File size: 2,790 Bytes
fd65baa |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 |
---
pretty_name: SABmark
language:
- en
tags:
- protein
- sequence-alignment
- structural-biology
task_categories:
- other
configs:
- config_name: twi
description: Twilight subset (<25% identity)
data_files:
- split: test
path: twi.jsonl
- config_name: sup
description: Superfamilies subset
data_files:
- split: test
path: sup.jsonl
- config_name: twi_fp
description: Twilight with added false positives
data_files:
- split: test
path: twi_fp.jsonl
- config_name: sup_fp
description: Superfamilies with added false positives
data_files:
- split: test
path: sup_fp.jsonl
- config_name: all
description: All subsets combined
data_files:
- split: test
path: all.jsonl
---
# SABmark (Hugging Face Port)
Benchmark for protein sequence alignment under remote homology.
## Subsets
- `sup`: low-to-intermediate identity
- `twi`: very low identity (<25%)
- `sup_fp`: superfamilies with added false positives
- `twi_fp`: twilight with added false positives
- `all`: union of the above
## Features
- `pair_id`, `group_id`, `set_name`
- `seq1_id`, `seq2_id`, `seq1`, `seq2`
- `ref_alignment`: list of index pairs `[[i, j], ...]` (0-based)
- `percent_identity`, `scop_labels`, `meta`
## Usage
```python
from datasets import load_dataset
ds = load_dataset("DeepFoldProtein/SABmark", name="twi", split="test")
ex = ds[0]
print(ex["seq1"], ex["seq2"], ex["ref_alignment"][:5])
```
## Citation
Please cite:
```bibtex
@article{VanWalle2004SABmark,
title={SABmark---a benchmark for sequence alignment that covers the entire known fold space},
author={Van Walle, Ivan and Lasters, Ignace and Wyns, Lode},
journal={Bioinformatics},
volume={21},
number={7},
pages={1267--1268},
year={2004},
publisher={Oxford University Press},
DOI = {10.1093/bioinformatics/bth493}
}
```
---
### `dataset_infos.json` — stub (optional but nice)
```json
{
"twi": {
"description": "SABmark Twilight subset (<25% identity).",
"features": {
"pair_id": {"dtype": "string", "_type": "Value"},
"group_id": {"dtype": "string", "_type": "Value"},
"set_name": {"names": ["twi", "sup", "twi_fp", "sup_fp"], "_type": "ClassLabel"},
"seq1_id": {"dtype": "string", "_type": "Value"},
"seq2_id": {"dtype": "string", "_type": "Value"},
"seq1": {"dtype": "string", "_type": "Value"},
"seq2": {"dtype": "string", "_type": "Value"},
"ref_alignment": {"feature": {"feature": {"dtype": "int32", "_type": "Value"}, "_type": "Sequence"}, "_type": "Sequence"},
"percent_identity": {"dtype": "float32", "_type": "Value"},
"scop_labels": {"feature": {"dtype": "string", "_type": "Value"}, "_type": "Sequence"},
"meta": {"dtype": "string", "_type": "Value"}
}
}
}
```
|