Fix loading instructions and add dataset metadata

- Fix repo_type="dataset" missing from hf_hub_download (was causing 404)
- Fix variable name collision (ad overwrote import)
- Fix import consistency (import anndata, not as ad)
- Add matrix format info (Sparse CSC, float32)
- Add data splits table (train/valid/test counts)
- Document missing .var columns (width, cre_idx)
- Add cre_class possible values
- Fix title typo (human-catlas-atac -> human-atac-catlas)

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

README.md

@@ -10,7 +10,7 @@ size_categories:
 - 1M<n<10M
 ---
 
-# human-catlas-atac-data
+# human-atac-catlas-data
 
 ## Dataset Summary
 This dataset provides a binary accessibility matrix of candidate Cis-Regulatory Elements (cCREs) across human cell types. It is derived from the CATlas project (https://decoder-genetics.wustl.edu/catlasv1/catlas_humanenhancer/#!/). The matrix identifies which genomic regions are accessible (1) or inaccessible (0) across 204 distinct cell types. The data is derived from https://decoder-genetics.wustl.edu/catlasv1/humanenhancer/data/cCRE_by_cell_type/. The hg38 genome build was used.
@@ -24,6 +24,14 @@ This dataset provides a binary accessibility matrix of candidate Cis-Regulatory
 ### AnnData Object Dimensions
 - **n_obs (Cell Types):** 204
 - **n_vars (cCREs):** 1,121,319
+- **Matrix format:** Sparse CSC (scipy.sparse.csc_matrix), float32
+
+### Data Splits
+| Split | Count |
+|-------|-------|
+| train | 977,014 |
+| valid | 71,798 |
+| test | 72,507 |
 
 ### Data Fields
 - **`.X`**: A sparse binary matrix where rows are cell types and columns are genomic regions (cCREs).
@@ -31,22 +39,25 @@ This dataset provides a binary accessibility matrix of candidate Cis-Regulatory
   - `cell type`: The descriptive name of the human cell type/cluster.
 - **`.var` (Genomic Feature Metadata):**
   - `chrom`, `start`, `end`: Genomic coordinates (hg38).
-  - `cre_class`: Classification of the regulatory element.
-  - `in_fetal` / `in_adult`: indicators of activity in developmental stages.
+  - `cre_class`: Classification of the regulatory element (Promoter, Promoter Proximal, or Distal).
+  - `in_fetal` / `in_adult`: Indicators of activity in developmental stages.
   - `cre_module`: Associated regulatory module.
+  - `width`: Width of the cCRE region in base pairs.
+  - `cre_idx`: Index of the cCRE.
   - `enformer_split`: Overlap with the data splits used for training the Enformer model.
-  - `split`: Splits used for downstream modeling (training/validation/test).
+  - `split`: Splits used for downstream modeling (train/valid/test).
 
 ## Usage
 
 ```python
-import anndata as ad
+import anndata
 from huggingface_hub import hf_hub_download
 
 file_path = hf_hub_download(
-    repo_id="Genentech/human-atac-catlas-data", 
-    filename="data.h5ad"
+    repo_id="Genentech/human-atac-catlas-data",
+    filename="data.h5ad",
+    repo_type="dataset"
 )
 
-ad = anndata.read_h5ad(file_path)
-```
+adata = anndata.read_h5ad(file_path)
+```