rcsb_database

Name: rcsb_database
Author: ai4protein

244stars35forks

How to install rcsb_database?

ai4protein/venusfactory2/rcsb_database

$npx -y skills add ai4protein/VenusFactory2 --skill rcsb_database

Installs into the current project.

Summary

RCSB Protein Data Bank (PDB) — experimentally determined 3D biomolecular structures. Search by full-text/sequence/structure/attribute, fetch entry metadata, download coordinate files (PDB/mmCIF). Use when the user provides a PDB ID, asks for structures of a protein, wants to find

Files · 1

View on GitHub

SKILL.md

1	# RCSB Protein Data Bank
2
3	## Overview
4
5	The RCSB PDB hosts >200K experimentally-determined biomolecular structures (X-ray, cryo-EM, NMR, etc.). This skill exposes 3 download tools and 1 search tool.
6
7	## Project Tools (VenusFactory2)
8
9	\| Tool \| Args \| Returns \| Description \|
10	\|------\|------\|---------\|--------------\|
11	\| download_rcsb_search_by_query \| `query` (JSON string or dict — Search API v2 query block or full payload), `out_dir` (required), `return_type` (default `"entry"`; `entry` \\| `assembly` \\| `polymer_entity` \\| `non_polymer_entity` \\| `polymer_instance` \\| `mol_definition`), `page_start`/`rows` (pagination; default = return all hits), `sort_by` (e.g. `score`, `rcsb_accession_info.initial_release_date`), `sort_direction` (`asc` \\| `desc`), `count_only` (skip results, return only `total_count`), `timeout` (default `60`s) \| JSON: `{status, file_info {file_path → rcsb_search_<return_type>.json}, content_preview (first 25 ids), biological_metadata {return_type, total_count, result_count, page_start, rows, sort_by, search_time_ms}}` \| Run an RCSB Search API v2 query (text / sequence / structure / attribute). \|
12	\| download_rcsb_entry_metadata_by_pdb_id \| `pdb_id` (required, e.g. `4HHB`), `out_path` (required, JSON path) \| rich JSON envelope; saves metadata JSON to `out_path` \| Fetch full entry metadata for one PDB ID. \|
13	\| download_rcsb_structure_by_pdb_id \| `pdb_id` (required), `out_dir` (required), `file_type` (default `pdb`; `pdb` \\| `cif` \\| `xml`) \| rich JSON envelope; structure file at `file_info.file_path` \| Download coordinate file. \|
14
15	## When to Use This Skill
16
17	- User provides a PDB ID and asks for the structure / metadata
18	- "Find structures of X" / "find structures similar to this sequence" → `download_rcsb_search_by_query` with the right `service`
19	- Build a workflow that pipes RCSB structures into PyMOL (`render_protein_structure`) or Foldseek
20
21	## Search API v2 Query Cookbook
22
23	Pass `query` as a dict / JSON-string of a query block (auto-wrapped into `{"query": ...}`) or a full request payload.
24
25	Full-text:
26	```python
27	{"type": "terminal", "service": "full_text", "parameters": {"value": "hemoglobin"}}
28	```
29
30	By sequence (BLAST-like, fast, e-value cutoff):
31	```python
32	{
33	"type": "terminal", "service": "sequence",
34	"parameters": {
35	"evalue_cutoff": 0.1, "identity_cutoff": 0.95,
36	"sequence_type": "protein",
37	"value": "MVHLTPEEKSAVTALWGKVNVDEVGGEALGRLLVVYPWTQRFF..."
38	}
39	}
40	```
41
42	By attribute (e.g. resolution, organism, year):
43	```python
44	{
45	"type": "group", "logical_operator": "and",
46	"nodes": [
47	{"type": "terminal", "service": "text",
48	"parameters": {"attribute": "rcsb_entry_info.resolution_combined",
49	"operator": "less", "value": 2.0}},
50	{"type": "terminal", "service": "text",
51	"parameters": {"attribute": "rcsb_entity_source_organism.taxonomy_lineage.name",
52	"operator": "exact_match", "value": "Homo sapiens"}}
53	]
54	}
55	```
56
57	Boolean combination: use `{"type": "group", "logical_operator": "and"\|"or", "nodes": [...]}`.
58
59	## Return Type → Identifier Shape
60
61	\| return_type \| identifier shape \|
62	\|---\|---\|
63	\| `entry` \| `4HHB` (PDB ID) \|
64	\| `assembly` \| `4HHB-1` \|
65	\| `polymer_entity` \| `4HHB-1` (entity ID) \|
66	\| `polymer_instance` \| `4HHB.A` (chain) \|
67	\| `mol_definition` \| `HEM` (ligand 3-letter code) \|
68
69	## Common Mistakes
70
71	- Passing a Python query dict expecting URL-encoding: dicts are serialized & url-encoded automatically; pass the dict, not a hand-crafted URL.
72	- Forgetting `return_type`: defaults to `entry` (PDB IDs). If you want chain identifiers, pass `polymer_instance`.
73	- Using `count_only=True` and then trying to read `result_set`: count-only mode returns `{"total_count": N}` only, no `result_set`.
74	- Sequence searches with `identity_cutoff` too strict: try lower (e.g. 0.3) if you get few hits.
75
76	## References
77
78	- [RCSB Search API v2 docs](https://search.rcsb.org/)
79	- [Query attribute schema](https://search.rcsb.org/rcsbsearch/v2/metadata/schema)
80	- [Data API docs](https://data.rcsb.org/)

