← graph

park-off-target

Park off-target manuscript-session materials in paper_prep/_brainstorming/

Metadata

Statusdone
Assignedagent-744
Agent identityf51439356729d112a6c404803d88015d5b44832c6c584c62b96732b63c2b0c7e
Created2026-05-05T23:10:18.166108593+00:00
Started2026-05-05T23:14:14.667469389+00:00
Completed2026-05-05T23:26:47.314493739+00:00
Tagspaper-prep, housekeeping, pre-audit, eval-scheduled
Eval score0.83
└ blocking impact0.85
└ completeness0.85
└ constraint fidelity0.25
└ coordination overhead0.85
└ correctness0.82
└ downstream usability0.80
└ efficiency0.80
└ intent fidelity0.76
└ style adherence0.85

Description

GOAL: Move all off-target materials produced during the prior 'manuscript' session into paper_prep/_brainstorming/ so they are clearly NOT canonical, while preserving them for possible salvage. The canonical abstract has already been saved at paper_prep/synthesis/ABSTRACT.md by the coordinator — do NOT modify it.

CONTEXT: The prior synthesis session hill-climbed on copy-number-weighted ORA / hypergeometric-test / OR4F-gene-family minutiae and produced a manuscript draft (rendered as MANUSCRIPT_DRAFT.pdf) that fundamentally does NOT match the actual paper described in paper_prep/synthesis/ABSTRACT.md. Read ABSTRACT.md FIRST before deciding what to move. The user (lead author) has explicitly asked for these off-target materials to be parked so the synthesis directory only contains canonical materials going forward.

THE FILTER: Read paper_prep/synthesis/ABSTRACT.md. Anything that is clearly NOT about (a) the implicit pangenome graph methodology, (b) 466-haplotype HPRC v2 dataset, (c) extended interchromosomal subtelomeric homology, (d) NJ-tree cladistic analysis of subtelomere similarity (Xp/Yp, Xq/Yq, acrocentrics, 10p–18p, the 22q/21q/19q/1q/13q/17q clade, 4q–10q DUX4), (e) PCA + community detection of similarity matrix, (f) Hi-C 3D genome maps testing nuclear-envelope-proximity recombination hypothesis — is off-target.

EXAMPLES OF OFF-TARGET MATERIAL (move these):

  • Anything matching ora, phyper, weighted_phyper, copy_number_enrichment, copy_weighted, hypergeometric
  • Anything specifically about OR4F gene family (or4f_*.md), GREAT/rGREAT, GAT, gprofiler, neuron projection, snRNP, silencing genes, miRNA enrichment
  • 'mathematical_framework_copy_weighted_ora.md' and similar copy-weighted-ORA theory docs
  • 'methodology_recommendations_section.md', 'manual_evaluation_report_statistical_validation_framework.md', 'comprehensive_*' if their content is ORA/enrichment-flavored
  • paper_prep/synthesis/MANUSCRIPT_DRAFT.md (the off-target draft itself), MANUSCRIPT_SKELETON.md, ARCHITECT_TASK_BRIEF.md, WORK_DECOMPOSITION.md, ACCEPTANCE_CHECKLIST.md, STATS_AUDIT.md, SCRIPT_INVENTORY.md, TALK_OUTLINE_15MIN.md, VERSIONS.md, CAPTIONS.md — open each one and check if it's anchored on the abstract; if not, move it.

EXAMPLES OF MATERIAL TO LEAVE IN PLACE (do NOT move):

  • paper_prep/synthesis/ABSTRACT.md (canonical anchor)
  • paper_prep/synthesis/REFERENCES.bib (will be reused)
  • paper_prep/synthesis/MANUSCRIPT_DRAFT.typ + render.log + pandoc_convert.log (will be regenerated; leave as historical record of what we just rendered)
  • paper_prep/figures/* (do NOT move figures — the audit task will assess them; leave them in place even if their captions look off)
  • chm13.phrs.bed, chm13.phrs.no_acro.bed, CHM13-HG002.sub-telo-phrs.bed, chm13-annotations.bed, chm13v2.0_RefSeq_Liftoff_v5.2.gff3.gz (these look like real subtelomere data assets)
  • .wg-worktrees/, .wg/, .wg.clean/ (workgraph internals)
  • CLAUDE.md, OCTOPUS_CLUSTER.md (project infra docs)

UNCERTAIN CASES: If you can't tell from filename + a quick read of file head whether something is off-target, leave it where it is. Better to under-move than to move something canonical.

TASK STEPS:

  1. Read paper_prep/synthesis/ABSTRACT.md fully.
  2. mkdir -p paper_prep/_brainstorming/
  3. Inventory candidate files (top-level repo root + paper_prep/synthesis/ only — do NOT recurse into other dirs). Use the filter above.
  4. For each tracked file, use 'git mv' to preserve history. For untracked files, use plain 'mv'.
  5. Write paper_prep/_brainstorming/README.md with:
    • Header: 'Off-target materials from prior manuscript session'
    • Explanation: these were produced during a prior synthesis run that hill-climbed on copy-number-weighted ORA / gene-family minutiae and did not anchor on the canonical abstract. Retained for possible salvage but are NOT part of the canonical manuscript pipeline.
    • Date of relocation
    • List of files moved, grouped by category (ORA/enrichment, gene-family-specific, off-target synthesis docs, etc.)
    • Pointer: 'For canonical material, see paper_prep/synthesis/ABSTRACT.md and downstream artifacts.'
  6. git add paper_prep/_brainstorming/ and any modified tracked files
  7. Commit: 'chore: park off-target manuscript-session materials in _brainstorming/'

ACCEPTANCE:

  • paper_prep/_brainstorming/ exists and contains the relocated files
  • paper_prep/_brainstorming/README.md exists, names every relocated file, explains why
  • paper_prep/synthesis/ABSTRACT.md is UNCHANGED (verify byte-equality before and after)
  • paper_prep/synthesis/REFERENCES.bib still in place
  • paper_prep/figures/ untouched
  • Top-level subtelomere data files (chm13.phrs.bed etc.) still in place
  • Single clean commit with message above
  • 'git log --oneline -1' shows the commit
  • 'wg artifact' records paper_prep/_brainstorming/README.md as an artifact

DO NOT in this task:

  • Modify ABSTRACT.md in any way
  • Touch paper_prep/figures/
  • Delete any file (only move)
  • Modify .gitignore
  • Run any data analysis or rewrite anything

Depends on

Required by

Log