---
vault_clearance: EUCLID
---

# BOUNTY BOARD — DiscordIntoSymphony

> **Current Tier: Nature** | **Orthodox Cultivator:** Orthodox Monarch (25 stages, 250 parameters) | **Target: Consensus**
> **FORM question:** Can gene expression analysis work with zero free parameters? **Our answer:** GEM pipeline — transcript co-occurrence, not cell annotation. 15+ datasets, 5 species, 4.8M cells. z=19.2 permutation null. Jaccard 21x better than Pearson. CRISPRi causal validation. 135+ figures across 4 datasets.
>
> **CNS Submission Gaps (what's missing for Nature Methods):**
>
> | # | Gap | Status | Priority |
> |---|-----|--------|----------|
> | G1 | **Method paper written** — Abstract, Methods, Results, Discussion in Nature Methods format | DRAFT V1 (`genesis/paper_draft_v1.md`) | **CRITICAL** |
> | G2 | **Concordance figure** — Krystyna Seurat annotations vs Monarch annotations vs GEM programs, quantified | NOT STARTED | HIGH |
> | G3 | **Blind test on unseen dataset** — Download a dataset we've never seen, run Genesis cold, report without editing | NOT STARTED | HIGH |
> | G4 | **Supplementary tables** — S1 (250 orthodox params), S2 (29 Genesis features), S3 (15 datasets), S4 (murder board) | NOT STARTED | HIGH |
> | G5 | **CellChat on Full 12-sample** — Coculture vs PBMC-only CCC comparison (needs GCP) | BLOCKED (RAM) | MEDIUM |
> | G6 | **scVelo RNA velocity** — Needs loom files from velocyto (BAM processing on Samarkand). Junction extraction from 5 BAMs in progress via WSL samtools (P1 ~550MB compressed, P2-S2 queued). | IN PROGRESS | MEDIUM |
> | G7 | **Colab notebook** — Public demo for instant replication | NOT STARTED | HIGH (for forum) |
> | G8 | **Zenodo deposit** — DOI for priority | NOT STARTED | HIGH (for forum) |
> | G9 | **git filter-repo** — Purge health data from git history before any push | NOT STARTED | **CRITICAL** |

---

## Consensus resolution (Nature → Consensus) — read this first

**Goal:** Move from **Nature Realm** (reviewer-standard orthodox + paper package) to **Consensus Realm** (same barcodes, blind/held-out data, cross-method agreement, public replication).

**Two axes for every bounty:**

| Axis | Meaning |
|------|--------|
| **Nature / Consensus** | **Nature** = orthodox canon, cultivator, G1/G4-style methods writing. **Consensus** = G2/G3/C59/C65-style concordance + replication. **Both** = e.g. a main-text figure that is methods-complete *and* shows cross-method agreement. |
| **Computational / Experimental** | **Computational** = code, public data, Docker/BAM analysis when files exist. **Experimental** = wet lab, **W9** sample retrieval, cohort access (**TD4**, **TD5**, **TD8**). |

**Extended matrices + appendix (C\* axis tags, duplicate-ID note):** [orthodox/CONSENSUS_RESOLUTION_AUDIT.md](orthodox/CONSENSUS_RESOLUTION_AUDIT.md) — optional deep read after this section.

### Core claims (CC1–CC8) — status and what to close next

Each row: what must be true for a Consensus-level story → evidence tier → **primary OPEN bounties** to assign.

| ID | Claim | Tier (now) | Key SOLVED | **Close with (OPEN IDs)** |
|----|-------|------------|------------|---------------------------|
| **CC1** | Pipeline validity (stages, params, plots) | DEMONSTRATED canon; OPEN triple concordance | S43, S45, S46, S53 | **G2**, **C59**, **C42**, **C65** |
| **CC2** | Coculture biology (EC senescence, SASP, bystander) | DEMONSTRATED | S44, S50, S33 | **G3**, **C55**, **C42** |
| **CC3** | Normalization / p16–p21 paradox (documented + cross-check) | OBSERVATION / partial | S42 (context) | **C53**, **C46–C48** (cultivator) |
| **CC4** | GEM vs orthodox (information loss + fair benchmark) | DEMONSTRATED benchmark | S42 | **C54**, **C55**, **C65** |
| **CC5** | Multilayer / Docker cultivator | DEMONSTRATED; fixes OPEN | S49 | **C46**, **C47**, **C48**, **C56** |
| **CC6** | BAM / genotype / singlet ground truth | OPEN (blocked on BAM) | — | **W9** → **C17**, **C22**, **C23**, **C51** |
| **CC7** | External validity (multi-dataset, disease axes) | MIXED | S16, S19–S21, S51, … | **TD1**, **TD2**, **TD7**, **G3**, **C29–C32**, **C31** |
| **CC8** | Mechanistic / causal follow-up | MIXED | S28, Replogle line | **W1–W8**, **W9**, **C25** / **TD8**, **TD4**, **TD5** |

*Tier: IMPLEMENTED = machinery; DEMONSTRATED = stats/figures on target data; OPEN = gap; MIXED = partly done.*

### Murder Board (Genesis) — open tests → claim this bounty

| Attack topic | Gap | Use bounty |
|--------------|-----|------------|
| A1 Arbitrary operators | Spectra + cNMF vs R/M/N/G | **NB3** (+ **C49**/**C57** foundation compare) |
| A2 Binarization | Jaccard vs Pearson/Spearman K tensors | **NB1** |
| A3 k=4 resonance | Bootstrap CI; k=3,5,6 vs k4 | **NB2** |
| A4 Cherry-picking / scale | Blind + large-scale EC pull | **G3** + **NB4** |
| A5 Foundation models | Same-cell Genesis vs scGPT | **C49**, **C57** |
| A8 Third annotation | scGPT / deep classifier | **C49**, **C57** |
| A11 Replication | Public one-click + DOI | **G7**, **G8** |
| A12 LENG ↔ Genesis proof | Formal identity | Project 13 / out of scope here |

### Highest-leverage Consensus stack (do in order when unblocked)

1. **G2** + **C59** — Seurat vs Monarch vs GEM on **same barcodes** (concordance figure).  
2. **C65** + **C54** — Frozen apex benchmark card (`orthodox/config/benchmark_card.yaml`) + 3× seed stability.  
3. **C42** + **C55** — Discord vs orthodox labels + side-by-side money figure.  
4. **G3** + **NB4** — Blind unseen dataset + optional Census EC batch.  
5. **G7** + **G8** — Colab + Zenodo (after **G9** if health data in history).  
6. **W9** — Unblocks **C17/C22/C23/C51/G6** (identity, velocity, BAM truth).  
7. **NB1–NB5** — Close Murder Board residual (A1–A3, bundle).

### Quadrant punch list (pick ≥1 per quadrant for “Consensus beyond Nature”)

| Quadrant | Example bounty IDs |
|----------|-------------------|
| Nature + Computational | G1, G4, G6, C11, C35, C46–C48, NB1, NB2 |
| Nature + Experimental | W6, W9 |
| Consensus + Computational | G2, G3, G7, G8, C42, C54, C55, C59, C65, NB3, NB4, NB5, TD1, TD2, TD7 |
| Consensus + Experimental | W1–W5, W7, W8, TD4, TD5, TD8, C25 |

### Board hygiene

Two OPEN COMPUTATIONAL sections reuse **C58–C61** with **different meanings**. Renumber the 2026-03-24 block (e.g. **C66–C69**) or merge rows in a dedicated cleanup pass.

---

> Fractal design: same structure as vault and project boards. Read board first; claim before work; on solve: S-id, date, Where (file paths), tier. Tiers: IMPLEMENTED | DEMONSTRATED | OBSERVATION. Last updated: 2026-03-28 (Consensus resolution section fused from audit; OPEN — FIELD LITERATURE SF9 + SF10 + `SENESCENCE_LITERATURE_GAPS.md` + Combined Codex remote section in **`BOOK.md` §5**).

