Harako-RNAseq v0.3.0-beta.2

Harako-RNAseq v0.3.0-beta.2

Release date: 2026-08-15

Summary

Harako-RNAseq v0.3.0-beta.2 corrects the scientific handoff between tximport
and DESeq2 for full-length RNA-seq and introduces explicit selection of the
library protocol. The existing differential and QC-only modes, checksum-pinned
Ensembl references, and public output paths remain in place.

Scientific correction

v0.3.0-beta.1 wrote original tximport gene-level estimated counts to a matrix
and passed that matrix to DESeq2 without the tximport-derived effective-length
offset. This could affect full-length RNA-seq when average effective gene
length differs across samples, including cases with differential isoform
usage. It does not mean that every beta.1 result was incorrect, and the size
or direction of any effect cannot be stated without reanalysis.

Beta.2 retains the complete tximport object in tximport/txi.rds and constructs
full-length analyses with DESeqDataSetFromTximport, allowing the
sample-specific effective-length correction to reach DESeq2. For 3′-tag data,
Harako uses the original estimated counts without transcript-length
correction. TPM is an abundance output and is never used as DESeq2 model input.

Reanalysis recommendation

Users who analyzed full-length RNA-seq with v0.3.0-beta.1 should rerun the
analysis with v0.3.0-beta.2 or later. 3′-tag protocols do not ordinarily
require transcript-length correction. Because beta.1 did not record an
explicit library protocol, users who are uncertain about their protocol should
verify it before deciding whether rerun is required. Legacy frozen runs remain
readable but are not silently reinterpreted.

New protocol selection

New runs must explicitly select full_length or three_prime_tag; Harako does
not infer the protocol from filenames, accessions, read length, library names,
platform, or sample metadata. The selection is frozen in the run configuration
and is execution-relevant in agent plan IDs and approval hashes.

Compatibility

Existing gene-level counts TSV, gene-level TPM TSV, library-size QC, DESeq2,
and report paths remain unchanged. tximport/txi.rds is the only new internal
handoff artifact. Old frozen runs remain readable. Old schema-v1 agent plans
retain their historical plan IDs and approval hashes, but must be regenerated
with an explicit protocol before dry-run or execution.

nf-core/rnaseq scope

Harako is a narrower, bilingual, GUI-first local workflow and is not intended
to replace nf-core/rnaseq. See the concise
scope comparison.

Container and verification

The prerelease publication workflow is prepared to publish:

• ghcr.io/do-shima/harako-rnaseq:v0.3.0-beta.2
• ghcr.io/do-shima/harako-rnaseq:beta

No latest tag is published for this prerelease. Image availability begins only after
the tag-driven publication workflow succeeds. The final digest is
recorded only after publication; no digest is asserted in these notes. The
workflow generates SBOM, BuildKit provenance, and GitHub attestation evidence
for the published image. Maintainers verify it with gh attestation verify
against the immutable digest.

The candidate and published image target linux/amd64. Harako-RNAseq source is
available under the PolyForm Noncommercial License 1.0.0; bundled tools retain
their own licenses.

Publication evidence

• Exact image: ghcr.io/do-shima/harako-rnaseq:v0.3.0-beta.2
• Moving beta channel: ghcr.io/do-shima/harako-rnaseq:beta
• Published digest: sha256:12461cc98ce4a55c973803b0adea7e17f1354ee8929151a22f1e6c32e2b7c287
• Platform: linux/amd64
• GHCR anonymous pull: passed
• SBOM and BuildKit provenance: published by the tag-driven workflow
• GitHub artifact attestation: verified
• Publication workflow: completed successfully
• Publication workflow URL: https://github.com/do-shima/harako-rnaseq/actions/runs/31875458992

No latest tag is published for this prerelease.

Attestation verification command:

gh attestation verify oci://ghcr.io/do-shima/harako-rnaseq:v0.3.0-beta.2 -R do-shima/harako-rnaseq