← graph

fig5-maternal-native-untangle-diff

Fig5 maternal untangle native-output plot diff

Metadata

Statusdone
Assignedagent-2560
Agent identity46f6237a65ec4f1002c4d3fb201dc8633638d0947c276be7008c227e1051ba5e
Created2026-06-19T08:25:43.444519978+00:00
Started2026-06-19T08:27:06.308300016+00:00
Completed2026-06-19T08:36:01.441341730+00:00
Tagspedigree, fig5, untangle, visualization, eval-scheduled
Eval score0.92
└ blocking impact0.94
└ completeness0.95
└ constraint fidelity0.10
└ coordination overhead0.89
└ correctness0.91
└ downstream usability0.91
└ efficiency0.88
└ intent fidelity0.76
└ style adherence0.93

Description

Focused objective: determine whether the native odgi BEDPE/PAF + sweepGA path changes the plotted untangle view of the proband maternal haplotype against maternal chromosomes, specifically the Fig. 5/PAN027_vs_PAN010 maternal panel. This is a visualization/provenance comparison task, not a manuscript rewrite and not a recombination mechanism analysis.

Question to answer:

  • If we redraw the PAN027 maternal haplotype (PAN027 hap1, inherited from PAN010) against PAN010 maternal/paternal target chromosomes using native odgi BEDPE/PAF and sweepGA-filtered PAF, does the visible Fig. 5-style untangle plot change relative to the current first-best/old source view?

Inputs:

  • Existing manuscript figure: submission/fig/MainFigures/Fig5_pedigree_untangle.pdf
  • Existing old untangle BED: /moosefs/guarracino/HPRCv2/PHR_III/pedigrees/washu/untangle/PAN027_vs_PAN010.e50000.m1000.bed.gz
  • Native Slurm outputs from job 1703959: /moosefs/erikg/phrs/pedigree_native_untangle_agent2556_slurm/PAN027_vs_PAN010.e50000.m1000.j0.8.n4.bedpe.gz and .paf.gz
  • sweepGA test/output directory if useful: /moosefs/erikg/phrs/pedigree_native_untangle_agent2556_sweepga_test/
  • Current helper code: scripts/pedigree/untangle_native_merge_tracts.py, scripts/pedigree/run_untangle_native_merge_tracts.sbatch, scripts/pedigree/untangle_multimap_tracts.py
  • Decision record: paper_prep/_brainstorming/pedigree_native_untangle_merge.md

Hard scope:

  • Do not edit submission/paper.tex.
  • Do not replace submission/fig/MainFigures/Fig5_pedigree_untangle.pdf in this task.
  • Do not run heavy odgi untangle on the head node. Use existing Slurm outputs; if missing, only write an sbatch command/report, do not run heavy work directly.
  • Do not make conversion/crossover mechanism claims.
  • The output is a visual comparison and a yes/no decision about whether the Fig. 5 maternal plot would change.

Required work:

  1. Inspect the current Fig. 5 asset/provenance enough to identify what the existing maternal panel is showing. If the exact original plotting recipe is unavailable, state that clearly and make a standardized comparison plot using the old and new data with the same visual grammar.
  2. Verify query/target semantics. Do not assume the old PAN027_vs_PAN010 BED is child-query; inspect the path names and document whether the old file is parent-query, child-query, or mixed relative to the current Fig. 5 claim.
  3. Build a compact comparison table for PAN027_vs_PAN010 only, aligning comparable query arms/intervals between:
    • old first-best/source view
    • native BEDPE n1/n4 view
    • native PAF n1/n4 view
    • sweepGA-filtered PAF views for 1:many, 2:many, 4:many if relevant and already available or cheap to generate from uncompressed PAF.
  4. Render a side-by-side visual artifact under paper_prep/_brainstorming/fig5_maternal_native_diff/ showing the same PAN027 maternal query arms/intervals across old/native/sweepGA views. The plot should make differences visible: changed donor chromosome/arm assignment, added alternate hits, removed hits, rank/n-best differences, and ambiguous multimapping.
  5. Write a concise decision record at paper_prep/_brainstorming/fig5_maternal_native_diff.md answering:
    • Does the plotted maternal untangle view change visually?
    • If yes, what changes: donor arms, additional equivalent targets, fewer/more ribbons, or changed segmentation?
    • Is the current Fig. 5 still defensible as a schematic/first-best view, or should a later figure-replacement task be opened?
    • What exact data source should a future regenerated Fig. 5 use?
  6. If a future replacement is warranted, recommend the next task in one paragraph, but do not perform it here.

Deliverables:

  • paper_prep/_brainstorming/fig5_maternal_native_diff.md
  • paper_prep/_brainstorming/fig5_maternal_native_diff/ with at least one PDF/SVG/PNG comparison and source TSV(s)
  • Optional script under scripts/pedigree/plot_fig5_maternal_native_diff.py or .R if needed for reproducibility

Acceptance criteria:

  • The answer is explicitly about the maternal PAN027_vs_PAN010 Fig. 5-style plot, not global tract distributions.
  • The task produces a visual diff artifact, not just counts.
  • The report distinguishes unchanged manuscript asset from changed-if-regenerated data content.
  • No large PAF/BEDPE intermediates are committed.

Depends on

Required by

Log