNA Assignment)
-Version: 0.2.2
+Version: 0.3.0
 Authors@R: 
     person("DotMatch Contributors", email = "donncha@example.com", role = c("aut", "cre"))
 Description: Provides R wrappers around the Python dotmatch package via

diff --git a/Dockerfile b/Dockerfile
@@ -5,7 +5,7 @@ LABEL org.opencontainers.image.title="DotMatch" \
       org.opencontainers.image.source="https://github.com/dnncha/dotmatch" \
       org.opencontainers.image.url="https://dotmatch.readthedocs.io/" \
       org.opencontainers.image.documentation="https://dotmatch.readthedocs.io/" \
-      org.opencontainers.image.version="0.2.2" \
+      org.opencontainers.image.version="0.3.0" \
       org.opencontainers.image.licenses="Apache-2.0" \
       org.opencontainers.image.authors="Donncha O'Toole"
 

diff --git a/README.md b/README.md
@@ -10,7 +10,7 @@ match, an ambiguous match, unmatched, or invalid.
 [![Documentation](https://readthedocs.org/projects/dotmatch/badge/?version=latest)](https://dotmatch.readthedocs.io/en/latest/)
 [![Bioconda](https://img.shields.io/conda/vn/bioconda/dotmatch?label=Bioconda)](https://anaconda.org/bioconda/dotmatch)
 [![License](https://img.shields.io/badge/license-Apache--2.0-blue.svg)](https://github.com/dnncha/dotmatch/blob/main/LICENSE)
-[![DOI](https://zenodo.org/badge/DOI/10.5281/zenodo.21511337.svg)](https://doi.org/10.5281/zenodo.21511337)
+[![DOI](https://zenodo.org/badge/DOI/10.5281/zenodo.20541628.svg)](https://doi.org/10.5281/zenodo.20541628)
 
 [Documentation](https://dotmatch.readthedocs.io/en/latest/) ·
 [Getting started](https://dotmatch.readthedocs.io/en/latest/getting-started.html) ·
@@ -148,6 +148,53 @@ dotmatch barcode autopsy \
 Open `autopsy/report.html` first. The tables beside it record offset scans,
 near-neighbour barcodes, correction safety, and frequent unmatched windows.
 
+### Build a cell-by-feature matrix from extracted observations
+
+When an upstream workflow has already extracted feature windows and attached an
+explicit cell identifier, DotMatch can write a sparse cells × features matrix:
+
+```bash
+dotmatch feature matrix \
+  --observations feature_observations.tsv \
+  --targets feature_library.tsv \
+  --id-column observation_id \
+  --cell-column cell_barcode \
+  --sequence-column feature_seq \
+  --metric hamming --k 1 \
+  --out-dir feature_matrix/
+```
+
+The output directory contains `matrix.mtx`, cell and feature axes, long-form
+counts, per-observation assignments, per-cell QC, and a JSON summary. Only
+unique assignments add a matrix count. This command does not perform FASTQ
+pairing, cell-barcode correction, UMI deduplication, or cell calling; those
+upstream steps should remain documented with the observation table.
+
+See the [scverse and feature-barcode tutorial](https://dotmatch.readthedocs.io/en/latest/tutorials/scverse-perturb-seq.html)
+for the file contract and AnnData handoff.
+
+### Count target pairs across R1 and R2
+
+Use `pair-count` when a left target and a right target are sequenced in
+synchronized FASTQ mates:
+
+```bash
+dotmatch pair-count \
+  --left-targets r1_targets.tsv \
+  --right-targets r2_targets.tsv \
+  --left-reads sample_R1.fastq.gz \
+  --right-reads sample_R2.fastq.gz \
+  --left-start 0 --left-length 20 \
+  --right-start 0 --right-length 20 \
+  --k 1 --metric hamming \
+  --out pair_counts.tsv \
+  --summary pair_summary.json
+```
+
+R1 and R2 must contain the same records in the same order. DotMatch checks the
+canonical read identifier before assignment and records input synchronization,
+side-specific unmatched totals, and side-specific invalid totals in the summary.
+
 ### Check a target library
 
 Before allowing mismatch correction, check whether neighbouring targets can

diff --git a/app/page.tsx b/app/page.tsx
@@ -14,7 +14,7 @@ const biocondaUrl = "https://anaconda.org/bioconda/dotmatch";
 const releaseUrl = `${repoUrl}/releases`;
 const containerUrl = `${repoUrl}/pkgs/container/dotmatch`;
 const workflowExamplesUrl = `${repoUrl}/tree/main/examples/workflows`;
-const doiUrl = "https://doi.org/10.5281/zenodo.21511337";
+const doiUrl = "https://doi.org/10.5281/zenodo.20541628";
 
