Improve dataset card: add paper link, code link, and task category

#2
by nielsr HF Staff - opened
Files changed (1) hide show
  1. README.md +56 -31
README.md CHANGED
@@ -1,40 +1,64 @@
1
  ---
 
 
2
  license: cc-by-4.0
3
  size_categories:
4
  - 1K<n<10K
5
- language:
6
- - en
7
  tags:
8
  - biology
9
-
10
  configs:
11
  - config_name: benchmark
12
  data_files:
13
  - split: test
14
  path:
15
- - "astex_benchmark.csv"
16
- - "posex_self_dock_benchmark.csv"
17
- - "posex_cross_dock_benchmark.csv"
18
  ---
19
- The file `posex_set.zip` contains processed dataset for docking evaluation, while `posex_cif.zip` contains the raw CIF files from RCSB PDB. For information about creating a dataset from scratch, please refer to the data construction [code](https://github.com/CataAI/PoseX/tree/master/dataset).
 
 
 
 
 
 
 
 
 
20
  The processed dataset comprises three sub-datasets in total:
21
- - The Astex Diverse set was curated by Hartshorn et al. [1] and contains 85 cases, it was reprocessed to adapt to our evaluation process.
22
- - The PoseX Self-Docking set was curated by us using the PDB and contains 718 cases.
23
- - The PoseX Cross-Docking set was curated by us using the PDB and contains 1312 pairs.
24
 
25
- For each protein-ligand complex, there is one folder named after the PDB identifier and the CCD identifier (PDB_CCD).
 
 
 
 
 
 
 
 
 
 
 
 
 
 
26
 
27
  Each folder contains the following files:
28
- 1. PDB file PDB_CCD_protein.pdb: The protein structure without the ligand, without solvents, and with all cofactors.
29
- 2. SDF file PDB_CCD_ligands.sdf: All instances of the ligand.
30
- 3. SDF file PDB_CCD_ligand.sdf: One instance of the ligand.
31
- 4. SDF file PDB_CCD_ligand_start_conf.sdf: One generated molecule conformation for the ligand.
32
- 5. JSON file PDB_CCD.json: Input JSON file for AlphaFold 3
33
- 6. PDB file PDB_CCD_ref_protein.pdb (unique to Cross-Docking set): The reference protein structure.
34
- 7. TXT file group_id.txt (unique to Cross-Docking set): Group information
35
- 8. In folder posex_self_docking_set and posex_cross_docking_set, a csv file named qtm.csv contains pocket structure similarity which is calculated for each evaluation data point relative to the training data.
36
-
37
- The folder structure is:
38
  ```plaintext
39
  posex_set/
40
  ├── astex_diverse_set/
@@ -67,19 +91,20 @@ The folder structure is:
67
  └── └── ...
68
  ```
69
 
70
- This dataset is associated with the following paper:
 
 
 
71
  ```bibtex
72
- @misc{jiang2025posexaidefeatsphysics,
73
- title={PoseX: AI Defeats Physics Approaches on Protein-Ligand Cross Docking},
74
- author={Yize Jiang and Xinze Li and Yuanyuan Zhang and Jin Han and Youjun Xu and Ayush Pandit and Zaixi Zhang and Mengdi Wang and Mengyang Wang and Chong Liu and Guang Yang and Yejin Choi and Wu-Jun Li and Tianfan Fu and Fang Wu and Junhong Liu},
75
- year={2025},
76
- eprint={2505.01700},
77
- archivePrefix={arXiv},
78
- primaryClass={cs.LG},
79
- url={https://arxiv.org/abs/2505.01700},
80
  }
81
  ```
82
 
83
- References:
84
 
85
  [1] Hartshorn MJ, Verdonk ML, Chessari G, Brewerton SC, Mooij WTM, Mortenson PN, et al. Diverse, high-quality test set for the validation of protein-ligand docking performance. J Med Chem. 2007 Feb 1;50(4):726–41.
 
