Release 0.3.0 + 0.3.1 (validated via ICA full-pipeline runs) by qclayssen · Pull Request #35 · umccr/bolt

qclayssen · 2026-07-13T05:39:06Z

Summary

Split out of Release 0.3.2 (pending full-pipeline validation) #33 to isolate the already-validated portion of the release from unvalidated 0.3.2 work.
Contains everything through the v0.3.1 tag, matching the ghcr.io/umccr/bolt:0.3.1* images already run against real ICA data.
See CHANGELOG.md ## 0.3.1 section for the full list of fixes.

Validation

Full-pipeline runs tracked in umccr/biodaily#199 (sash 0.7.0, bolt 0.3.1 images) across 11 samples.

Test plan

CI unit tests pass on this branch
Docker images for v0.3.1 already built and validated via ICA runs

…11) * add params input for sigrap tools

* update "HIGH_INF_POS" to "HIGH_I_I_POS" in VCF header for latest pcgr version * update constant for pcgr 2.2.5

* change sage vcf header to match 2024 version * move check function * add check in rescue for sage vcf * change for consistencency with pcgr

…arallelisation because OOM hypermutated samples

…'t use more core and so more memory

…e_command. Fixes #26 - Commands now fail immediately instead of continuing silently

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

….2.5 tabix check)

# Conflicts: # .bumpversion.cfg # docker/Dockerfile.gpgr # pyproject.toml

scwatts

Nice work, read over the code briefly and made a couple of minor comments, trust you've tested well :)

scwatts · 2026-07-21T01:01:57Z

 ## Variation selection (annotation) ##
 ######################################
-MAX_SOMATIC_VARIANTS = 500_000
+MAX_SOMATIC_VARIANTS = 450_000


Do we have a Slack thread (or similar) to link to for this change?

I believe this was an oral discussion, since PCGR's hard cap is 500k, we set 450k to avoid being right on the limit and allow margin for edge cases.

I updated ADR document for sash and bolt explaining rational

run_somatic_chunk -> merging_pcgr_files unconditionally called util.merge_vcf_files, which shells out to `bcftools merge`. That command requires 2+ inputs, so a sample whose variants fit in a single chunk (no splitting needed) produced a bcftools merge Usage error instead of a valid annotated VCF. Guard on len(pcgr_vcf_files) == 1: copy the single chunk's VCF directly to the expected merged-output path and tabix-index it, matching the indexed .vcf.gz contract that merge_vcf_files normally produces. Two or more chunks still go through the existing bcftools merge -> sort -> index path unchanged.

- Remove stray blank first line in bolt/common/pcgr.py - Reword comment in report.py to reference MAX_SOMATIC_VARIANTS constant instead of hardcoded '500,000' value

Explain why the threshold is 450k (not 500k): PCGR's hard cap is 500k, the 50k margin avoids edge-case overflows during tiered filtering. Notes this was an oral team decision with no written Slack thread. Addresses Stephen's review comment on constants.py:7.

All 11 conflicts resolved by taking release/0.3.0 (0.3.2) content: - Version bumps (.bumpversion.cfg, pyproject.toml, constants.py): keep 0.3.2 - CHANGELOG.md: keep full 0.3.2 history - build.yaml: keep pinned action version (v1.3.0) - logging_config.py: keep force=True addition - pcgr.py, util.py, annotate.py, report.py, rescue.py: keep 0.3.2 implementations Main only had the 0.3.1-validated subset (merged via #35); release/0.3.0 is strictly ahead with all 0.3.2 development work.

Conflicts resolved by taking oa-only-v2 content (superset of main): - CHANGELOG.md, constants.py, pcgr.py, report.py, test_pcgr_hypermutated.py

…nts) Merge resolution strategy: - Version/config (.bumpversion.cfg, pyproject.toml, build.yaml, CHANGELOG): take ours (0.3.2) - constants.py: take ours + add MAX_SOMATIC_VARIANTS rationale comment from #35 review - pcgr.py: take ours (has _normalise_tier, drops --estimate_signatures per sash#57) + incorporate documentation comments from #35 review - util.py: take ours + add logging setup comment - logging_config.py: take ours (force=True fix) - annotate.py, report.py, rescue.py: take ours + add review comments - docs/adr/001-max-somatic-variants-450k.md: new from main (kept) - tests/test_pcgr_hypermutated.py: new from main (kept)

qclayssen and others added 30 commits October 2, 2025 14:17

Add option for sash sigrap individual process input to cancer report (#…

d7ca46b

…11) * add params input for sigrap tools

Hypermutation Handling in PCGR (#9)

f33c48c

Feature/pcgr v2.2.1 (#10)

cb5ee4d

fix merge duplication and imports

9d8e4d1

bump pcgr version

8db35b3

bump gpgr version

c67b30a

change print to logging

f9a7d3b

add argument chunck size

b8ee4a0

fix version software pcgr container

8223ef9

add logging and filter fonction to read value and replace placeholder

f83c23a

fix indentation

22b5c12

read gz tsv from pcgr

529309c

change PCGR_CLINVAR_CLASSIFICATION

d3dd182

Change pcgr header check

a5fc923

tyo

bf930e7

typo

8ea1edc

add missing variable pcgr_output_dir

443c5be

Pcgr v2.2.5 (#22)