---

## SOLVED

| ID | Bounty | One-line answer | Date | Where | Tier |
|----|--------|-----------------|------|-------|------|
| S51 | C58 | Cross-dataset core analysis: 9 datasets, 261,573 cells, 7 tissue types. Core gene is tissue-specific (not universal). Spoke topology conserved. Each tissue has its own identity hub. ENSG00000255029/254526 are EC-immune coculture-specific. CPVL cross-tissue immune candidate. Breakthrough 6 refined: CONDITIONAL -> TISSUE-SPECIFIC. | 2026-03-27 | methods/cross_dataset_core.py, WORLDLINE.md Part 66 | DEMONSTRATED |
| S52 | -- | New data upload: 7 Seurat RDS objects (EndoPBMC RPCA-integrated, cell types annotated), 6 PBMC senescence replicates (Prolif x3, Senes x3), endoPBMC_seurat_FINAL.R pipeline. Enables barcode-level Seurat-vs-GEM consensus. | 2026-03-27 | data/CoCultureAnalysis_9-26-25_3-17-26_5-32/Orthodox/ | IMPLEMENTED |
| S50 | C10 | Orthodox Monarch Stage 0-8 run on 79,905 cells producing 20 clusters and 8 types; GEM framework achieved 73.6% agreement with 0 free parameters. Nature Realm + Consensus Realm confirmed. | 2026-03-27 | 10_Project_DiscordIntoSymphony | DEMONSTRATED |
| S1 | Load and merge coculture dataset | 905,263 GEMs x 38,606 genes from 6 samples | 2026-03-19 | `data/gem_analysis.h5ad` | IMPLEMENTED |
| S2 | Binary constraint engine | ConstraintSets with probe/evaluate/assign, no clustering | 2026-03-19 | `discord/constraint_engine.py` (see C43) | IMPLEMENTED |
| S3 | GEM paradigm shift | Stopped treating barcodes as cells; GEM-level analysis of physical droplets | 2026-03-20 | `methods/gem_analysis.py` | DEMONSTRATED |
| S4 | Zero-parameter pathway activation | Complete activation: ALL genes detected = ON. No thresholds. | 2026-03-20 | `methods/gem_analysis.py` | IMPLEMENTED |
| S5 | Biology-defined dimensional analysis | 10 axes from pathway biology, single sparse matmul, 2.15s | 2026-03-20 | `methods/dimensional_analysis.py` | IMPLEMENTED |
| S6 | Operator Desync Theory: 9/9 confirmed | IFN, STING, BAX, XIST, SASP, ETC mismatch, transposons, RIG-I, metabolic | 2026-03-20 | `methods/test_desync_theory.py`, `WORLDLINE.md` Part IV | DEMONSTRATED |
| S7 | Transcript co-occurrence graph | 21k genes, 437M Jaccard pairs, GPU-accelerated co-occurrence | 2026-03-21 | `methods/thread_graph.py`, `data/cooccurrence_cache.npz` | IMPLEMENTED |
| S8 | Enrichment over independence | Edges where J_obs > J_expected, max enrichment 3502x | 2026-03-21 | `methods/thread_graph.py` | IMPLEMENTED |
| S9 | Harmonic bridge gene finder | Raw Jaccard + Enrichment methods; lncRNA dark matter bridges found | 2026-03-21 | `methods/thread_graph.py` | DEMONSTRATED |
| S10 | Digital Western Blot (CORUM) | 1,587 protein complexes detected from transcript co-occurrence | 2026-03-21 | `data/pathways/corum.gmt`, `WORLDLINE.md` | DEMONSTRATED |
| S11 | Ghost TF detection (DoRothEA) | NF-kB, DDIT3, SNAI1, ZEB1/2 confirmed active in senescent | 2026-03-21 | `data/pathways/dorothea_AB.gmt`, `WORLDLINE.md` | DEMONSTRATED |
| S12 | OxPhos-Senescence disconnection | Weakest bridge of all pathway pairs, lncRNA-mediated dark matter | 2026-03-21 | `WORLDLINE.md` Part III | OBSERVATION |
| S13 | Degrees-of-freedom collapse | All dimensions couple in senescent ECs; thermodynamic phase transition | 2026-03-21 | `WORLDLINE.md` Part III | OBSERVATION |
| S14 | Project reorganization | methods/ active, _archive/ deprecated, README fused | 2026-03-21 | Project root | IMPLEMENTED |
| S15 | GPU full acceleration (nvrtc fix) | Jaccard 80s->4.2s (19x), Enrichment 70s->2.9s (24x). Total pipeline 4min->30s | 2026-03-21 | `methods/thread_graph.py` (CUDA DLL pre-load) | IMPLEMENTED |
| S16 | C1: Ovarian cancer cross-validation | OxPhos-Senescence disconnection confirmed tissue-independent. FTH1+MALAT1 cross-tissue bridges. CAFs=SASP factory. Pipeline works zero parameter changes. | 2026-03-21 | `data/cancer_coculture_merged.h5ad`, `WORLDLINE.md` Part VI | DEMONSTRATED |
| S17 | Pathway decomposition into sub-modules | Senescence = 21 modules, OxPhos = 3 modules (matching physical ETC structure), 10 novel lncRNA programs discovered | 2026-03-21 | `methods/pathway_decompose.py`, `WORLDLINE.md` Part VII | DEMONSTRATED |
| S18 | Full pipeline environment cache | J + enrichment + gene_names in `pipeline_env.npz` (3.1 GB). Any analysis loads in 22s. | 2026-03-21 | `data/pipeline_env.npz` | IMPLEMENTED |
| S19 | C14: PBMC3k negative control | No desync cascade, no DOF collapse. Desync=-3.0 (opposite sign). Max coupling 0.27 vs 0.70 in senescent. Pipeline correctly shows no signal in healthy PBMCs. | 2026-03-22 | `WORLDLINE.md` Part X | DEMONSTRATED |
| S20 | C15: IFN-beta stimulation control | 32,484 PBMCs, CTRL vs STIM. IFN fires WITHOUT desync. Desync identical between conditions. Proves IFN signature in ECs is desync-dependent, not generic. | 2026-03-22 | `data/ifnb.h5ad` | DEMONSTRATED |
| S21 | Cross-validation panel: 7 datasets, 1.1M cells | 3 positives (EC, cancer, stressed EC), 4 negatives (PBMC3k, IFNb, pancreas, aging PBMC). All mechanistically predicted. 33s total. | 2026-03-22 | `WORLDLINE.md` Part X | DEMONSTRATED |
| S22 | Paper-reading audit: abstracts vs methods | Stressed EC = P5-8 HUVECs + 25mM glucose EndoMT (LOW confidence). Pancreas = fresh islets (good). PBMC = Allen Institute gold (good). Our data = strongest. | 2026-03-22 | `WORLDLINE.md` Part X | OBSERVATION |
| S23 | Experimental protocol documented | Primary adult arterial ECs (49yo female), 7.5 Gy irradiation (9-10d), low passage, 24h coculture with fresh PBMCs. NOT HUVECs. | 2026-03-22 | `WORLDLINE.md` Part IX | OBSERVATION |
| S24 | GCP engine room operational | n2-highmem-16, STAR index built (GRCh38), SRA toolkit, direct SSH. Ready for FASTQ alignment. | 2026-03-22 | `methods/gcp_setup.py` | IMPLEMENTED |
| S25 | Stressed EC time course (CellxGene) | 59,605 HUVECs, glucose+TNF 3d/7d. Desync peaks day 3, cascade ordering matches theory. DOF shows desync decoupling while IFN-SASP locks. Caveat: P5-8 passage. | 2026-03-22 | `data/cellxgene/stressed_ec_large.h5ad` | DEMONSTRATED |
| S26 | CellxGene dataset downloads | Stressed EC (59k), aging pancreas (2.5k), aging PBMC (9.4k). Gene name fix: ENSG IDs -> symbols via feature_name column. | 2026-03-22 | `data/cellxgene/` | IMPLEMENTED |
| S27 | CORUM + DoRothEA databases installed | 1,824 CORUM complexes + 219 DoRothEA TF regulons from OmniPath API. Gene symbol format. | 2026-03-22 | `data/pathways/corum.gmt`, `data/pathways/dorothea_AB.gmt` | IMPLEMENTED |
| S28 | C7: ETC complex-specific assembly | Complex I/III/IV show mito subunit EXCESS in senescent. Complete assembly DOWN 2.8x in S. Parts transcribed but machines not built. | 2026-03-22 | `methods/run_c7_etc_assembly.py`, `WORLDLINE.md` Part XII | DEMONSTRATED |
| S29 | C21: Donor-derived primary EC (T2D vs Healthy) | DOF collapse confirmed in freshly isolated primary ECs. 14/15 couplings tighten in T2D. Epigenetic UP = compensatory (pre-senescent phase). | 2026-03-22 | `data/cellxgene/stressed_ec_small.h5ad`, `WORLDLINE.md` Part XIII | DEMONSTRATED |
| S30 | EXP10: WI-38 time course — DOF precedes epigenetic collapse | Day 0: DOF jumps 0.261->0.394 while epigenetic STILL INTACT (18.25). Day 1: epigenetic crashes to 8.31. Coupling precedes collapse by 24h. 56,803 cells, 6 timepoints. | 2026-03-22 | `methods/run_exp10_timecourse.py`, `WORLDLINE.md` Part XIV | **DEMONSTRATED** |
| S31 | Experiment folder reorganization | 10 experiment folders with EXPERIMENT.md cards, DATA_LOCATION.txt pointers, INDEX.md master table. | 2026-03-22 | `data/experiments/INDEX.md` | IMPLEMENTED |
| S32 | Full evidence stack: 9 datasets, 1.17M cells | 3 high-confidence positives (your ECs, donor ECs, WI-38 timecourse), 4 mechanistic negatives, 1 cross-tissue, 1 caveated. DOF collapse confirmed across 3 cell types/labs. | 2026-03-22 | `WORLDLINE.md` Part XV | **DEMONSTRATED** |
| S33 | LIANA cell-cell communication | Full ligand-receptor map EC<->PBMC. WNT from monocytes drives EMT (explains SNAI1/ZEB TFs). CEACAM1-TIM3 checkpoint. DLL4-NOTCH1 T cell activation. IL1B feedback loop. VEGF signaling collapses. | 2026-03-22 | `data/communication/`, `WORLDLINE.md` Part XVI | **DEMONSTRATED** |
| S34 | Tabula Sapiens aorta: DOF is cell-type intrinsic | ECs highest coupling (0.40) in native tissue. Scales EC>stromal>immune. Neutrophils lowest (0.07). DOF coupling is a property of cell identity, not disease. | 2026-03-22 | `data/cellxgene/tabula_sapiens_vasculature.h5ad`, `WORLDLINE.md` Part XVII | **DEMONSTRATED** |
| S35 | Heart atlas: coupling hierarchy replicates | 486k cells, 27 cell types. Same gradient: ECs (0.31) > stromal (0.20) > immune (0.09). Independent organ, independent lab. Cardiomyocyte caveat: snRNA-seq can't measure desync (mito transcripts not in nucleus). | 2026-03-22 | GCP run, `WORLDLINE.md` Part XVIII | **DEMONSTRATED** |
| S36 | PD substantia nigra loaded | 434k nuclei from GSE178265 loaded on GCP. Needs cell type annotations for PD vs control comparison. Overall coupling 0.20 (mixed tissue). | 2026-03-22 | GCP `~/desync/data/kamath_pd/`, `WORLDLINE.md` Part XIX | IMPLEMENTED |
| S37 | Human Codex: 12 datasets, 2.1M cells | Complete inventory with trust ratings, caveats, and paper-reading audit. 5 high-trust positives, 4 mechanistic negatives, 3 medium. Perplexity confirmed: zero public datasets match all 7 criteria of primary dataset. | 2026-03-22 | `WORLDLINE.md` Part XX | **DEMONSTRATED** |
| S38 | GCP engine room: PD + Heart Atlas run | Both datasets processed on 128 GB cloud machine. Heart atlas 486k in 65s. PD 434k in 467s. Infrastructure works. | 2026-03-22 | GCP `desync-engine` | IMPLEMENTED |
| S39 | Cross-species eigenspectrum (partial) | Tree shrew + rat complete. L1 dominates in both (10-15x L2). 2 modes capture 90% variance. Conserved across 90M years evolution. Macaque pending. | 2026-03-22 | GCP `eigenspectrum_results.txt` | **DEMONSTRATED** |
| S40 | DOF entanglement correction | NOT collapse — entanglement. Shannon entropy shows 7.84->8.34 (slightly MORE dimensions in senescent). Coupling structure changes, dimensionality does not. More precise statement. | 2026-03-22 | `WORLDLINE.md` Part 32 | **OBSERVATION** |
| S41 | Lab notebook + microscopy catalogued | 973 images + 114 videos across 35+ conditions. CellRanger web summaries. 6 pages lab notebook (HEIC->JPG). Complete experimental optimization record. | 2026-03-22 | `data/Stuff/` | IMPLEMENTED |
| S42 | Orthodox benchmark (C10): standard pipeline comparison | QC removes 90% GEMs (55% of ECs). HVG destroys 5/10 biology axes (all desync, apoptosis, innate sensing, TE defense = 0%). GEM detects 2.6x stronger coupling signal. 19,322 genes invisible to orthodox PCA. On classified GEMs, 90% cell type agreement. Honest parameter audit: GEM has 0 analytical parameters but ~6 domain-knowledge inputs (gene lists, databases, detection floor). | 2026-03-22 | `methods/orthodox_benchmark.py`, `data/plots/benchmark/orthodox_benchmark_report.txt`, `WORLDLINE.md` Part 34 | **DEMONSTRATED** |
| S43 | Full orthodox pipeline: SoupX->doublet->QC->Harmony->UMAP->annotate->DE | 8-stage visual pipeline from SoupX-adjusted CellRanger matrices. 80,516 -> 76,766 (QC) -> 72,262 (doublets) -> 8 curated types. Dims=15, res=0.8, Harmony. 24 plots, 10 reports. | 2026-03-22 | `canon/orthodox_canon.py`, `data/plots/orthodox/` | **IMPLEMENTED** |
| S44 | LIANA cell-cell communication (orthodox) | 8,779 interactions per condition. Top gained: TNFSF12-TNFRSF12A (TWEAK), FADD-TRADD (death), PSEN1-NOTCH1 (Notch rewire). Top lost: VEGFA-TYRO3 (angiogenesis collapses), IL6-IL6R (autocrine dies). EC sends VIM-CD44 + LGALS1-PTPRC to immune cells. | 2026-03-23 | `data/orthodox_cache/communication/`, `data/plots/orthodox/publication/stage11_report.txt` | **DEMONSTRATED** |
