BindingDB Query Skill
Search BindingDB binding affinity records by any entity. Auto-detects type by pattern:
| Input Pattern |
Detected As |
Example |
API Endpoint |
P35355, Q9Y233 |
UniProt ID |
P00533 (EGFR) |
getLigandsByUniprots |
1Q0L, 3ANM |
PDB ID (4-char, digit-leading) |
1Q0L |
getLigandsByPDBs |
contains =()#[]@/\ |
SMILES string |
CC(=O)Oc1ccccc1C(O)=O |
getTargetByCompound |
| fallback |
treated as UniProt |
— |
getLigandsByUniprots |
API
| Function |
Input |
Returns |
search(entity, cutoff) |
single entity string |
dict with entity, type, hits, affinities |
search_batch(entities, cutoff) |
list of entity strings |
dict[str, search_result] |
summarize(result) |
search() output |
compact multi-line text |
to_json(result) |
search() output |
list[dict] of affinity records |
query_by_uniprot(ids, cutoff) |
UniProt ID(s), nM cutoff |
list[dict] |
query_by_pdb(ids, cutoff, identity) |
PDB ID(s), nM cutoff, % identity |
list[dict] |
query_by_smiles(smiles, cutoff) |
SMILES, similarity 0–1 |
list[dict] |
Parameters
cutoff (int): affinity threshold in nM (default 10 000). Entries with IC50/Ki/Kd ≤ cutoff are returned.identity (int, PDB only): sequence-identity cutoff in percent (default 92).- Results are capped at 50 per query for LLM readability.
Usage
See if __name__ == "__main__" block in 26_BindingDB.py for runnable examples covering: UniProt single query, PDB query, SMILES compound query, batch query, and JSON output.
Key Fields in Each Affinity Record
| Field |
Description |
query |
Target protein name |
monomerid |
BindingDB compound ID |
smile |
SMILES structure of ligand |
affinity_type |
Ki, IC50, Kd, or EC50 |
affinity |
Value in nM |
pmid |
PubMed ID of source |
doi |
DOI of source publication |
Data Source
- Database: BindingDB (https://www.bindingdb.org/)
- Size: 3.2M data points, 1.4M compounds, 11.4K targets
- Access: Public REST API (JSON), no authentication required
- Citation: BindingDB in 2024: a FAIR knowledgebase of protein-small molecule binding data. Nucleic Acids Research, 53(D1), D1633 (2025). DOI: 10.1093/nar/gkae1199
1---2name: bindingdb-query3description: Query the BindingDB drug-target binding affinity database. Use whenever the user asks about protein-ligand binding data, affinity measurements (Ki, IC50, Kd, EC50), or wants to look up binding partners for a UniProt ID, PDB ID, or compound SMILES string.4---5
6# BindingDB Query Skill
7
8Search BindingDB binding affinity records by any entity. Auto-detects type by pattern:
9
10| Input Pattern | Detected As | Example | API Endpoint |
11|---|---|---|---|
12| `P35355`, `Q9Y233` | UniProt ID | `P00533` (EGFR) | `getLigandsByUniprots` |
13| `1Q0L`, `3ANM` | PDB ID (4-char, digit-leading) | `1Q0L` | `getLigandsByPDBs` |
14| contains `=()#[]@/\` | SMILES string | `CC(=O)Oc1ccccc1C(O)=O` | `getTargetByCompound` |
15| fallback | treated as UniProt | — | `getLigandsByUniprots` |
16
17## API
18
19| Function | Input | Returns |
20|---|---|---|
21| `search(entity, cutoff)` | single entity string | dict with `entity`, `type`, `hits`, `affinities` |
22| `search_batch(entities, cutoff)` | list of entity strings | dict[str, search_result] |
23| `summarize(result)` | search() output | compact multi-line text |
24| `to_json(result)` | search() output | list[dict] of affinity records |
25| `query_by_uniprot(ids, cutoff)` | UniProt ID(s), nM cutoff | list[dict] |
26| `query_by_pdb(ids, cutoff, identity)` | PDB ID(s), nM cutoff, % identity | list[dict] |
27| `query_by_smiles(smiles, cutoff)` | SMILES, similarity 0–1 | list[dict] |
28
29**Parameters**
30
31- `cutoff` (int): affinity threshold in nM (default 10 000). Entries with IC50/Ki/Kd ≤ cutoff are returned.
32- `identity` (int, PDB only): sequence-identity cutoff in percent (default 92).
33- Results are capped at 50 per query for LLM readability.
34
35## Usage
36
37See `if __name__ == "__main__"` block in `26_BindingDB.py` for runnable examples covering: UniProt single query, PDB query, SMILES compound query, batch query, and JSON output.
38
39## Key Fields in Each Affinity Record
40
41| Field | Description |
42|---|---|
43| `query` | Target protein name |
44| `monomerid` | BindingDB compound ID |
45| `smile` | SMILES structure of ligand |
46| `affinity_type` | Ki, IC50, Kd, or EC50 |
47| `affinity` | Value in nM |
48| `pmid` | PubMed ID of source |
49| `doi` | DOI of source publication |
50
51## Data Source
52
53- **Database**: BindingDB (https://www.bindingdb.org/)
54- **Size**: 3.2M data points, 1.4M compounds, 11.4K targets
55- **Access**: Public REST API (JSON), no authentication required
56- **Citation**: BindingDB in 2024: a FAIR knowledgebase of protein-small molecule binding data. *Nucleic Acids Research*, 53(D1), D1633 (2025). DOI: 10.1093/nar/gkae1199
gabrielmoreira