 const basePath = process.env.NEXT_PUBLIC_BASE_PATH ?? "";
 const assignmentWorkflowImage = `${basePath}/dotmatch-read-assignment-v2.webp`;
@@ -37,7 +37,7 @@ const structuredData = {
       name: "DotMatch",
       applicationCategory: "Bioinformatics software",
       operatingSystem: "Linux, macOS",
-      softwareVersion: "0.2.2",
+      softwareVersion: "0.3.0",
       softwareHelp: docsUrl,
       codeRepository: repoUrl,
       downloadUrl: pypiUrl,

diff --git a/codemeta.json b/codemeta.json
@@ -9,8 +9,8 @@
   "issueTracker": "https://github.com/dnncha/dotmatch/issues",
   "license": "https://spdx.org/licenses/Apache-2.0",
   "identifier": "https://doi.org/10.5281/zenodo.20541628",
-  "version": "0.2.2",
-  "softwareVersion": "0.2.2",
+  "version": "0.3.0",
+  "softwareVersion": "0.3.0",
   "programmingLanguage": [
     "C",
     "Python",

diff --git a/docs/assayspec.md b/docs/assayspec.md
@@ -165,9 +165,25 @@ Demux mode uses `mode = "demux"`, `barcodes`, `reads`, `[extract]`, and writes
 `demuxed/`, `summary.json`, optional `assignments.tsv`, `ambiguous.fastq`, and
 `unmatched.fastq`.
 
-Pair mode uses `mode = "pair-count"`, `left_targets`, `right_targets`, `reads`,
-`[left]`, and `[right]`. It writes `pair_counts.tsv`, `pair_summary.json`, and
-optional `pair_assignments.tsv`.
+Pair mode uses `mode = "pair-count"`, `left_targets`, `right_targets`, `[left]`, and `[right]`.
+It accepts one of two input layouts:
+
+```toml
+# Both target windows occur in one read.
+reads = "reads.fastq.gz"
+
+# Or, remove reads and use synchronized mates.
+left_reads = "sample_R1.fastq.gz"
+right_reads = "sample_R2.fastq.gz"
+```
+
+For paired inputs, left extraction coordinates apply to `left_reads` and right
+coordinates apply to `right_reads`. The two files must have the same number of
+complete records in matching order. DotMatch compares canonical read IDs and
+stops on the first mismatch. It writes `pair_counts.tsv`, `pair_summary.json`, and
+optional `pair_assignments.tsv`. The summary records the input layout and
+side-specific unmatched and invalid outcomes; methods and assay reports list
+both FASTQ inputs.
 
 ## Safety Policy
 

diff --git a/docs/command-reference.md b/docs/command-reference.md
@@ -46,6 +46,54 @@ dotmatch audit --targets guides.tsv --k 3 --audit-mode exact --out-dir audit/
 Proceed only when `safe_at_hamming_k2` or `safe_at_hamming_k3` is true for the
 radius you plan to use.
 
+## Pre-extracted Feature Matrices
+
+```bash
+dotmatch feature matrix \
+  --observations feature_observations.tsv \
+  --targets feature_library.tsv \
+  --id-column observation_id \
+  --cell-column cell_barcode \
+  --sequence-column feature_seq \
+  --metric hamming --k 1 --ambiguity-policy radius \
+  --out-dir feature_matrix/
+```
+
+Use `feature matrix` when an upstream workflow has already made a headered
+observation table with an explicit cell identifier and a feature sequence
+window. It writes a deterministic sparse Matrix Market matrix with cells on
+rows and features on columns, plus feature/cell axes, long-form counts,
+per-observation assignments, per-cell QC, and a JSON summary.
+
+Only unique assignments add matrix counts. The command does not pair FASTQ
+reads, correct cell barcodes, deduplicate UMIs, or call cells; retain those
+upstream decisions and provenance with the input table.
+
+## Pair Counting Across Paired FASTQs
+
+`pair-count` assigns a left target and a right target for each record pair.
+Use `--reads` when both windows are present in one read, or give synchronized
+R1 and R2 files:
+
+```bash
+dotmatch pair-count \
+  --left-targets r1_targets.tsv \
+  --right-targets r2_targets.tsv \
+  --left-reads sample_R1.fastq.gz \
+  --right-reads sample_R2.fastq.gz \
+  --left-start 0 --left-length 20 \
+  --right-start 0 --right-length 20 \
+  --k 1 --metric hamming \
+  --out pair_counts.tsv \
+  --summary pair_summary.json \
+  --assignments pair_assignments.tsv
+```
+
+Paired inputs must contain the same number of complete FASTQ records in the
+same order. DotMatch compares the first header token after removing a terminal
+`/1` or `/2`; a mismatch stops the command before counts are written. The
+summary records the input mode and side-specific unmatched and invalid totals.
+
 ## AssaySpec Workflows
 
 ```bash

diff --git a/docs/conf.py b/docs/conf.py
@@ -8,7 +8,7 @@
 project = "DotMatch"
 author = "Donncha O'Toole"
 copyright = "2026, Donncha O'Toole"
-release = "0.2.2"
+release = "0.3.0"
 version = release
 
 extensions = [