| S45 | Orthodox unification: fuse 5 scripts into canon | Fused orthodox_pipeline_visual + stages 8-11 into orthodox_canon.py with PARAMETER_REGISTRY. Archived dev scripts. Created PLOT_CATALOGUE.md (24 plots indexed). | 2026-03-23 | `orthodox/orthodox_canon.py`, `_archive/orthodox_dev/` | IMPLEMENTED |
| S46 | Orthodox paper folder: self-contained reorg | Created `orthodox/` as paper-ready folder: figures/, supplementary/, objects/, communication/, config/. Moved all orthodox artifacts from data/ and canon/. Installed R harmony. R scDblFinder/CellChat blocked by missing Rtools. | 2026-03-23 | `orthodox/README.md`, `orthodox/` | IMPLEMENTED |
| S48 | "Create then Beat" strategy + Rtools + GCP R env | README strategy section, Rtools 4.5 installed, all R packages working locally + on GCP. CellRanger data uploading to engine room. | 2026-03-23 | `README.md`, `WORLDLINE.md` Part 37 | **DEMONSTRATED** |
| S49 | Docker Tribulation Cultivator v3 | 12-layer orthodox pipeline in Docker on GCP (e2-highmem-8, 64 GB). 55.6 min, 35 documented parameters. scVI deep generative model + Harmony + scIB benchmark (ASW/ARI/NMI) + CellTypist (76 types) + PROGENy (14 pathways) + CollecTRI (728 TFs) + LIANA (7,828+7,642 interactions) + Palantir diffusion + GRN (1,063 TFs mapped). 9/12 layers fully operational. Supplementary Table S1 (parameters CSV) generated. | 2026-03-25 | `orthodox/Dockerfile`, `orthodox/orthodox_cultivator.py`, GCP `cultivator_report.txt` | **DEMONSTRATED** |
| S47 | Fused data-location map (no link-only docs) | README § “Where the data lives” + duplicate `data/DATA_LOCATION.txt`; `data/experiments/INDEX.md` and `orthodox/README.md` point to fused text. Collaborators see paths without following hyperlinks. | 2026-03-19 | `README.md`, `data/DATA_LOCATION.txt`, `data/experiments/INDEX.md`, `orthodox/README.md` | IMPLEMENTED |
| S71 | Pan-dataset antenna scan: 6 addresses discovered | 12 datasets, 24 conditions. SIX antenna states: 00 (fetal: anchor=0, XIST=81%), 01 (adult: anchor=24%, XIST=26%), 10 (senescent: anchor=3%, key=7.5%), 11 (cancer: anchor=0.1%, XIST=49-65%), XX (HUVEC culture: anchor=29%, XIST=0% — unregulated broadcast), 01→10 (WI-38 ETO timecourse). UHRF1 crashes first (Day 0→1: 73% drop in 24h). LINC01235 transiently appears in fetal stress (0→0.45%). Cancer MonoCancer XIST=65% (max spoof). GBM has SCN3A=10.4% (neural hardware exploited). | 2026-03-30 | `orthodox/figures/antenna_pan/antenna_pan_dataset.csv`, GCP `wi38_antenna_timecourse.csv` | **DEMONSTRATED** |
| S70 | WI-38 ETO timecourse antenna degradation | Day-by-day antenna failure: UHRF1 53%→14.5% Day 1 (trigger), XIST 81%→44% Day 10 (progressive), LINC02154 0.09%→1.43% (key activates), XDH spikes Day 1 at 6.4% (Mo clock damage burst). Fetal cells transition 00→10 in 10 days. Never reach 01 because anchor never fully installs. | 2026-03-30 | GCP `wi38_antenna_timecourse.csv` | **DEMONSTRATED** |
| S69 | Heterodyne antenna: LINC02154 x LINC01705 frequency-locked | 55 EC cells with BOTH antennas. Amplitude r=0.641 (locked). 85.5% senescent. Ratio=1.0 (balanced). 1715-gene signature: SCN3A 72x (Na channel), TPH2 62x (serotonin), ANO3 43x (Cl channel), ATP6V0D2 38x (H+ pump). Cells building primitive electrochemical transducer. Neural-like program in senescent ECs. Connects to Schumann/ELF coupling via voltage-gated channel stochastic resonance. | 2026-03-30 | `orthodox/figures/heterodyne/` (2 CSV) | **DEMONSTRATED** |
| S68 | MoS2 nanoparticles restore UHRF1 + Mo pathway (GSE235996) | Bulk RNA-seq from hMSCs treated with MoS2 nanoflowers (Nature Comms 2024). UHRF1 1.7x UP, DNMT1 1.3x UP, FTH1 6.5x UP (iron sequestration), SOD2 2.3x UP (mito protection), MOCS2 2.1x UP (cofactor synthesis), GPHN 1.7x UP. Mo supplementation restores the keystone (UHRF1) and activates iron management. Confirms the therapeutic prediction: restore Mo → restore UHRF1 → restore coupling. MOCS1 0.6x DOWN = feedback (enough cofactor). | 2026-03-30 | GSE235996, eye-genesis | **DEMONSTRATED** |
| S67 | Mo clock single-cell confirmation | XDH drops 0.6x, GPHN 0.3x, SUOX 0.3x — entire Mo machinery collapses in senescence. Backup: LINC02154 8.6x UP, AOX1 2.1x UP, SOD2 2.2x UP. XIST correlates with maintenance (NFE2L2 r=0.28, DNMT1 0.20, SUOX 0.21). XIST anti-correlates with FTH1 (r=-0.27): iron accumulates when Xi fails. LINC02154 anti-correlates with ILF3 (r=-0.048): hijacking confirmed. Validates MOLYBDENUM_CLOCK.md. | 2026-03-29 | cross-ref 12_BloodyEchoes | **DEMONSTRATED** |
| S66 | X-chromosome cascade: LINC02154 is X-linked + double hit | LINC02154 at chrX:13.3M — Martin's gene IS on the X chromosome. 18 disease genes change with XIST: 7 immune activators ESCAPE (FOXP3 123x, TASL 8.5x, TLR7 2.3x), 11 protective genes COLLAPSE (KDM6A 0.30x, HDAC8 0.42x, G6PD 0.61x). Double hit: immune amplification UP + tumor suppression DOWN. Explains AD drug failure in women, autoimmune bias, longevity paradox. | 2026-03-29 | `orthodox/figures/xist/` (2 CSV), `XIST_CLINICAL_MAP.md` updated | **DEMONSTRATED** |
| S65 | XIST-stratified coupling tensor + inflammatory cascade + WI-38 replication | XIST+ ECs trace=2.70 vs XIST- trace=1.81 (33% drop). XIST status as powerful as P/S condition. 355 genes explode when XIST collapses: S100A8/A9 (24,687x), TNF (80x), IL1B (8.9x), ENSG286034 (8.1x dark matter). WI-38 validates: XIST 86%->18% in ETO/IR. LINC02154 replicates in fibroblasts (12.8x). 3 dark matter genes are universal, 3 are EC-specific. | 2026-03-29 | `orthodox/figures/xist/` (3 PNG + 2 CSV), `XIST_BREAKTHROUGH.md` updated | **DEMONSTRATED** |
| S64 | XIST sex-dimorphic aging breakthrough | XIST drops 26.4%->1.7% (15.5x) in senescent ECs. Female donor ECs vs male donor PBMCs confirmed by XIST pattern. XIST x UHRF1 J=0.094 (strongest epigenetic co-occurrence). Dark matter module turns ON as XIST turns OFF. MITO operator couples more with XIST than RIBO. Mechanism: UHRF1 failure → XIST collapse → X reactivation (females only). Tomo-Seq (lab-developed) confirms spatial patterning. | 2026-03-29 | `XIST_BREAKTHROUGH.md` | **DEMONSTRATED** |
| S62 | Senescent EC Dark Matter Module | 6 co-occurring lncRNAs/novel genes in senescent ECs. LINC02154 (Martin's) + ENSG286034 (9.1% EC_Sen) + ENSG288781 (FC=9.0x) + LINC01705 (LINC02154's partner J=0.079) + LINC01811 (B_cell bridge) + ENSG254337 (most EC-specific). Module cohesion J=0.030. SASP-coupled (INHBA, FGF2, CDKN2A). No orthodox opponent — nobody has looked. 13 bounties set (DM1-DM13). | 2026-03-29 | `DARK_MATTER_MODULE.md`, `orthodox/figures/lncrna_teardown/` (3 PNG + 2 CSV) | **DEMONSTRATED** |
| S63 | lncRNA + novel gene comprehensive screen | 4,364 dark genes screened (629 LINC, 3,735 novel ENSG). Volcano, EC heatmap, enrichment distribution. LINC02154 = #1 LINC. ENSG254526 (chr11 hub) drops to 0.23x in senescence — the hub that DISAPPEARS. | 2026-03-29 | `orthodox/figures/lncrna_teardown/all_linc_de.csv`, `all_novel_de.csv` | **DEMONSTRATED** |
| S58 | LINC02154 deep dive (Martin's lncRNA, 4-quadrant analysis) | Martin's primary senescent EC lncRNA. Orthodox: OR=4.37, p=4.2e-28, NK bystander 23x. Rogue: #12/20,010 genes for senescence enrichment (top 0.1%), co-occurrence rewires from dark matter to EC identity network in senescence. Righteous: co-occurs with SASP (INHBA J=0.048, FGF2 J=0.043), tracks CDKN2A (J=0.039), NOT TE silencing. Daemon: blind ranking places it at #12, neighbors are IFN/chemokine/alarmin. All 4 quadrants converge. | 2026-03-29 | `orthodox/figures/linc02154/` (4 PNG + 2 CSV) | **DEMONSTRATED** |
| S59 | Casella bulk Genesis (GSE130727) | 37 bulk RNA-seq samples, 4 cell types, 4 senescence triggers. Gene name conversion via Ensembl REST API. trace=3.04, z=5.8. Completes the LGG lab validation set. | 2026-03-29 | eye-genesis `results/GSE130727/` | **DEMONSTRATED** |
| S60 | Per-cell junction matrix (Righteous splice stack) | 330,715 cells x 5,000 junctions extracted from P1 BAM on desync-engine (128GB). Manual BRIE2-fallback approach. Leafcutter pending (regtools build issue). | 2026-03-29 | desync-engine `splice_righteous/P1/brie2/per_cell_junctions.tsv.gz` (15 MB) | **IMPLEMENTED** |
| S61 | D3 BLIND condition clustering | 20 conditions from 5 datasets clustered by coupling tensor features. PC1=93.8% variance (trace axis). ARI=0.23 at k=4 (partial recovery). Healthy/senescent/immune separate blind. Silhouette=0.63. | 2026-03-29 | `orthodox/figures/daemon_d3/` (3 PNG + 2 files) | **DEMONSTRATED** |
| S57 | DAEMON D1: UNNAMED spectral operator discovery | Spectral clustering of 750-gene Jaccard matrix (250 operator + 500 other) at k=2..10. Result: ALL genes cluster together (ARI=0.011). BUT RIBO purity=99%, MITO purity=100% within the single cluster. NUCLEAR and GOLGI do NOT separate from OTHER. Eigengap favors k=2 (RIBO vs rest), not k=4. INTERPRETATION: The 4-operator structure is a useful simplification, not a natural partition. RIBO and MITO are real topological objects. NUCLEAR and GOLGI are biology-imposed categories that add value but don't emerge blind. Righteous classification ADDS information beyond what Daemon discovers. Honest. | 2026-03-29 | `orthodox/figures/daemon_d1/` (4 PNG + 1 CSV) | **DEMONSTRATED** |
| S56 | Per-cell-type permutation null | 200-shuffle null test on each of 8 cell types independently. ALL z > 45. Lowest: EC_Prolif z=45.0, highest: T_cell z=92.2. Real trace is 27-45x random. Operator structure is non-random in every cell type (p < 10^-200). Murder Board Attack 1 destroyed. | 2026-03-28 | `orthodox/figures/celltype_tensors/permutation_null_celltype.png`, `permutation_null_celltype.csv` | **DEMONSTRATED** |
| S55 | Per-cell-type coupling tensors + 200 bootstrap CI | K[4,4] for each of 8 cell types x 2 conditions. EC_Prolif trace=2.395 (highest), EC_Sen=1.379 (42% drop). Immune cells INCREASE coupling under SASP (bystander activation: T_cell +9%, NK +6%, DC +8%, Mono +5%). RIBO diagonal stable across all types. MITO decouples in DC/Monocyte. Bootstrap CIs confirm: immune P/S differences small but real, EC difference massive. | 2026-03-28 | `orthodox/figures/celltype_tensors/` (3 PNG + 2 CSV) | **DEMONSTRATED** |
| S54 | Genesis Splice Eigenspectrum | Coupling tensor K[4,4] computed separately on spliced, unspliced, and total layers (80,516 cells x 38,584 genes). Key: coupling drops uniformly in senescence (trace 0.75->0.66 spliced). MITO and GOLGI decouple most; RIBO stable. Unspliced 2x weaker than spliced. det(K) stable across conditions. Lambda ratio ~6 conserved. Nascent RNA already shows decoupling. | 2026-03-28 | `genesis/run_genesis_splice.py`, `orthodox/figures/genesis_splice/` | **DEMONSTRATED** |
| S53 | Eyeball calibration pass + EC deep senescence | Full diagnostic on all 10 pipeline stages. Key findings: (1) PCA elbow at PC6, n_comps=50 correct for scRNA-seq. (2) Harmony n_dims=15 captures 28% variance -- standard. (3) SASP_Core is dominant signal (d=0.47 population, d=0.74 monocytes, d=0.92 ECs). (4) Paracrine bystander effect: immune cells respond to senescent EC SASP (monocytes d=0.55). (5) TE_Silencing flat at transcriptomic level -- changes are epigenetic. (6) Senescence_Core weak -- p16/p21 are post-transcriptionally regulated. (7) Proliferation score bug (overlaps EC_Prolif annotation). Pipeline parameters are defensible. Weakness is dilution by PBMC majority; fix is stratified analysis not parameter tuning. | 2026-03-28 | `orthodox/figures/eyeball/EYEBALL_CALIBRATION_REPORT.md`, `orthodox/figures/eyeball/ec_senescence_*.png`, `orthodox/figures/eyeball/ec_senescence_effect_sizes.csv` | **DEMONSTRATED** |

