|
|
--- |
|
|
dataset_info: |
|
|
features: |
|
|
- name: Entry |
|
|
dtype: string |
|
|
- name: domain |
|
|
dtype: string |
|
|
- name: Sequence |
|
|
dtype: string |
|
|
- name: labels |
|
|
dtype: int64 |
|
|
splits: |
|
|
- name: train |
|
|
num_bytes: 101549794 |
|
|
num_examples: 251620 |
|
|
- name: valid |
|
|
num_bytes: 2025747 |
|
|
num_examples: 5000 |
|
|
- name: test |
|
|
num_bytes: 2019305 |
|
|
num_examples: 5000 |
|
|
download_size: 101889618 |
|
|
dataset_size: 105594846 |
|
|
configs: |
|
|
- config_name: default |
|
|
data_files: |
|
|
- split: train |
|
|
path: data/train-* |
|
|
- split: valid |
|
|
path: data/valid-* |
|
|
- split: test |
|
|
path: data/test-* |
|
|
--- |
|
|
|
|
|
UniProt search (7/22/2025): * (Swiss-Prot reviewed entries only) download TSV with Taxonomic lineage (Ids), sequence, and length. |
|
|
|
|
|
Curation steps: |
|
|
|
|
|
- Filtered for sequence length >=20 and <=2048 |
|
|
- Extract Taxonomic IDs (domain, kingdom, phylum, class, order, family, genus, species) from taxonomic_lineage_ids column |
|
|
- Keep entry, domain, and sequence information |
|
|
- Drop N/A |
|
|
- CD-HIT using Docker at 80% similarity threshold and n=5 |
|
|
- Keep representative domains |
|
|
- Create labels based on domain |
|
|
- Remove domains/examples with fewer than 100 samples |
|
|
- Stratified split: first get test set (5,000), then eval (5,000), rest is train |
