review-zoom-v3-human-3d-dotplot-correlation

Metadata

Status	done
Assigned	`agent-1021`
Agent identity	`3577bc75d6ed4f1947509aa5c086c91ce7c997c7806dab6bf6affac647452647`
Created	2026-05-07T00:58:43.481514002+00:00
Started	2026-05-07T01:01:02.272470694+00:00
Completed	2026-05-07T01:10:17.840768826+00:00
Tags	`bog-this-week`, `bog-deck`, `review-zoom`, `review-zoom-v3`, `eval-scheduled`
Eval score	0.97
└ blocking impact	0.98
└ completeness	0.97
└ coordination overhead	0.96
└ correctness	0.97
└ downstream usability	0.97
└ efficiency	0.94
└ intent fidelity	0.97
└ style adherence	0.96

Description

User feedback: slide 12 mouse dot plot is awesome; do we have anything like this for regular human data, the dot plot and correlation specifically?

Goal: determine whether we can build a human analog of the mouse zygotene scatter: per arm-pair or per PHR-pair sequence similarity versus Hi-C/Pore-C contact, with a correlation/statistic, and generate a slide-ready candidate if valid.

Requirements:

Inspect human 3D source data in paper_prep/figures/fig3, ED5, and HPRCv2 analysis/human paths.
Identify the appropriate unit: arm pair, arm-haplotype pair, or PHR pair. Be explicit about which is available and comparable to mouse.
For at least one strong human dataset, preferably HG002 Pore-C and/or CHM13 Hi-C at 50 kb, build a scatter/dot plot of sequence similarity against contact or O/E contact.
Compute Spearman rho or the appropriate correlation. Distinguish pointwise Spearman from Mantel rho.
If only arm-pair matrices are available and not per-PHR-pair data, say so and make a correct arm-pair analog instead of overclaiming.
Keep outputs under slides/v2-review-zoom/_revision_assets/v3/human_3d_dotplot/.

Validation:

README states whether a human analog exists and at what unit.
Candidate plot includes axes, n, correlation, sample/platform/resolution.
The plot does not reuse the mouse statistic or call pointwise Spearman a Mantel test.

User feedback: slide 12 mouse dot plot is awesome; do we have anything like this for regular human data, the dot plot and correlation specifically?

Goal: determine whether we can build a human analog of the mouse zygotene scatter: per arm-pair or per PHR-pair sequence similarity versus Hi-C/Pore-C contact, with a correlation/statistic, and generate a slide-ready candidate if valid.

Requirements:
- Inspect human 3D source data in paper_prep/figures/fig3, ED5, and HPRCv2 analysis/human paths.
- Identify the appropriate unit: arm pair, arm-haplotype pair, or PHR pair. Be explicit about which is available and comparable to mouse.
- For at least one strong human dataset, preferably HG002 Pore-C and/or CHM13 Hi-C at 50 kb, build a scatter/dot plot of sequence similarity against contact or O/E contact.
- Compute Spearman rho or the appropriate correlation. Distinguish pointwise Spearman from Mantel rho.
- If only arm-pair matrices are available and not per-PHR-pair data, say so and make a correct arm-pair analog instead of overclaiming.
- Keep outputs under slides/v2-review-zoom/_revision_assets/v3/human_3d_dotplot/.

Validation:
- README states whether a human analog exists and at what unit.
- Candidate plot includes axes, n, correlation, sample/platform/resolution.
- The plot does not reuse the mouse statistic or call pointwise Spearman a Mantel test.

Depends on

Required by

Log

2026-05-07T00:58:43.457286383+00:00 Task paused
2026-05-07T00:59:08.835946445+00:00 Task published
2026-05-07T01:01:02.272475373+00:00 Spawned by coordinator --executor codex --model gpt-5.5
2026-05-07T01:01:21.021235581+00:00 Starting human 3D dot plot review: inspecting prior WIP, data availability, and task validation requirements.
2026-05-07T01:04:49.389034120+00:00 Inspected fig3/ED5 and HPRCv2 human paths: main sources are arm-haplotype contact matrices plus arm-pair Jaccard/contact summaries; community_free sequence-level PHR-pair controls also exist under exclusion_controls.
2026-05-07T01:09:15.997710409+00:00 Validated: regenerated plots with Rscript; git diff --check passes; README covers unit decision, source inspection, pointwise Spearman vs Mantel distinction, and candidate plot requirements. Cargo build/test not applicable for this documentation/asset task.
2026-05-07T01:09:52.236659927+00:00 Committed: 116734c — pushed to origin/wg/agent-1021/review-zoom-v3-human-3d-dotplot-correlation
2026-05-07T01:10:17.840776601+00:00 Task pending eval (agent reported done; awaiting `.evaluate-*` to score)
2026-05-07T01:15:35.885695448+00:00 PendingEval → Done (evaluator passed; downstream unblocks)