---

## OPEN — MURDER BOARD CLOSURE (NB\*) 

> From [`orthodox/MURDER_BOARD_GENESIS.md`](orthodox/MURDER_BOARD_GENESIS.md). Mapped in **Consensus resolution** above. Extended tables: [`orthodox/CONSENSUS_RESOLUTION_AUDIT.md`](orthodox/CONSENSUS_RESOLUTION_AUDIT.md).

| ID | Bounty | Status | Assigned | Impact |
|----|--------|--------|----------|--------|
| NB1 | **Jaccard vs Pearson vs Spearman** — Build K\[4,4\] from all three on coculture; compare Prolif/Sen condition discrimination (Murder Board A2). | OPEN | — | HIGH |
| NB2 | **k4 resonance controls** — Bootstrap CI on k4 ratio; show k=3,5,6 do not discriminate as cleanly as k4 (Murder Board A3). | OPEN | — | HIGH |
| NB3 | **Spectra + cNMF vs operators** — Data-driven factors aligned to RIBO/MITO/NUCLEAR/GOLGI; convergent validation (Murder Board A1). | OPEN | — | MEDIUM |
| NB4 | **CELLxGENE Census EC batch** — Scripted pull + Genesis cold run; complements single-dataset blind **G3** (Murder Board A4 scale). | OPEN | — | HIGH |
| NB5 | **Murder Board closure pack** — Single deliverable folder bundling NB1+NB2 outputs under `orthodox/figures/murder_board_closure/`. | OPEN | — | MEDIUM |

---

## OPEN — WET LAB

> Assigned: user (physical lab). Claude supports with pipeline runs on resulting data.

| ID | Bounty | Status | Assigned | Impact |
|----|--------|--------|----------|--------|
| W1 | **Forced desynchronization: rotenone** — Treat proliferative ECs with rotenone (50nM, Complex I). Harvest 24h/48h/72h. Run 10x. Does cascade follow predicted order? **THE causation experiment.** | OPEN | user | CRITICAL |
| W2 | **Forced desynchronization: antimycin A** — Complex III inhibitor. Independent replication on different ETC target. | OPEN | user | HIGH |
| W3 | **Bridge gene knockdown: SMCHD1** — siRNA in senescent ECs. Does DOF collapse loosen? | OPEN | user | HIGH |
| W4 | **Bridge gene knockdown: LCN2** — siRNA Lipocalin-2. Does SASP-IFN coupling break? Iron as senescence router. | OPEN | user | HIGH |
| W5 | **lncRNA validation: CHASERR** — ASO knockdown. Measure CHD2 + FAM172A. Does thread-graph pair break? | OPEN | user | HIGH |
| W6 | **Protein confirmation: Western blots** — Chromosomal Passenger Complex (DOWN in S), GM-CSF receptor (UP), NKG2D (UP). | OPEN | user | MEDIUM |
| W7 | **NF-kB inhibition: BAY 11-7082** — Does SASP collapse? Does DOF coupling change? | OPEN | user | HIGH |
| W8 | **FTH1/iron chelation experiment** — C1 found FTH1 as cross-tissue bridge. DFO (iron chelator) on senescent ECs. Does bridge topology change? Iron metabolism as universal senescence routing. | OPEN | user | HIGH |
| W9 | **Retrieve BAM files from sequencing core** — CellRanger BAMs for 6 coculture samples. Enables SNP-based cell identity (souporcell/vireo), singlet/doublet proof (allele ratios), mtDNA heteroplasmy (mgatk). 50-200 GB/sample. | OPEN | user | CRITICAL |

---

## OPEN — COMPUTATIONAL

> Pipeline runs ~30s from cache (GPU), ~15 min from raw 10x.
> Available local datasets: EC/PBMC coculture (done), Ovarian cancer coculture (done), PBMC3k control, IFN-beta stimulated PBMCs.
> Any public GEO dataset with 10x output can be ingested directly.