1
  ---
2
+ language:
3
+ - en
4
  license: cc-by-4.0
5
  size_categories:
6
  - 1K<n<10K
7
+ task_categories:
8
+ - other
9
  tags:
10
  - biology
11
+ - protein-ligand-docking
12
  configs:
13
  - config_name: benchmark
14
  data_files:
15
  - split: test
16
  path:
17
+ - astex_benchmark.csv
18
+ - posex_self_dock_benchmark.csv
19
+ - posex_cross_dock_benchmark.csv
20
  ---
21
+
22
+ # PoseX: AI Defeats Physics Approaches on Protein-Ligand Cross Docking
23
+
24
+ [**Paper**](https://huggingface.co/papers/2505.01700) | [**GitHub**](https://github.com/CataAI/PoseX) | [**Leaderboard**](http://dock-lab.tech)
25
+
26
+ PoseX is a comprehensive benchmark dataset designed to evaluate molecular docking algorithms for predicting protein-ligand binding poses. It includes curated datasets for both self-docking and cross-docking scenarios.
27
+
28
+ The file `posex_set.zip` contains the processed dataset for docking evaluation, while `posex_cif.zip` contains the raw CIF files from RCSB PDB. For information about creating a dataset from scratch, please refer to the data construction [code](https://github.com/CataAI/PoseX/tree/master/dataset).
29
+
30
+ ## Dataset Subsets
31
  The processed dataset comprises three sub-datasets in total:
32
+ - **The Astex Diverse set**: Curated by Hartshorn et al. [1] and contains 85 cases, reprocessed to adapt to our evaluation process.
33
+ - **The PoseX Self-Docking set**: Curated using the PDB, containing 718 cases.
34
+ - **The PoseX Cross-Docking set**: Curated using the PDB, containing 1312 pairs.
35
 
36
+ ## Sample Usage
37
+
38
+ To generate the benchmark CSV files (containing protein sequences, ligand SMILES, and metadata) from the scripts provided in the [official repository](https://github.com/CataAI/PoseX):
39
+
40
+ ```bash
41
+ # Example: Generate benchmark CSV for PoseX Self-Docking
42
+ bash ./scripts/generate_docking_benchmark.sh posex_self_dock
43
+
44
+ # Example: Generate benchmark CSV for PoseX Cross-Docking
45
+ bash ./scripts/generate_docking_benchmark.sh posex_cross_dock
46
+ ```
47
+
48
+ ## File Structure
49
+
50
+ For each protein-ligand complex, there is one folder named after the PDB identifier and the CCD identifier (`PDB_CCD`).
51
 
52
  Each folder contains the following files:
53
+ 1. `PDB_CCD_protein.pdb`: The protein structure without the ligand, without solvents, and with all cofactors.
54
+ 2. `PDB_CCD_ligands.sdf`: All instances of the ligand.
55
+ 3. `PDB_CCD_ligand.sdf`: One instance of the ligand.
56
+ 4. `PDB_CCD_ligand_start_conf.sdf`: One generated molecule conformation for the ligand.
57
+ 5. `PDB_CCD.json`: Input JSON file for AlphaFold 3.
58
+ 6. `PDB_CCD_ref_protein.pdb` (unique to Cross-Docking set): The reference protein structure.
59
+ 7. `group_id.txt` (unique to Cross-Docking set): Group information.
60
+ 8. `qtm.csv`: Contains pocket structure similarity calculated for each evaluation data point relative to the training data.
61
+
 
62
  ```plaintext
63
  posex_set/
64
  ├── astex_diverse_set/
 
91
  └── └── ...
92
  ```
93
 
94
+ ## Citation
95
+
96
+ If you use this dataset or code, please cite:
97
+
98
  ```bibtex
99
+ @inproceedings{jiang2026posex,
100
+ title={PoseX: {AI} Defeats Physics-based Methods on Protein Ligand Cross-Docking},
101
+ author={Yize Jiang and Xinze Li and Yuanyuan Zhang and Jin Han and Youjun Xu and Ayush Pandit and ZAIXI ZHANG and Mengdi Wang and Mengyang Wang and Chong Liu and Guang Yang and Yejin Choi and Yingzhou Lu and Wu-Jun Li and Tianfan Fu and Fang Wu and Junhong Liu},
102
+ booktitle={The Fourteenth International Conference on Learning Representations},
103
+ year={2026},
104
+ url={https://openreview.net/forum?id=qqzxKudD4T}
 
 
105
  }
106
  ```
107
 
108
+ ## References
109
 
110
  [1] Hartshorn MJ, Verdonk ML, Chessari G, Brewerton SC, Mooij WTM, Mortenson PN, et al. Diverse, high-quality test set for the validation of protein-ligand docking performance. J Med Chem. 2007 Feb 1;50(4):726–41.