fig5-sweepga-f16-validated-chop-sensitivity — octopus01:/moosefs/erikg/phrs

Metadata

Status	open ‖ paused
Agent identity	`46f6237a65ec4f1002c4d3fb201dc8633638d0947c276be7008c227e1051ba5e`
Created	2026-06-22T16:05:51.981803091+00:00
Started	2026-06-22T16:29:08.894733553+00:00
Tags	`pedigree`, `fig5`, `sweepga`, `pafchop`, `slurm`, `validation`, `plot`, `eval-scheduled`

Description

Run only after the PAF chopper semantics validation and sweepGA identity-scoring audit are complete. Do not use any chopped PAF generated before validation unless PAF_SEMANTICS_VALIDATION.md explicitly marks it valid for identity-sensitive filtering.

Task:

Use the validated pafchop-rs behavior and the audited sweepGA command for per-chunk identity filtering.
Re-run the f16 whole-genome chop sensitivity for chop lengths at least 10000, 5000, 2000, and 1000 bp.
Use Slurm shared-node array jobs only: no --exclusive, modest per-element resources (--cpus-per-task about 4 unless justified, --mem about 16G-32G), and array concurrency cap so jobs can share nodes.
Heavy gzip/chop/sweepGA/PAF scanning must occur inside Slurm jobs, not on the head/login node.
Use --scaffold-jump 0 and the audited identity-scoring option, likely --scoring ani, if the audit confirms it.
Summarize chr3 survival for PAN027 and PAN028 candidate windows, including chr3 rows, chr3 summed overlap, chr3 query-union bp, chr9 query-union bp, other target union bp, and yes/no survival.
Generate/update the focused Fig5/SweepGA chop sensitivity SVG/PDF.

Acceptance:

Direct answer: after validated chopping and identity scoring, does smaller chopping improve strict 1:1 chr3 survival relative to 10kb?
Direct answer: which chop length is best for the visual evidence layer?
Provenance shows Slurm worker nodes ran all heavy steps.
Large PAFs remain ignored; scripts/summaries/plot are tracked.
No submission/ files modified.
Commit and push with WG provenance.

Run only after the PAF chopper semantics validation and sweepGA identity-scoring audit are complete. Do not use any chopped PAF generated before validation unless `PAF_SEMANTICS_VALIDATION.md` explicitly marks it valid for identity-sensitive filtering.

Task:
- Use the validated `pafchop-rs` behavior and the audited sweepGA command for per-chunk identity filtering.
- Re-run the f16 whole-genome chop sensitivity for chop lengths at least 10000, 5000, 2000, and 1000 bp.
- Use Slurm shared-node array jobs only: no `--exclusive`, modest per-element resources (`--cpus-per-task` about 4 unless justified, `--mem` about 16G-32G), and array concurrency cap so jobs can share nodes.
- Heavy gzip/chop/sweepGA/PAF scanning must occur inside Slurm jobs, not on the head/login node.
- Use `--scaffold-jump 0` and the audited identity-scoring option, likely `--scoring ani`, if the audit confirms it.
- Summarize chr3 survival for PAN027 and PAN028 candidate windows, including chr3 rows, chr3 summed overlap, chr3 query-union bp, chr9 query-union bp, other target union bp, and yes/no survival.
- Generate/update the focused Fig5/SweepGA chop sensitivity SVG/PDF.

Acceptance:
- Direct answer: after validated chopping and identity scoring, does smaller chopping improve strict 1:1 chr3 survival relative to 10kb?
- Direct answer: which chop length is best for the visual evidence layer?
- Provenance shows Slurm worker nodes ran all heavy steps.
- Large PAFs remain ignored; scripts/summaries/plot are tracked.
- No `submission/` files modified.
- Commit and push with WG provenance.

Depends on

Required by

open .flip-fig5-sweepga-f16-validated-chop-sensitivity

✉ Messages 1 message

#1user2026-06-22T19:52:53.569999659+00:00sent

STOP: sweepGA PAF identity scoring audit deliverables are absent; current validated rerun is not validly gated. Cancel any Slurm array/jobs and do not resubmit. Preserve scripts/logs only.

Log

2026-06-22T16:28:18.347794784+00:00 Lightweight assignment: agent=random-agent-46f6237a (46f6237a), exec_mode=full, context_scope=graph, reason=The Pipeline Operator is the closest fit for a Slurm-heavy rerun that must validate inputs, execute array jobs on shared nodes, update the Fig5 artifact, and produce a provenance-backed commit.
2026-06-22T16:29:08.894738673+00:00 Spawned by coordinator --executor codex --model gpt-5.5
2026-06-22T16:29:21.784745948+00:00 Starting validated f16 chop sensitivity rerun; checking prior work, validation artifacts, and existing pipeline layout.
2026-06-22T16:33:42.644906009+00:00 Prepared validated Slurm array workflow: lengths 10000/5000/2000/1000 for PAN027 and PAN028 candidate comparisons; filter command pins --num-mappings 1:1 --scaffold-jump 0 --scoring ani.
2026-06-22T19:52:53.784336342+00:00 STOP: cancelled Slurm array 1705408 because audit-sweepga-paf-filter-identity-scoring did not produce required audit deliverables; rerun was invalidly unblocked.
2026-06-22T19:53:15.583879877+00:00 Agent 'agent-2670' killed — task auto-paused (use 'wg resume' to re-enable dispatch)