purvansh01 commited on
Commit
9c2d63f
·
verified ·
1 Parent(s): 404c8f2

docs: dataset card - configs, quick start, intended use, version history

Browse files
Files changed (1) hide show
  1. README.md +47 -0
README.md CHANGED
@@ -3,6 +3,27 @@ license: other
3
  task_categories: [text-classification]
4
  tags: [genomics, regulatory-elements, ENCODE, cCRE, dna]
5
  pretty_name: ENCODE regulatory elements (T cell, K562, GM12878), v2.1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
6
  ---
7
  # ENCODE regulatory elements - confound-aware tables (v2.1)
8
 
@@ -10,6 +31,19 @@ pretty_name: ENCODE regulatory elements (T cell, K562, GM12878), v2.1
10
  biosamples, each with three negative strategies. Built for the Interpretable Regulatory Genomics project
11
  (github.com/PxA-Labs/interpretable-regulatory-genomics).
12
 
 
 
 
 
 
 
 
 
 
 
 
 
 
13
  ## Biosamples (identities checked against ENCODE file metadata)
14
  | key | ENCODE file | biosample |
15
  |---|---|---|
@@ -53,3 +87,16 @@ similarity to any training sequence is >= 0.3 (report metrics with and without t
53
  ## Sources and licences
54
  ENCODE (open data, https://www.encodeproject.org/), reference GRCh38. Check ENCODE data-use terms before redistribution.
55
  Source snapshot: git SHA ac3e4156be59a868885099c707400ee107b0ce53, seed 42.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3
  task_categories: [text-classification]
4
  tags: [genomics, regulatory-elements, ENCODE, cCRE, dna]
5
  pretty_name: ENCODE regulatory elements (T cell, K562, GM12878), v2.1
6
+ size_categories: ["100K<n<1M"]
7
+ configs:
8
+ - config_name: k562_gc_matched
9
+ default: true
10
+ data_files: k562_gc_matched.parquet
11
+ - config_name: k562_random_unmatched
12
+ data_files: k562_random_unmatched.parquet
13
+ - config_name: k562_flanking
14
+ data_files: k562_flanking.parquet
15
+ - config_name: gm12878_gc_matched
16
+ data_files: gm12878_gc_matched.parquet
17
+ - config_name: gm12878_random_unmatched
18
+ data_files: gm12878_random_unmatched.parquet
19
+ - config_name: gm12878_flanking
20
+ data_files: gm12878_flanking.parquet
21
+ - config_name: tcell_gc_matched
22
+ data_files: tcell_gc_matched.parquet
23
+ - config_name: tcell_random_unmatched
24
+ data_files: tcell_random_unmatched.parquet
25
+ - config_name: tcell_flanking
26
+ data_files: tcell_flanking.parquet
27
  ---
28
  # ENCODE regulatory elements - confound-aware tables (v2.1)
29
 
 
31
  biosamples, each with three negative strategies. Built for the Interpretable Regulatory Genomics project
32
  (github.com/PxA-Labs/interpretable-regulatory-genomics).
33
 
34
+ ## Quick start
35
+ ```python
36
+ from datasets import load_dataset
37
+ ds = load_dataset("PxALabs/encode-regulatory-elements", "k562_gc_matched") # private repo: log in with `hf auth login` first
38
+ train = ds["train"].filter(lambda r: r["split"] == "train")
39
+ ```
40
+ Each table has one `train` split in the Hugging Face sense; use the `split` column (train / val / test, by chromosome) for modelling.
41
+ Report test metrics both on all rows and with `homology_flag == False`.
42
+
43
+ ## Intended use
44
+ Benchmarking sequence-to-regulatory-element classifiers (k-mer models, CNNs) with explicit control of composition shortcuts
45
+ (GC content, CpG). Not intended for clinical or variant-interpretation use. Negatives are not validated non-regulatory elements.
46
+
47
  ## Biosamples (identities checked against ENCODE file metadata)
48
  | key | ENCODE file | biosample |
49
  |---|---|---|
 
87
  ## Sources and licences
88
  ENCODE (open data, https://www.encodeproject.org/), reference GRCh38. Check ENCODE data-use terms before redistribution.
89
  Source snapshot: git SHA ac3e4156be59a868885099c707400ee107b0ce53, seed 42.
90
+
91
+ ## Baseline results
92
+ Reference results, protocols and caveats are in the project repository (`docs/results/`, run records under
93
+ `experiments/runs/`), not in this card, so that numbers have a single source of truth. Random negatives inflate scores;
94
+ use `gc_matched` as the minimum standard.
95
+
96
+ ## Version history
97
+ - **v2.1** (2026-11-18): deduplicated tables, homology flags, canonical chromosome split, hash manifest.
98
+ - **v1** (project start): tables named `k562_*` that were actually T-cell elements; superseded, do not use.
99
+
100
+ ## Citation
101
+ Cite the project repository (github.com/PxA-Labs/interpretable-regulatory-genomics) and the ENCODE / SCREEN resources
102
+ following the citation guidance at https://www.encodeproject.org/help/citing-encode/ .