| ID | Bounty | Status | Assigned | Impact |
|----|--------|--------|----------|--------|
| C2 | ~~Cross-validation: ImmuneVascular~~ | RETRACTED | -- | -- |
| C3 | ~~Cross-validation: Senescence Golgi~~ | RETRACTED | -- | -- |
| C4 | **Pre-test W1: Public rotenone scRNA-seq** — Jerber et al 2021 (HipSci, 1M+ iPSC-neurons). GCP engine room ready with STAR index. Needs correct E-MTAB-9662 accessions + FASTQ download. | OPEN | -- | CRITICAL |
| C6 | **Novel gene characterization** — Lookup 5,632 novel genes vs Ensembl/GENCODE/lncRNAdb. Build annotation table. | OPEN | -- | MEDIUM |
| C7 | **ETC complex-specific CORUM** — CORUM Complex I/III/IV/V assembly completeness, P vs S per complex. Direct protein-level desync. | OPEN | -- | HIGH |
| C8 | **Condition-split bridges** — Bridge finder on P-only and S-only GEMs separately. Which bridges rewire? | OPEN | -- | HIGH |
| C9 | **Drug target cross-reference** — Bridge genes vs DrugBank/ChEMBL. Druggable bridges? ASO-targetable lncRNAs? | OPEN | -- | HIGH |
| C11 | **Figure set for publication** — Camera-ready: method schematic, ribosome control, desync cascade, DOF heatmap, bridge network, CORUM, DoRothEA, cancer cross-val. | OPEN | -- | CRITICAL |
| C12 | **Method paper draft** — GEM paradigm, complete activation, Jaccard, enrichment, harmonic bridges, CORUM/DoRothEA. Target: Nature Methods / Genome Biology. | OPEN | -- | CRITICAL |
| C13 | **Temporal ordering from bridge decay** — Bridge scores at each desync quartile. Order of decoupling = cascade timing. | OPEN | -- | MEDIUM |
| C14 | ~~PBMC3k negative control~~ | SOLVED (S19) | -- | -- |
| C15 | ~~IFN-beta stimulation dataset~~ | SOLVED (S20) | -- | -- |
| C16 | **Unified pipeline script** — Single entry point: raw 10x -> full results. Auto-detects samples, builds co-occurrence, runs all layers, generates all plots. `python run_pipeline.py /path/to/10x/` | OPEN | -- | CRITICAL |
| C17 | **BAM-level analysis pipeline** — Given BAMs: souporcell (donor deconvolution), mgatk (mtDNA heteroplasmy), allele-ratio singlet proof. Build proven-singlet ground truth, compare to GEM-level results. | BLOCKED (W9) | -- | CRITICAL |
| C18 | **Partial pathway activation analysis** — Using decomposed sub-modules (S17), score each GEM for partial activation of each sub-module. Map sub-module landscape across conditions. | OPEN | -- | HIGH |
| C19 | **Novel program functional annotation** — Cross-reference 10 discovered programs (S17) with ENCODE, lncRNAdb, FANTOM5. What are the unannotated ENSGs? | OPEN | -- | HIGH |
| C20 | **Find low-passage primary EC scRNA-seq** — Search for published datasets using primary arterial/microvascular ECs at P0-P2 with stress/senescence. May not exist — most labs use P5+ HUVECs. | OPEN | -- | HIGH |
| C21 | **Stressed EC donor subset** — Calandrelli 2020 validated in 4 freshly isolated donor-derived ECs (2 healthy + 2 T2D). Check if those cells are in the CellxGene download. If yes, run pipeline on ONLY those cells. | OPEN | -- | HIGH |
| C22 | **BAM-level analysis: mtDNA heteroplasmy** — When BAM files retrieved (W9), run mgatk to extract per-cell mtDNA variant allele frequencies. Correlate mtDNA mutation burden with desync index. Direct mechanistic test. | BLOCKED (W9) | -- | CRITICAL |
| C23 | **BAM-level analysis: SNP deconvolution** — souporcell on BAMs to prove which GEMs are singlet EC, singlet PBMC, or doublet. Build ground-truth set, compare to GEM-level results. | BLOCKED (W9) | -- | CRITICAL |
| C24 | **Partial pathway activation as signatures** — Treat partially-activated pathways as their own programs. Gene subsets that co-occur form the real biological modules, not the full MSigDB sets. | OPEN | -- | HIGH |
| C25 | **SenCat early access** — Carlos from your lab worked on SenCat (14 primary cell types x 30 paradigms). Your friend is doing human SenCat. Get early access. Test DOF collapse across ALL 14 cell types. THE definitive cross-validation. | OPEN | user | **CRITICAL** |
| C26 | **WI-38 time course: bridge rewiring over time** — Run condition-split bridges at each day (d0-d10). Which bridges form/break during the cascade? Tests C8 with temporal resolution. | OPEN | -- | HIGH |
| C27 | **DOF as senescence clock** — Mean coupling score as a continuous senescence metric. Validate across all 9 datasets. Compare to known senescence clocks (SenCID, hUSI). | OPEN | -- | HIGH |
| C28 | **IR vs ETO vs RS coupling comparison** — WI-38 data has all 3 senescence types. Does coupling severity match insult severity? (IR > ETO > RS predicted from our data) | OPEN | -- | HIGH |
| C29 | **Brain vasculature atlas (Winkler 2024)** — 606k FACS-sorted fresh primary ECs from 68 donors, 5 CNS pathologies. Download from CellxGene. Test DOF in brain ECs. | OPEN | -- | HIGH |
| C30 | **SEA-AD Alzheimer's** — 240k nuclei, 84 donors aged 65-102, full AD spectrum. CellxGene. Test DOF in neurodegeneration (snRNA-seq caveat applies). | OPEN | -- | HIGH |
| C31 | **PD substantia nigra: annotate + split** — Download cell type + disease annotations from Broad Portal. Split PD vs control. Test DOF in dopamine neurons specifically. | OPEN | -- | CRITICAL |
| C32 | **Heart failure (DCM) atlas** — GSE183852 (Reichart). 700k nuclei, 27 healthy + 18 DCM. Test DOF in failing vs healthy hearts. snRNA-seq caveat for cardiomyocytes. | OPEN | -- | HIGH |
| C33 | **LIANA deep dive** — Run LIANA on EC-specific GEMs only (not all cell types). Analyze receptor expression changes on EC surface in P vs S. Map complete senescent EC surface proteome from transcripts. | OPEN | -- | HIGH |
| C34 | **snRNA-seq desync alternative** — For cardiomyocytes/neurons where mito transcripts are missing: compute nuclear-only assembly score using just nuclear-encoded ETC subunits. Does the nuclear side alone predict coupling? | OPEN | -- | HIGH |
| C35 | **Paper draft: structure + figures** — Target: Nature Methods (method) or Cell Systems (method + biology). Structure: GEM paradigm -> dimensional analysis -> co-occurrence -> bridges -> CORUM/DoRothEA -> cross-validation -> communication -> coupling hierarchy. | OPEN | -- | **CRITICAL** |
| C36 | **Clique analysis on Jaccard graph** — Find maximal cliques in the enrichment graph. Cliques = groups of genes that ALL co-occur tightly with each other. These are the natural "modules" without any clustering algorithm. | OPEN | -- | HIGH |
| C37 | **Cross-species eigenspectrum: fair comparison** — Re-run human at 5k genes to match tree shrew/rat. Get macaque results from GCP. Compare L1/L2 ratios at matched gene counts. | OPEN | -- | HIGH |
| C38 | **Eigenspectrum of Jaccard: what ARE the modes?** — The top 5 eigenvectors of J correspond to 5 fundamental co-occurrence patterns. Project genes onto these eigenvectors to identify what each mode represents biologically. | OPEN | -- | HIGH |
| C39 | **DOF entanglement as phase transition** — Formalize: at what desync threshold does coupling jump? Is there a critical point? Use WI-38 time course to map coupling vs desync continuously. | OPEN | -- | HIGH |
| C40 | **LIANA on EC-only GEMs** — Current LIANA run includes all cell types. Re-run on EC-classified GEMs only. What are senescent ECs specifically presenting on their surface? | OPEN | -- | HIGH |
| C41 | ~~Quick Start / `gem_analysis.h5ad` prerequisite~~ | SOLVED (S47) | -- | -- |
| C42 | **Discord cascade vs orthodox annotation — comparison** — `orthodox/objects/discord_annotated.h5ad` (constraint labels) vs `orthodox/objects/stage6_annotated.h5ad` (Leiden/marker pipeline): barcode alignment, confusion matrix / agreement, narrative for where they disagree. Unblocks interpreting both methods side-by-side. **Functional cluster:** pairs with **C59** (three-way Seurat/Monarch/GEM consensus) and **C65** (fixed-task orthodox vs GEM metrics). | OPEN | -- | **CRITICAL** |
| C43 | **SOLVED row “Where” audit** — Several SOLVED rows reference moved paths (`_archive/discord/`, `data/gem_analysis.h5ad`). Sweep S1–S46: each **Where** must resolve to canonical path or “local artifact; regenerate via …”. Fix table + add one-line pointer in project README. | OPEN | -- | MEDIUM |
| C44 | **`discord/README.md` — refresh “Current Results”** — Text still describes an early strict-only pass (80,516 cells, 43.9% assigned). Orthodox + relaxed cascade + `stage2_qc` input are now documented elsewhere; sync counts, point to `discord_annotated.h5ad` provenance. | OPEN | -- | MEDIUM |
| C45 | **Resolve overlap: OPEN C7 vs SOLVED S28** — C7 (ETC complex-specific CORUM) remains OPEN while S28 documents ETC complex assembly / C7 analysis in `methods/run_c7_etc_assembly.py`. Either mark C7 SOLVED with pointer to S28, retract C7, or redefine C7 as a *new* scope explicitly. | OPEN | -- | MEDIUM |
| C46 | **Fix DESeq2 in Docker** — pandas `nonzero` deprecation. Fix: `to_numpy().nonzero()` or pin pandas version. One-line fix in cultivator, rebuild Docker. | OPEN | -- | MEDIUM |
| C47 | **Fix CytoTRACE2 in Docker** — internal bug `'float' has no attribute 'upper'`. Version pin or input format change needed. AI potency scoring for 72k cells. | OPEN | -- | MEDIUM |
| C48 | **Fix Milo in Docker** — add `BiocManager::install("edgeR")` to Dockerfile R layer. Enables KNN-based differential abundance testing (pertpy). | OPEN | -- | MEDIUM |
| C49 | **Add scGPT foundation model to cultivator** — 33M-cell pretrained transformer. Cell embeddings + annotation without markers. Compare with CellTypist (76 types) and GEM programs (5 types). | OPEN | -- | HIGH |
| C50 | **Add SCENIC+ GRN to cultivator** — Full cis-regulatory network inference. pySCENIC or SCENIC+. Maps TF -> target gene -> chromatin accessibility prediction. | OPEN | -- | HIGH |
| C51 | **Add RNA velocity to cultivator** — STARsolo on BAMs (W9) for spliced/unspliced layers. Then scVelo dynamical mode inside Docker. Requires BAM files. | BLOCKED (W9) | -- | HIGH |
| C52 | **Add cell2location deconvolution** — Spatial deconvolution scoring without spatial data. Tests how well orthodox cell types separate vs GEM programs. | OPEN | -- | MEDIUM |
| C53 | **Add SCTransform v2 normalization** — Third normalization method (variance stabilization via pySCTransform). Compare with log-norm and scVI using scIB metrics. | OPEN | -- | HIGH |
| C54 | **Cultivator reproducibility test** — Run Docker cultivator 3x with different random seeds. Document which results change, which are stable. Parameter sensitivity analysis for the paper. **Functional cluster:** use seed list from [orthodox/config/benchmark_card.yaml](orthodox/config/benchmark_card.yaml); stabilizes orthodox-side metrics claimed in **C65**. | OPEN | -- | **CRITICAL** |
| C55 | **Orthodox vs GEM comparison figure** — Side-by-side main figure for paper. Left: cultivator outputs (55.6 min, 35 params). Right: GEM outputs (10 sec, 0 params). The money shot. | OPEN | -- | **CRITICAL** |
| C56 | **Cultivator on cross-validation datasets** — Run Docker cultivator on WI-38, donor-derived EC, Tabula Sapiens. Does orthodox find the same biology across datasets? GEM already validated on 12 datasets. | OPEN | -- | HIGH |
| C57 | **Foundation model cell type transfer** — scGPT/Geneformer to predict cell types WITHOUT any marker genes. Compare with CellTypist (76 types) and GEM programs (5 types). Ultimate "how many parameters does annotation need?" test. | OPEN | -- | HIGH |
| C58 | ~~Cross-dataset core: universal vs tissue-specific~~ | SOLVED (S51) | -- | -- |
| C59 | **Seurat-vs-GEM barcode-level consensus** — Convert RPCA cell type labels from RDS to Python. Match barcodes across Seurat (7 objects), Monarch (CellTypist), and GEM programs. Three-way confusion matrix on same 79,905 cells. | OPEN | -- | **CRITICAL** |
| C60 | **PBMC senescence standalone pipeline** — Run GEM framework on PBMC Prolif vs Senes replicates (no ECs). Test whether desync cascade fires in immune cells alone. Test whether TARBP1 hub (cross-dataset aging PBMC) replicates in experimental senescence. | OPEN | -- | HIGH |
| C61 | **Cross-dataset core: sparse Jaccard for large datasets** — Rotenone hepatocytes (43k cells) OOM at dense Jaccard. Implement chunked or sparse Jaccard to handle 40k+ cell datasets. Re-run hepatocyte core. | OPEN | -- | MEDIUM |
| C62 | **CPVL cross-tissue validation** — CPVL appears in top-5 hubs in two independent immune-containing datasets. Run targeted CPVL analysis: co-occurrence neighborhood, condition split (P vs S), pathway membership, DrugBank lookup. | OPEN | -- | HIGH |
| C63 | **Hub gene stability test** — Run dependency graph at k=5, 10, 20, 50 on all 9 datasets. Does the #1 hub change with k? If stable across k, the hub is intrinsic. If it changes, the hub is scale-dependent. | OPEN | -- | HIGH |
| C65 | **FORM orthodox apex benchmark (§E proof)** — On one **fixed public dataset**, run a named integration/QC workflow from [16_Project_Constellation/FORM_ORTHODOX_APEX_TOOLING.md](../16_Project_Constellation/FORM_ORTHODOX_APEX_TOOLING.md) Chapter 10 §F (e.g. Scanpy + scVI/scANVI or Seurat-class RPCA) vs **GEM pipeline**; log concordance metrics (ARI/NMI/ASW or agreed biology-first metrics) in `WORLDLINE.md` + figure under `orthodox/figures/`. **Pre-registered card (frozen task + seeds):** [orthodox/config/benchmark_card.yaml](orthodox/config/benchmark_card.yaml). **Functional cluster:** **C54** (seed stability), **C42**/**C59** (concordance matrices). | OPEN | -- | **CRITICAL** |

---

## OPEN — THEORY

| ID | Bounty | Status | Assigned | Impact |
|----|--------|--------|----------|--------|
| T1 | **Formalize desync index with CORUM** — ETC complex assembly scores instead of transcript ratios. | OPEN | -- | HIGH |
| T2 | **Derive phase transition threshold** — Desync ~2.5 from ETC stoichiometry, not curve fitting. | OPEN | -- | HIGH |
| T3 | **Cancer immune evasion via desync** — C1 partially tested. Need immune cells + cancer together. Pull public tumor scRNA-seq (e.g., TISCH database). | OPEN | -- | CRITICAL |
| T4 | **Biological age from desync** — DOF coupling score as scRNA-seq aging clock. | OPEN | -- | HIGH |
| T5 | **Bridge genes as drug targets** — Map which bridge disruptions halt which downstream programs. | OPEN | -- | HIGH |
| T6 | **FTH1/MALAT1 as universal senescence routing** — C1 found these bridge both EC and cancer datasets. Literature search + functional hypothesis for cross-tissue mechanism. | OPEN | -- | HIGH |

---

## OPEN — FIELD LITERATURE (senescence)

> Full curated catalog (peer-reviewed only, non-exhaustive): **[SENESCENCE_LITERATURE_GAPS.md](SENESCENCE_LITERATURE_GAPS.md)**

| ID | Bounty | Status | Assigned | Impact |
|----|--------|--------|----------|--------|
| SF1 | **Map GEM programs to literature gaps** — Table: each major GEM/SASP-related program ↔ row/theme in `SENESCENCE_LITERATURE_GAPS.md` (which open question it addresses). | OPEN | -- | HIGH |
| SF2 | **Biomarker × dataset matrix** — Senescence markers (transcript/protein) × public datasets we use × limitation (tissue, trigger, tech). Short Methods paragraph. | OPEN | -- | MEDIUM |
| SF3 | **SASP module ↔ SASP review crosswalk** — Map DoRothEA/SASP_Core modules to Nat Rev Mol Cell Biol SASP framing; document match or tension. | OPEN | -- | HIGH |
| SF4 | **Senotherapy ↔ operator tensor (hypothesis)** — Could senolytic/senomorphic intervention appear as a shift in K[4,4] coupling? What would falsify? Narrative only, not clinical claim. | OPEN | -- | MEDIUM |
| SF5 | **Quarterly literature refresh** — Add 3–5 bullets to `SENESCENCE_LITERATURE_GAPS.md` from agreed journals (*Nat Aging*, *Nat Rev Mol Cell Biol*, *Cell Metab*); update revision log. Complements **SF9** (international index scan → section 8). | OPEN | -- | LOW |
| SF6 | **Clock vs senescence disambiguation** — Which aging-clock genes overlap senescence signatures in our stacks vs stress programs? Short literature-tied note. | OPEN | -- | MEDIUM |
| SF7 | **Immune surveillance mini-memo** — Senescent cell recognition/clearance; tie to EC–PBMC LIANA (ligands presented vs immune receptors). | OPEN | -- | HIGH |
| SF8 | **Cite-the-gap for paper** — Pull 1–2 sentences per theme from `SENESCENCE_LITERATURE_GAPS.md` into C35/C12 discussion outline. | OPEN | -- | **CRITICAL** |
| SF9 | **Annual international literature scan** — Add 2–4 new DOIs from **Europe PMC**, **J-STAGE** (OA), **SciELO** (PMC-linked), or **SpringerLink** senescence or aging reviews; append to section 8 of `SENESCENCE_LITERATURE_GAPS.md`; update revision log. | OPEN | -- | LOW |
| SF10 | **Quarterly RNA API catalogue refresh** — Add or update at least one row in **`BOOK.md` §5** subsection *Remote data access* (new API, auth change, or deprecation); spot-check linked documentation URLs; append **Remote data revision log** table there. | OPEN | -- | LOW |

---

## OPEN — COMPUTATIONAL (new 2026-03-24)

| ID | Bounty | Status | Assigned | Impact |
|----|--------|--------|----------|--------|
| C58 | **Kamath PD with cell type annotations** — 434k nuclei on GCP. Download annotations from Broad Single Cell Portal. Run pipeline on DA neuron subset. Test coupling hierarchy in most OxPhos-dependent brain cells. | OPEN | -- | HIGH |
| C59 | **UHRF1 cross-species (BE-C1)** — Compare repeat element density at chr19:4.9-10.1 Mb in naked mole rat, bowhead whale, and human. Tests whether long-lived species have cleaner TE silencing HQ. | OPEN | -- | HIGH |
| C60 | **C8: Condition-split co-occurrence** — Build separate Jaccard matrices for P and S GEMs. Find gene pairs that FORM or BREAK during senescence. The rewiring map. ~30 min GPU. | OPEN | -- | HIGH |
| C61 | **Annotate 5,632 novel genes** — Batch Ensembl API query. Classify: lncRNA, pseudogene, NUMT, TE-derived, protein-coding. Map genomic locations. Find which are near TE silencing loci. | OPEN | -- | MEDIUM |
| C62 | **GBM Core Map viral lens** — 338k cells, 110 patients. Already started by Antigravity. UHRF1 8.2x up in malignant. STING1 silenced. Cancer keeps the prison guards but fires the alarm. | IN PROGRESS | Antigravity | HIGH |
| C63 | **Auto-shutdown for GCP jobs** — Every job script must `gcloud compute instances stop` when done. Cost control. No more $1.10/hr burning while we sleep. | OPEN | -- | LOW |
| C64 | **Consolidate imaging folders** — 75% duplication between old-naming and HALO-naming. Deduplicate, keep HALO names, archive old structure. | OPEN | -- | LOW |

## OPEN — THREE-DISEASE PATH (Cancer + Aging + Alzheimer's)

> **The deliverable:** One tensor, three diseases, zero parameters. See [DATA_x_METHODS.md](DATA_x_METHODS.md) for the full gap matrix.

| ID | Bounty | Status | Priority | Disease |
|----|--------|--------|----------|---------|
| TD1 | **TCGA pan-cancer coupling map** — Download open-access HTSeq counts (9,264 tumors, 33 cancer types). Run Genesis K[4,4] per cancer type. Build the pan-cancer operator coupling atlas. Cross-reference with Replogle for drug targets. | OPEN (downloadable NOW) | **CRITICAL** | Cancer |
| TD2 | **Replogle Ambrosia screen** — Which of 2,285 CRISPRi knockdowns RESTORE healthy coupling? Rank all perturbations by tensor shift toward healthy. The drug target list. | OPEN (data exists) | **CRITICAL** | All three |
| TD3 | **Heteroplasmy x coupling correlation** — Merge per-cell mtDNA mutation burden (260 variants, 241K cells) with per-cell MITO coupling. Does mtDNA damage CAUSE decoupling or is it a response? | OPEN (data exists on desync-engine) | **HIGH** | Aging + AD |
| TD4 | **Apply for ROSMAP access** — Synapse DUC. Bulk RNA at baseline + decades of cognitive outcomes. Tensor predicts AD conversion. THE longitudinal proof. | OPEN (application) | **TRUTH-BLOCKING** | Alzheimer's |
| TD5 | **Apply for Framingham dbGaP** — Blood RNA-seq + 70yr cardiovascular outcomes. Tensor predicts CVD. | OPEN (application) | **TRUTH-BLOCKING** | Cardiovascular |
| TD6 | **Download SEA-AD (CellxGene)** — 240K nuclei, 84 donors, full AD spectrum. Tensor tracks across pathological stages. | OPEN (downloadable) | **HIGH** | Alzheimer's |
| TD7 | **Download TCGA counts** — GDC open access. Immediate. | OPEN (downloadable NOW) | **CRITICAL** | Cancer |
| TD8 | **Ask Carlos for SenCat** — 14 cell types x 30 paradigms. Internal access. THE senescence atlas. | OPEN (internal) | **CRITICAL** | Aging |
| TD9 | **Dark matter cross-dataset** — Do the 6 EC_Sen lncRNAs appear in WI-38 (fibroblast senescence)? EC-specific or universal? | OPEN (data on eye-genesis) | **HIGH** | Aging |
| TD10 | **OneK1K normal reference** — 982 healthy donors. Build the "normal coupling tensor" baseline. Open access. | OPEN (downloadable) | **HIGH** | All three |
| TD11 | **COVID longitudinal** — 139 patients, 2 timepoints. Fast disease = fast validation. Tensor at T1 predicts severity at T2. | OPEN (downloadable) | **HIGH** | Validation |

---

## OPEN — EYE-SAMARKAND (Prediction Engine, 2026-03-27)

| ID | Bounty | Status | Assigned | Impact |
|----|--------|--------|----------|--------|
| E1 | **RMT denoising of Jaccard matrix** — Apply Marchenko-Pastur law to eigenvalues. Tracy-Widom test separates signal from noise. Parameter-free cutoff for real GEM programs. | OPEN | -- | **CRITICAL** |
| E2 | **Spectral graph: Fiedler + GRSBM** — Laplacian eigenvalues, algebraic connectivity, parameter-free recursive spectral bisection for community detection. Replaces Leiden resolution parameter. | OPEN | -- | **CRITICAL** |
| E3 | **Persistent homology (Ripser)** — Vietoris-Rips filtration on Jaccard distance. Persistence diagrams + Betti numbers + persistent entropy. Topological invariants of coupling structure. | OPEN | -- | HIGH |
| E4 | **Hodge decomposition (ddHodge)** — Decompose gene flow into gradient/curl/harmonic. Curl = feedback loops. Harmonic = global oscillation. On RNA velocity or desync gradient. | OPEN | -- | HIGH |
| E5 | **Forman-Ricci curvature** — Combinatorial curvature on Jaccard graph. Negative = bottleneck (bridge gene). Positive = tight module. O(E), trivial to implement. | OPEN | -- | HIGH |
| E6 | **Mutual information + transfer entropy (PIDC/TENET)** — Directional information flow between operators. On binarized data = fully parameter-free. | OPEN | -- | HIGH |
| E7 | **Download TCGA counts** — Open access HTSeq counts for 9,264 tumors. Cancer tensor development. Immediate. | OPEN | -- | **CRITICAL** |
| E8 | **Download OneK1K** — Open access PBMC scRNA-seq for 982 healthy donors. Build the "normal tensor" reference. | OPEN | -- | **CRITICAL** |
| E9 | **Apply for ROSMAP access** — Synapse DUC. Blood RNA-seq + decades of cognitive outcomes + neuropathology. GOLD STANDARD for neurodegeneration prediction. | OPEN | -- | **CRITICAL** |
| E10 | **Apply for Framingham dbGaP** — Blood RNA-seq on 2,115 + 70 years cardiovascular outcomes. GOLD STANDARD for CVD prediction. | OPEN | -- | **CRITICAL** |
| E11 | **Apply for All of Us Researcher Workbench** — Spring 2026 transcriptomics release: 10,000 individuals + EHR, diverse population. | OPEN | -- | HIGH |
| E12 | **Cameron County T2D proof of concept** — Blood RNA-seq at baseline → T2D conversion over 5 years. 24 converters + 34 controls. EXACTLY the right design: expression BEFORE disease. Small N, perfect concept. | OPEN | -- | **CRITICAL** |
| E13 | **4-operator tensor on TCGA** — Compute K[4,4] for each cancer type. Eigendecomposition per cancer. Which modes distinguish cancer types? Does Mode 2 (translation/ribo escape) dominate? | OPEN | -- | HIGH |
| E14 | **COVID-19 longitudinal validation** — 139 patients, 2 timepoints, scRNA-seq + CITE-seq. Fast disease course = fast validation cycle. Tensor at timepoint 1 predicts severity at timepoint 2. | OPEN | -- | HIGH |
| E15 | **Sheaf Laplacian on Jaccard graph** — Hansen-Ghrist construction. Detects global inconsistencies in co-expression. Custom implementation. | OPEN | -- | MEDIUM |
| E16 | **Discrete Morse theory** — Critical points of coupling landscape. Stable states (minima) vs transitions (saddles). Parameter-free cell-state decomposition. | OPEN | -- | MEDIUM |

---

## OPEN — WING-BLOCKING (must complete before any paper submits)

| ID | Bounty | Status | Assigned | Impact |
|----|--------|--------|----------|--------|
| W1 | **SenCat full stack** — Get GSE302792 data from Carlos (internal access, private until 2029). Run Genesis Eigenspectrum on all 14 cell types × 30 senescence paradigms. Map which cell types take which bifurcation branch. **BLOCKS Papers 2 and 5.** | OPEN | -- | **WING-BLOCKING** |
| W2 | **Weinreb LARRY fate prediction** — Download complete (4.7 GB). Run tensor on each clone's pre-fate snapshot. Compare predicted fate to actual clonal outcome. Beat 60% accuracy. **BLOCKS Paper 1.** | OPEN (downloading) | -- | **WING-BLOCKING** |
| W3 | **BAM spliceosome coverage shapes** — Junction extraction running via WSL samtools on 5 BAMs (P1-P3, S1-S2; S3 BAM deleted). P1 at ~550MB compressed junctions. Next: upload to GCP, install MARVEL, run spliceosome analysis. Also: molecule_info.h5 umi_type gives spliced/unspliced binary (extract_splice_fast.py written). **BLOCKS Paper 3.** | IN PROGRESS | EYE-Spliceosome | **WING-BLOCKING** |
| W4 | **One non-senescence disease dataset** — Run full stack on cardiovascular (MESA/ARIC), autoimmune, or fibrosis scRNA-seq. Prove the theory generalizes beyond senescence/cancer. **BLOCKS Paper 5.** | OPEN | -- | **WING-BLOCKING** |
| W5 | **Longitudinal prediction** — Tensor at time 1 predicts outcome at time 2. Requires ROSMAP (AD), Framingham (CVD), or Cameron County (T2D). **BLOCKS Paper 1 for top journals.** | OPEN | -- | **WING-BLOCKING** |
| W6 | **ENCODE total vs poly(A) K562** — Compute Genesis Eigenspectrum on both. Quantify what poly(A) selection hides about the Golgi operator. **BLOCKS Paper 4.** | OPEN | -- | **WING-BLOCKING** |
| W7 | **Lab presentation stack** — 3-min pitch + 15-min lab meeting + 45-min seminar + reproducibility package. Get Myriam/lab buy-in before external submission. **BLOCKS ALL papers.** | OPEN | -- | **WING-BLOCKING** |
| W8 | **DrugBank operator mapping (HALO Ambrosia)** — Map drug targets to operators. Predict tensor shifts. Rank interventions. **BLOCKS clinical utility claims.** | OPEN | -- | HIGH |

---

## OPEN — CONSTELLATION: SPLICEOSOME ATLAS

| ID | Bounty | Status | Impact |
|----|--------|--------|--------|
| SA1 | **Build operator-resolved splice diversity database** — Per-gene isoform count grouped by RIBO/MITO/NUCLEAR/GOLGI. Shannon entropy of isoform usage per operator. From GENCODE (baseline), SG-NEx nanopore (observed), BAM P vs S (disease). Waiting on current agents to return data. | OPEN (agents running) | **CRITICAL** |
| SA2 | **Build weird transcript registry** — Reads-per-UMI anomalies from molecule_info.h5. Genes that resist sequencing = structured/circular/truncated RNA candidates. Flag per operator. | OPEN (agent running) | **CRITICAL** |
| SA3 | **Map splice-operator coupling** — Which retained introns convert mRNA → lncRNA per operator? Which alternative poly(A) events change 3'UTR length → change miRNA coupling → change Jaccard hub status? | OPEN | HIGH |
| SA4 | **Connect to Ambrosia drug vectors** — Do splice-targeting drugs (e.g., spliceostatin, pladienolide, risdiplam) produce operator-specific effects? Which operator's genes are most affected by global splice inhibition? | OPEN | HIGH |
| SA5 | **Design Constellation website architecture** — Preprint + GitHub + u-os.dev deployment. Analysis layer only (no raw data redistribution). Interactive coupling tensor explorer, drug vector calculator, splice atlas browser. Deploy simultaneously with bioRxiv preprint. | OPEN | HIGH |
| SA6 | **Legal/license review** — Confirm all source database licenses permit derived analysis hosting. GENCODE CC, ENCODE open, GEO public domain, SG-NEx open. Document attribution requirements. | OPEN | MEDIUM |

---

## OPEN — DAEMON / INFRASTRUCTURE (new 2026-03-24)

| ID | Bounty | Status | Assigned | Impact |
|----|--------|--------|----------|--------|
| D1 | **HALO + WING + Daemon protocol** — Formalize as vault-level README. Living documents, production criteria, daemon architecture. The operating system for the lab. | OPEN | -- | HIGH |
| D2 | **Wire Jules to u-os.dev** — Overnight GCP jobs. Jules claims bounties from mailbox, runs on engine, posts results. Stop babysitting SSH. | OPEN | -- | MEDIUM |
| D3 | **Image annotation daemon** — Auto-classify new images by condition as they appear in imaging/. Read folder name, apply HALO naming, log to annotations file. | OPEN | -- | MEDIUM |
| D4 | **Mimetic shield protocol (Project 11)** — Design information containment that works even through this interface. Public worldline, private consciousness. KETER-level work invisible in public trace. | OPEN | -- | **CRITICAL** |

---

## OPEN — GENESIS SIGNAL / QIP (2026-03-31)

| ID | Bounty | Status | Assigned | Impact |
|----|--------|--------|----------|--------|
| GC1 | **Restore LINC01235 in HeLa** — CRISPRa activate from endogenous locus. Measure K[4,4] before/after. Does antenna installation change coupling? | OPEN | — | **CRITICAL** |
| GC3 | **Schumann-shielded cell culture** — Grow cells in mu-metal room. Compare K to unshielded. Does the signal exist? | OPEN | — | HIGH |
| GC5 | **Pan-cancer antenna survey (TCGA)** — Run antenna profiler on all TCGA tumor types. Does every cancer spoof Address 00? | OPEN | — | HIGH |
| GC6 | **SEA-AD brain antenna** — Download 1.4M nuclei (84 donors age 65-102). Run XIST-stratified coupling. Does K predict AD pathology? | OPEN | — | HIGH |
| GC7 | **OneK1K PBMC antenna** — 982 donors (416M/565F, age 19-97). Largest single-cohort XIST aging trajectory. | OPEN | — | HIGH |
| GC8 | **SenCat internal access** — Ask Carlos for GSE302792 processed counts. 14 cell types x 30 paradigms. | OPEN | — | HIGH |
| QIP1 | **Menstrual cycle coupling tensor** — Longitudinal scRNA-seq across 29.5 days in one female donor. Does trace(K) oscillate? | OPEN | — | **CRITICAL** |
| QIP2 | **CAIS immune phenotype** — Clinical collab to measure autoimmune markers in 46,XY females. Does QIP sex follow chromosomes or hormones? | OPEN | — | **CRITICAL** |
| QIP3 | **Exogenous hormones → XIST** — Measure XIST expression before/after OCP, HRT, or testosterone therapy. Does it shift? | OPEN | — | HIGH |
| QIP4 | **Dirac coupling tensor** — Replace Jaccard K with gradient-based K. Does Z3 structure emerge in the differential? | IN PROGRESS | Claude Code | HIGH |
| QIP5 | **Naked mole rat antenna profile** — Census or direct data. Does antenna quality (not just architecture) predict 30yr lifespan? | OPEN | — | MEDIUM |
| QIP6 | **LINC01235/02154/01705 structural validation** — SHAPE-seq or DMS-seq to confirm GNRA tetraloops and Z3 braid. | OPEN | — | HIGH |

## SOLVED — GENESIS SIGNAL SESSION (2026-03-30 to 2026-03-31)

| ID | Bounty | One-line answer | Date | Where | Tier |
|----|--------|-----------------|------|-------|------|
| S88 | GTEx sex-dimorphic antenna | 2,205 dimorphic gene x tissue combos. LINC02154 cerebellum gap=0.451. SERPINE1 F-faster in 73% tissues. XIST mid-life dip + 70s rebound. | 2026-03-30 | GTEx_antenna_*.csv on eye-genesis | DEMONSTRATED |
| S89 | GESTALT 82-donor aging trajectory | LINC01705 FC=1.93 males, PTGS2 opposite by sex, RIBO-GOLGI anti-coupling female-specific. Composite age score rho=0.35. | 2026-03-30 | GESTALT_*.csv on eye-genesis | DEMONSTRATED |
| S90 | WI-38 X-reactivation: passage vs damage | Passage: X-linked genes ESCAPE (KDM5C 1.63x). Damage: ALL X-linked genes CO-DOWNREGULATE. Two fundamentally different programs. | 2026-03-30 | WI38_X_deep_*.csv on eye-genesis | DEMONSTRATED |
| S91 | Mouse cross-species antenna (230K cells) | Kdm6a halves in male mice only. S100a8 inverted vs human. Architecture conserved 90 Mya, inflammatory arm species-specific. 6/8 tissues DIVERGE by sex with age. | 2026-03-30 | Mouse_*.csv on eye-genesis | DEMONSTRATED |
| S92 | GESTALT circRNA sex dimorphism | LINC02154 locus circRNAs: male rho=+0.43, female flat. 47K unique circRNAs. Age-correlated circRNAs sex-non-overlapping. | 2026-03-30 | GESTALT_circRNA_*.csv | DEMONSTRATED |
| S93 | Trophoblast sex dimorphism (98K cells) | IDO1 172x, HLA-G 15x, SERPINE1 8.5x, ESR1 78x girl>boy. Boy placenta: S100A8 7x, CDKN2A 7x. Two completely different organs. 9x SERPINE1 is NOVEL at single-cell resolution. | 2026-03-31 | Trophoblast_sex_dimorphic.csv, BOY_MOM_GIRL_MOM.md | DEMONSTRATED |
| S94 | Intersex dosage ladder | 9/10 sex chromosome variants match X-dosage prediction. Turner paradox explained by U-curve. Triple X = 2.5x female SLE. | 2026-03-31 | HALO_DUALITYSPECTRUM.md, literature review | DEMONSTRATED |
| S95 | XIST-stratified coupling spectrum (500K cells) | XIST-low=0.29, Male=0.40, XIST-high=0.59 across 8 tissues. Universal ordering. KDM6A 127x by XIST status. | 2026-03-31 | Intersex_*.csv on eye-genesis | DEMONSTRATED |
| S96 | Fetal antenna characterization | trace=2.45, RIBO 83%, antenna ABSENT, maintenance MAXIMUM. Loci open but silent. SOX11 silenced during installation. Fetal vs cancer: MKI67 72% vs <3%, DNMT1 92% vs 21%. | 2026-03-31 | Fetal_antenna_*.csv (18 files) | DEMONSTRATED |
| S97 | Prenatal Census (4.9M fetal cells) | Fetal XIST 60-98%. Trophoblast XIST 37% (deliberately low). Placental XIST OSCILLATES across Carnegie stages. UHRF1 52x lower in trophoblast than fetal soma. | 2026-03-31 | Prenatal_antenna_*.csv | DEMONSTRATED |
| S98 | GTEx prediction: pooled vs sex-stratified | 96% of gene x tissue combos show resolution gain. Brain Putamen R² gain +0.54. 90% tissue x age pairs sex-discordant for antenna address. | 2026-03-31 | GTEx_prediction_*.csv | DEMONSTRATED |
| S99 | Casella 4 cell types x triggers | HRAS crashes XIST 20x, LINC02154 explodes 105x. TET2/3 vs UHRF1/DNMT1 epigenetic mirror by fetal sex in trophoblast. | 2026-03-31 | Casella_antenna_deep.csv | DEMONSTRATED |
| S100 | SERPINE1 audit (corrected) | Men start 18% higher (20s). F/M ratio FLIPS from 0.85 to 1.78 by 70s. WI-38 97% is culture artifact (<6% in vivo fetal). Girl trophoblast 9x more than boy. Mouse is male-biased (inverted). | 2026-03-31 | SERPINE1_AUDIT.md | OBSERVATION |
| S101 | Sex-specific antenna modulators | 30+ interventions cataloged. Aspirin: reduces stroke in F, MI in M. AKG/VitC: KDM6A cofactors (XX-amplified). Iron-Mn competition at SOD2. Thymoquinone degrades UHRF1. | 2026-03-31 | QUANTUM_IDENTITY_PATTERNS.md | OBSERVATION |
| S102 | Cross-cultural behavioral mapping | Mental rotation d=1.0, E-S d=0.9 (57 countries). Gender equality paradox. 12% women tetrachromat (Xq28 mosaic). Circadian 6-min sex offset. | 2026-03-31 | HALO_QIP.md | OBSERVATION |

## DOCUMENTS PRODUCED (2026-03-30 to 2026-03-31)

| Document | Location | Classification |
|----------|----------|---------------|
| HALO_GENESISSIGNAL.md Sections I-VI | 10_Project_DiscordIntoSymphony/ | APOLLYON |
| QUANTUM_IDENTITY_PATTERNS.md | 10_Project_DiscordIntoSymphony/ | APOLLYON |
| HALO_DUALITYSPECTRUM.md | 10_Project_DiscordIntoSymphony/ | APOLLYON-KETER-THAUMIEL |
| BOY_MOM_GIRL_MOM.md | 10_Project_DiscordIntoSymphony/ | APOLLYON |
| SERPINE1_AUDIT.md | 10_Project_DiscordIntoSymphony/ | EUCLID |
| GAP_MATRIX.md | 10_Project_DiscordIntoSymphony/ | EUCLID |
| HALO_QIP.md | 13_Project_MemoryOfMind/ | APOLLYON-KETER-THAUMIEL |

## DATA PRODUCTS (73 CSV files on eye-genesis)

All at `~/lgg_genesis/results/` on GCP VM eye-genesis (us-central1-a):
- GTEx: 7 files (antenna, dimorphic, brain, prediction x4)
- GESTALT: 12 files (age correlation, deep x6, circRNA x6)
- WI-38 X: 9 files (gene survey, XIST-stratified, reactivation, LINC02154, immune, passage)
- Fetal: 18 files (expression, coupling, installation, XIST, operators, maintenance, loci, cancer)
- Prenatal Census: 5 files (stages, trajectory, young adult, fetal expression, trophoblast)
- Mouse: 6 files (tissue sex age, deep x5)
- Intersex: 7 files (categories, diseases, XIST dosage, antenna, K tensor, summary, escape)
- Trophoblast: 1 file (sex dimorphic 95 genes)
- Casella: 1 file (antenna deep)
- Unified: 2 files (master, deep analysis)
- Other: 5 files

---

## RETRACTED / WONTFIX

| ID | What | Why |
|----|------|-----|
| R1 | ETC stoichiometry as binary death marker | Ratio SHIFT matters, not absolute level |
| R2 | Leiden clustering on program space | Thread signature IS the clustering |
| R3 | Diffusion pseudotime | Parametric. Replaced by desync gradient. |
| R4 | Dijkstra bridge finding | Distance=0 collapse. Replaced by harmonic Jaccard/enrichment. |
| R5 | Standard doublet removal | Kills real biology (1,625 macrophage-EC GEMs). |
| R6 | O1-O9 from March 19 board | Superseded by GEM paradigm. |
| R7 | C2: ImmuneVascular cross-val | qPCR data only, not scRNA-seq. |
| R8 | C3: Senescence Golgi cross-val | qPCR data only, not scRNA-seq. |