People who liked this also viewed

6 picks

Type

rcsb_database

244stars35forks

View on GitHub

How to install rcsb_database?

ai4protein/venusfactory2/rcsb_database

$npx -y skills add ai4protein/VenusFactory2 --skill rcsb_database

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# RCSB Protein Data Bank
2
3	## Overview
4
5	The RCSB PDB hosts >200K experimentally-determined biomolecular structures (X-ray, cryo-EM, NMR, etc.). This skill exposes 3 download tools and 1 search tool.
6
7	## Project Tools (VenusFactory2)
8
9	\| Tool \| Args \| Returns \| Description \|
10	\|------\|------\|---------\|--------------\|
11	\| download_rcsb_search_by_query \| `query` (JSON string or dict — Search API v2 query block or full payload), `out_dir` (required), `return_type` (default `"entry"`; `entry` \\| `assembly` \\| `polymer_entity` \\| `non_polymer_entity` \\| `polymer_instance` \\| `mol_definition`), `page_start`/`rows` (pagination; default = return all hits), `sort_by` (e.g. `score`, `rcsb_accession_info.initial_release_date`), `sort_direction` (`asc` \\| `desc`), `count_only` (skip results, return only `total_count`), `timeout` (default `60`s) \| JSON: `{status, file_info {file_path → rcsb_search_<return_type>.json}, content_preview (first 25 ids), biological_metadata {return_type, total_count, result_count, page_start, rows, sort_by, search_time_ms}}` \| Run an RCSB Search API v2 query (text / sequence / structure / attribute). \|
12	\| download_rcsb_entry_metadata_by_pdb_id \| `pdb_id` (required, e.g. `4HHB`), `out_path` (required, JSON path) \| rich JSON envelope; saves metadata JSON to `out_path` \| Fetch full entry metadata for one PDB ID. \|
13	\| download_rcsb_structure_by_pdb_id \| `pdb_id` (required), `out_dir` (required), `file_type` (default `pdb`; `pdb` \\| `cif` \\| `xml`) \| rich JSON envelope; structure file at `file_info.file_path` \| Download coordinate file. \|
14
15	## When to Use This Skill
16
17	- User provides a PDB ID and asks for the structure / metadata
18	- "Find structures of X" / "find structures similar to this sequence" → `download_rcsb_search_by_query` with the right `service`
19	- Build a workflow that pipes RCSB structures into PyMOL (`render_protein_structure`) or Foldseek
20
21	## Search API v2 Query Cookbook
22
23	Pass `query` as a dict / JSON-string of a query block (auto-wrapped into `{"query": ...}`) or a full request payload.
24
25	Full-text:
26	```python
27	{"type": "terminal", "service": "full_text", "parameters": {"value": "hemoglobin"}}
28	```
29
30	By sequence (BLAST-like, fast, e-value cutoff):
31	```python
32	{
33	"type": "terminal", "service": "sequence",
34	"parameters": {
35	"evalue_cutoff": 0.1, "identity_cutoff": 0.95,
36	"sequence_type": "protein",
37	"value": "MVHLTPEEKSAVTALWGKVNVDEVGGEALGRLLVVYPWTQRFF..."
38	}
39	}
40	```
41
42	By attribute (e.g. resolution, organism, year):
43	```python
44	{
45	"type": "group", "logical_operator": "and",
46	"nodes": [
47	{"type": "terminal", "service": "text",
48	"parameters": {"attribute": "rcsb_entry_info.resolution_combined",
49	"operator": "less", "value": 2.0}},
50	{"type": "terminal", "service": "text",
51	"parameters": {"attribute": "rcsb_entity_source_organism.taxonomy_lineage.name",
52	"operator": "exact_match", "value": "Homo sapiens"}}
53	]
54	}
55	```
56
57	Boolean combination: use `{"type": "group", "logical_operator": "and"\|"or", "nodes": [...]}`.
58
59	## Return Type → Identifier Shape
60
61	\| return_type \| identifier shape \|
62	\|---\|---\|
63	\| `entry` \| `4HHB` (PDB ID) \|
64	\| `assembly` \| `4HHB-1` \|
65	\| `polymer_entity` \| `4HHB-1` (entity ID) \|
66	\| `polymer_instance` \| `4HHB.A` (chain) \|
67	\| `mol_definition` \| `HEM` (ligand 3-letter code) \|
68
69	## Common Mistakes
70
71	- Passing a Python query dict expecting URL-encoding: dicts are serialized & url-encoded automatically; pass the dict, not a hand-crafted URL.
72	- Forgetting `return_type`: defaults to `entry` (PDB IDs). If you want chain identifiers, pass `polymer_instance`.
73	- Using `count_only=True` and then trying to read `result_set`: count-only mode returns `{"total_count": N}` only, no `result_set`.
74	- Sequence searches with `identity_cutoff` too strict: try lower (e.g. 0.3) if you get few hits.
75
76	## References
77
78	- [RCSB Search API v2 docs](https://search.rcsb.org/)
79	- [Query attribute schema](https://search.rcsb.org/rcsbsearch/v2/metadata/schema)
80	- [Data API docs](https://data.rcsb.org/)

People who liked this also viewed

6 picks

Type