966967f

* update "HIGH_INF_POS" to "HIGH_I_I_POS" in VCF header for latest pcgr version * update constant for pcgr 2.2.5

Uptade sage vcf header check (#24)

6a93f6b

* change sage vcf header to match 2024 version * move check function * add check in rescue for sage vcf * change for consistencency with pcgr

add missing HMF_HOTSPOT constant

ed3b11b

Fix quotation

a3d16f7

Fix test

9ddeac8

remove PCGR html report from chunk (hypermutated) to reduce memory use

41b83bb

change remove pcgr parallelisation of chunks, and fix pcgr internal p…

c42d725

…arallelisation because OOM hypermutated samples

free disk worker

03bba1d

remove exception to stop at fail process

cb1c39b

make thread you by pcgr constant, so in case of OOM, on the retry don…

27403cb

…'t use more core and so more memory

Fix silent command failures by integrating error handling into execut…

2902428

…e_command. Fixes #26 - Commands now fail immediately instead of continuing silently

Extract PCGR errors from log files

15db538

Update bolt/workflows/other/cancer_report.py

9bf22f5

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

qclayssen added 5 commits May 19, 2026 11:27

Fix remove unused package and log pcgr limit

c53557a

fix: tabix-index VCF chunks after split_vcf so PCGR can read them

8a35851

test: assert split_vcf chunks have .tbi index (regression for PCGR v2…

cfac848

….2.5 tabix check)

fix: bump r-gpgr to 2.3.1 in gpgr Docker image

d2c390f

Merge remote-tracking branch 'origin/main' into release/0.3.1-validated

ef00deb

# Conflicts: # .bumpversion.cfg # docker/Dockerfile.gpgr # pyproject.toml

This was referenced Jul 17, 2026

Release/0.7.0 umccr/sash#39

Merged

Release 0.3.2 #37

Closed

Release 0.3.2 (pending full-pipeline validation) #33

Open

qclayssen changed the title ~~Release 0.3.1 (validated via ICA full-pipeline runs)~~ Release 0.3.0 + 0.3.1 (validated via ICA full-pipeline runs) Jul 17, 2026

qclayssen requested a review from scwatts July 17, 2026 04:06

scwatts approved these changes Jul 21, 2026

View reviewed changes

qclayssen added 2 commits July 21, 2026 12:36

style: address Stephen's review comments (bolt #35)

afc46a5

- Remove stray blank first line in bolt/common/pcgr.py - Reword comment in report.py to reference MAX_SOMATIC_VARIANTS constant instead of hardcoded '500,000' value

qclayssen force-pushed the release/0.3.1-validated branch from 8f2d975 to b463534 Compare July 21, 2026 23:22

qclayssen force-pushed the release/0.3.1-validated branch from b463534 to 8f264d9 Compare July 21, 2026 23:34

qclayssen force-pushed the release/0.3.1-validated branch from 8f264d9 to 1a2a5ca Compare July 21, 2026 23:46

qclayssen merged commit 99c4a7f into main Jul 22, 2026
5 checks passed

qclayssen deleted the release/0.3.1-validated branch July 22, 2026 00:15

qclayssen added a commit that referenced this pull request Jul 22, 2026

fix: merge main into oa-only-v2 (resolve conflicts post-#35 merge)

a89dbca

Conflicts resolved by taking oa-only-v2 content (superset of main): - CHANGELOG.md, constants.py, pcgr.py, report.py, test_pcgr_hypermutated.py

Navigation Menu

Uh oh!

Release 0.3.0 + 0.3.1 (validated via ICA full-pipeline runs) - #35

Release 0.3.0 + 0.3.1 (validated via ICA full-pipeline runs)#35
qclayssen merged 63 commits into
mainfrom
release/0.3.1-validated

qclayssen commented Jul 13, 2026

Uh oh!

scwatts left a comment

Uh oh!

scwatts Jul 21, 2026

Uh oh!

qclayssen Jul 21, 2026

Uh oh!

qclayssen Jul 21, 2026

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Conversation

qclayssen commented Jul 13, 2026

Summary

Validation

Test plan

Uh oh!

scwatts left a comment

Choose a reason for hiding this comment

Uh oh!

scwatts Jul 21, 2026

Choose a reason for hiding this comment

Uh oh!

qclayssen Jul 21, 2026

Choose a reason for hiding this comment

Uh oh!

qclayssen Jul 21, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants