Add dataset README (reproduction guide)

README.md

@@ -10,80 +10,132 @@ tags:
 - zarr
 ---
 
-# pevo-msa-grch38-19way (dataset)
+# pevo-msa-grch38-19way (dataset Hub)
 
-**EN:** Data and thesis reproduction artifacts for primate 19-way / 10-primate MSA variant-effect modeling.  
-**中文：** 灵长类 19-way / 10 物种 MSA 变异效应建模的数据与论文复现产物。
+**EN:** Training data, thesis tables, and reproduction artifacts for primate MSA variant-effect modeling.  
+**中文：** 灵长类 MSA 变异效应建模的**训练数据**与**论文复现数据**（不含模型权重）。
 
 | | |
 | --- | --- |
-| **Code (training & eval)** | [github.com/jasperyeoh/pevo-msa-primate-genomics](https://github.com/jasperyeoh/pevo-msa-primate-genomics) |
+| **Code & experiment map** | [github.com/jasperyeoh/pevo-msa-primate-genomics](https://github.com/jasperyeoh/pevo-msa-primate-genomics) — see `EXPERIMENT_MAP.md` |
 | **Model checkpoints** | [jasperyeoh2/pevo-msa-mlm-19way](https://huggingface.co/jasperyeoh2/pevo-msa-mlm-19way) |
-| **Layout map (full tree)** | `docs/HUGGINGFACE_DATASET_LAYOUT.md` in this repo |
+| **Full file tree** | `docs/HUGGINGFACE_DATASET_LAYOUT.md` (this repo) |
 | **Upload audit** | `docs/UPLOAD_AUDIT.md` on GitHub |
 
+---
+
+## Dataset vs model — which Hub?
+
+```text
+  GitHub (code + small CSV)          THIS REPO (data)              MODEL REPO (weights)
+  ─────────────────────────          ────────────────              ────────────────────
+  scripts, configs, thesis CSV  →    Zarr, MAF, thesis_repro  +    base_models/*/checkpoint-best
+                                     windows, conservation          configs/, inventory/
+```
+
+---
+
 ## Quick download / 快速下载
 
 ```bash
 pip install -U huggingface_hub
 
-# Everything (large — 100GB+ depending on includes)
-hf download jasperyeoh2/pevo-msa-grch38-19way --repo-type dataset --local-dir data_hf
-
-# Paper tables only (~hundreds of MB)
+# A) Paper reproduction only (~hundreds of MB)
 hf download jasperyeoh2/pevo-msa-grch38-19way --repo-type dataset --local-dir data_hf \
   --include 'thesis_repro/**' 'docs/**'
 
-# Training Zarr (per-chromosome tar bundles — extract after download)
+# B) Training corpus (Zarr per-chromosome tar — extract after download)
 hf download jasperyeoh2/pevo-msa-grch38-19way --repo-type dataset --local-dir data_hf \
   --include 'new_dataset_27feb/msaasr_dataset_10primates_full/**'
+
+# C) Full snapshot (100GB+)
+hf download jasperyeoh2/pevo-msa-grch38-19way --repo-type dataset --local-dir data_hf
 ```
 
-## Main folders / 主要目录
+---
+
+## `thesis_repro/` — paper tables & scores
+
+Restore into a Git clone (paths expected by `thesis_manu_v2_6/scripts/`):
+
+```bash
+git clone https://github.com/jasperyeoh/pevo-msa-primate-genomics.git
+cd pevo-msa-primate-genomics
+hf download jasperyeoh2/pevo-msa-grch38-19way --repo-type dataset --local-dir ../data_hf \
+  --include 'thesis_repro/**'
+
+mkdir -p phylo_msa1/outputs/vep_all_base_models
+cp ../data_hf/thesis_repro/phylo_msa1_outputs/VEP_*.csv phylo_msa1/outputs/
+cp ../data_hf/thesis_repro/phylo_msa1_outputs/vep_all_base_models/*.parquet \
+   phylo_msa1/outputs/vep_all_base_models/
 
-| Prefix | Contents |
+cp -a ../data_hf/thesis_repro/dna_foundation_benchmark_results_final \
+      downstream_tasks/dna_foundation_benchmark/results_final
+
+mkdir -p thesis_manu_v2_6/artifacts
+cp ../data_hf/thesis_repro/thesis_manu_v2_6_artifacts/* thesis_manu_v2_6/artifacts/
+```
+
+**Tip:** Many **summary CSVs** are also on GitHub now — `git clone` alone is enough for Tier-A LaTeX regeneration (see `EXPERIMENT_MAP.md`).
+
+| `thesis_repro/` path | Used for |
 | --- | --- |
-| `thesis_repro/` | VEP aggregate CSVs, per-run score parquets, Feng benchmark **summary** CSVs, manuscript bootstrap artifacts |
-| `new_dataset_27feb/msaasr_dataset_10primates_full/bundles/` | **24×** `msaasr.zarr.bundle_*.tar` + README — restore full training Zarr |
-| `new_dataset_27feb/gpnmsa_1024bp_pipeline/` | 1024bp window parquets, VEP ablation refs (large VCFs may be `.partNNNNofMMMMM`) |
-| `window_workspace/neutral_model_10primates/` | **`merged_10primates_training.maf`** (~23GB) for neutral-model / window pipeline |
-| `conservation_pipeline/scores/` | Per-chromosome phastCons / phyloP `.wig` tracks |
-| Root `train_*.parquet`, `windows_10primates_*.parquet` | Multi-resolution window snapshots |
-| `phylo_msa1/outputs/strictv2_frozen_1000_eval_20260506/` | Matched-width strict-v2 eval (May 2026) |
+| `phylo_msa1_outputs/VEP_*.csv` | VEP leaderboards, strict-v2 tables (`tab:vep_*`, `tab:strictv2_*`) |
+| `phylo_msa1_outputs/vep_all_base_models/*.parquet` | Per-checkpoint VEP vectors, ablation registry |
+| `dna_foundation_benchmark_results_final/` | Feng AUROC tables/figures (`tab:feng_*`, `fig:feng_*`) |
+| `thesis_manu_v2_6_artifacts/` | Matched-width strict-v2 bootstrap (May 2026) |
 
-**Not in this dataset (use GitHub code + model Hub or cluster):** raw MAF stage trees (`maf_workspace/`), Songlab `variants_and_preds.parquet`, full gnomAD-tail 1M score panels, `msaasr_dataset_v2/`.
+---
 
-## Restore into a code clone / 恢复到 Git 克隆目录
+## Main data prefixes / 主要数据目录
 
-After `git clone` and `hf download`:
+| Prefix | Contents | Size |
+| --- | --- | --- |
+| `thesis_repro/` | Paper CSV + VEP parquets + Feng summaries | small |
+| `new_dataset_27feb/msaasr_dataset_10primates_full/bundles/` | 24× `msaasr.zarr.bundle_*.tar` | ~132 GB total |
+| `new_dataset_27feb/gpnmsa_1024bp_pipeline/` | 1024bp window parquets, VEP ablation (VCFs may be chunked) | mixed |
+| `window_workspace/neutral_model_10primates/` | `merged_10primates_training.maf` | ~23 GB |
+| `conservation_pipeline/scores/` | phastCons / phyloP `.wig` per chromosome | moderate |
+| Root `train_*.parquet`, `windows_10primates_*.parquet` | Multi-resolution window panels | moderate each |
+| `msaasr.zarr/` (legacy) | Old chunk tree at repo root | prefer **bundles** |
 
-```bash
-REPO=pevo-msa-primate-genomics
-DATA=data_hf
+---
 
-mkdir -p $REPO/phylo_msa1/outputs/vep_all_base_models
-cp $DATA/thesis_repro/phylo_msa1_outputs/VEP_*.csv $REPO/phylo_msa1/outputs/
-cp $DATA/thesis_repro/phylo_msa1_outputs/vep_all_base_models/*.parquet $REPO/phylo_msa1/outputs/vep_all_base_models/
+## Thesis figure/table index (data side)
 
-mkdir -p $REPO/downstream_tasks/dna_foundation_benchmark
-cp -a $DATA/thesis_repro/dna_foundation_benchmark_results_final $REPO/downstream_tasks/dna_foundation_benchmark/results_final
+| LaTeX label | Data on this Hub |
+| --- | --- |
+| `tab:vep_*`, `tab:strictv2_*`, `tab:vep_leaderboard_full` | `thesis_repro/phylo_msa1_outputs/VEP_*.csv` (+ parquets) |
+| `fig:feng_*`, `tab:feng_*` | `thesis_repro/dna_foundation_benchmark_results_final/` |
+| `fig:gnomad-tail-*` | CSV on **GitHub**; score parquets cluster-only |
+| Retrain from scratch | `bundles/*.tar` → extract `msaasr.zarr/` |
 
-mkdir -p $REPO/thesis_manu_v2_6/artifacts
-cp $DATA/thesis_repro/thesis_manu_v2_6_artifacts/* $REPO/thesis_manu_v2_6/artifacts/ 2>/dev/null || true
-```
+Checkpoints for scoring: **model Hub** `base_models/full_training_*/checkpoint-best/`.  
+Full mapping: GitHub `EXPERIMENT_MAP.md`.
 
-See `thesis_repro/README.md` and GitHub `REPRODUCE.md` for the full workflow.
+---
 
-## MSAASR Zarr bundles / 训练语料恢复
+## MSAASR Zarr restore / 训练语料恢复
 
 1. Download `new_dataset_27feb/msaasr_dataset_10primates_full/bundles/msaasr.zarr.bundle_*.tar`
-2. Follow `README_MSAASR_BUNDLES.md` in that folder to extract into `msaasr.zarr/`
-3. Prefer bundles over the legacy root `msaasr.zarr/` chunk tree (millions of small files)
+2. Read `README_MSAASR_BUNDLES.md` in that folder
+3. Extract to `new_dataset_27feb/msaasr_dataset_10primates_full/msaasr.zarr/`
+4. Do **not** rely on root `msaasr.zarr/` chunk trees (millions of files; slow)
+
+---
 
-## Citation / 引用
+## Regenerate thesis LaTeX (minimal)
 
-If you use this dataset, cite the thesis/preprint and link this dataset URL. Pin the dataset commit hash from the Hub **History** tab for exact reproduction.
+```bash
+cd pevo-msa-primate-genomics/thesis_manu_v2_6/scripts
+python3 gen_strictv2_ci_artifacts.py
+python3 gen_feng_combined_seq_and_l7_auroc.py
+cd .. && latexmk -pdf main.tex
+```
+
+---
 
-## License
+## Citation & license
 
-MIT (see dataset card). Third-party references (gnomAD, ClinVar, EPO alignments) remain under their original terms.
+Cite the thesis/preprint and pin this dataset commit from **History**.  
+License: MIT (dataset card). gnomAD / ClinVar / EPO data remain under their original terms.