Ncbi Datasets ApiSAFE

🔬 A curated collection of 23,000+ agent skills for empirical research across 8 social science disciplines. | 精选 23,000+ AI Agent 技能库,覆盖8大社会科学学科的实证研究。CoPaper.AI 20分钟完成一篇可复现的规范实证论文,并支持用户上传 Skills。-- Maintained by CoPaper.AI from Stanford REAP.

The instruction file, verbatim from the audited commit — this is the text the model reads, and the surface the audit's instruction layer examines. Quoted here so you can judge it without cloning anything.

---
name: ncbi-datasets-api
description: "Access genomes, genes, and taxonomy data via NCBI Datasets v2 API"
metadata:
  openclaw:
    emoji: "🧬"
    category: "domains"
    subcategory: "biomedical"
    keywords: ["NCBI", "genome data", "gene data", "taxonomy", "RefSeq", "GenBank"]
    source: "https://www.ncbi.nlm.nih.gov/datasets/"
---

# NCBI Datasets v2 API

## Overview

NCBI Datasets is the modern API for accessing NCBI's genomic, gene, and taxonomic data — replacing older E-utilities for sequence data retrieval. It provides clean REST endpoints for genome assemblies, gene records, taxonomy trees, and sequence downloads. Covers all organisms in NCBI's databases including RefSeq and GenBank. Free, no authentication required.

## API Endpoints

### Base URL

```
https://api.ncbi.nlm.nih.gov/datasets/v2
```

### Genome Data

```bash
# Search genome assemblies by organism
curl "https://api.ncbi.nlm.nih.gov/datasets/v2/genome/taxon/9606?page_size=5"

# Get assembly by accession
curl "https://api.ncbi.nlm.nih.gov/datasets/v2/genome/accession/GCF_000001405.40"

# Download genome package
curl -o genome.zip \
  "https://api.ncbi.nlm.nih.gov/datasets/v2/genome/accession/GCF_000001405.40/download?\
include_annotation_type=GENOME_FASTA,GENOME_GFF"
```

### Gene Data

```bash
# Search genes by symbol
curl "https://api.ncbi.nlm.nih.gov/datasets/v2/gene/symbol/TP53/taxon/human"

# Get gene by NCBI Gene ID
curl "https://api.ncbi.nlm.nih.gov/datasets/v2/gene/id/7157"

# Search genes by keyword
curl "https://api.ncbi.nlm.nih.gov/datasets/v2/gene/search?query=BRCA&taxon=9606&page_size=20"

# Download gene data package
curl -o gene.zip \
  "https://api.ncbi.nlm.nih.gov/datasets/v2/gene/id/7157/download?include_annotation_type=FASTA_GENE"
```

### Taxonomy

```bash
# Get taxonomy info
curl "https://api.ncbi.nlm.nih.gov/datasets/v2/taxonomy/taxon/9606"

# Search taxonomy by name
curl "https://api.ncbi.nlm.nih.gov/datasets/v2/taxonomy/name_report?taxon_query=Homo+sapiens"

# Get taxonomy tree (subtree)
curl "https://api.ncbi.nlm.nih.gov/datasets/v2/taxonomy/taxon/9443/subtree"
```

### Query Parameters

| Parameter | Description | Example |
|-----------|-------------|---------|
| `page_size` | Results per page | `page_size=20` |
| `page_token` | Pagination token | From previous response |
| `include_annotation_type` | Download content | `GENOME_FASTA`, `GENOME_GFF`, `PROT_FASTA` |
| `filters.assembly_level` | Assembly quality | `complete_genome`, `chromosome` |
| `filters.refseq_only` | RefSeq assemblies | `true` |

## Response Structure (Gene)

```json
{
  "genes": [
    {
      "gene": {
        "gene_id": 7157,
        "symbol": "TP53",
        "description": "tumor protein p53",
        "taxname": "Homo sapiens",
        "tax_id": 9606,
        "type": "PROTEIN_CODING",
        "chromosomes": ["17"],
        "genomic_ranges": [
          {
            "accession_version": "NC_000017.11",
            "range": [{"begin": 7668402, "end": 7687550, "orientation": "minus"}]
          }
        ],
        "nomenclature": {
          "symbol": "TP53",
          "name": "tumor protein p53"
        },
        "annotations": [
          {"release_date": "2024-03-15", "release_name": "GRCh38.p14"}
        ]
      }
    }
  ]
}
```

## Python Usage

```python
import requests
import zipfile
import io

BASE_URL = "https://api.ncbi.nlm.nih.gov/datasets/v2"


def search_genes(query: str, taxon: str = "human",
                 page_size: int = 20) -> list:
    """Search NCBI genes by keyword."""
    resp = requests.get(
        f"{BASE_URL}/gene/search",
        params={"query": query, "taxon": taxon,
                "page_size": page_size},
    )
    resp.raise_for_status()
    data = resp.json()

    results = []
    for item in data.get("genes", []):
        gene = item.get("gene", {})
        results.append({
            "gene_id": gene.get("gene_id"),
            "symbol": gene.get("symbol"),
            "description": gene.get("description"),
            "type": gene.get("type"),
            "chromosomes": gene.get("chromosomes", []),
            "taxname": gene.get("taxname"),
        })
    return results


def get_gene(gene_id: int) -> dict:
    """Get detailed gene information."""
    resp = requests.get(f"{BASE_URL}/gene/id/{gene_id}")
    resp.raise_for_status()
    genes = resp.json().get("genes", [])
    return genes[0].get("gene", {}) if genes else {}


def search_genomes(taxon: str, refseq_only: bool = True,
                   page_size: int = 10) -> list:
    """Search genome assemblies by organism."""
    params = {"page_size": page_size}
    if refseq_only:
        params["filters.refseq_only"] = "true"

    resp = requests.get(
        f"{BASE_URL}/genome/taxon/{taxon}",
        params=params,
    )
    resp.raise_for_status()
    data = resp.json()

    results = []
    for report in data.get("reports", []):
        assembly = report.get("assembly_info", {})
        stats = report.get("assembly_stats", {})
        results.append({
            "accession": report.get("accession"),
            "name": assembly.get("assembly_name"),
            "level": assembly.get("assembly_level"),
            "organism": report.get("organism", {}).get("organism_name"),
            "total_length": stats.get("total_sequence_length"),
            "contig_n50": stats.get("contig_n50"),
        })
    return results


# Example: search cancer-related genes
genes = search_genes("tumor suppressor", taxon="human")
for g in genes[:5]:
    print(f"{g['symbol']} (ID: {g['gene_id']}): {g['description']}")
    print(f"  Type: {g['type']} | Chr: {', '.join(g['chromosomes'])}")

# Example: find reference genomes
genomes = search_genomes("Mus musculus", refseq_only=True)
for g in genomes[:3]:
    print(f"{g['accession']}: {g['name']} ({g['level']})")
    print(f"  Length: {g['total_length']:,} bp")
```

## CLI Tool

NCBI also provides a command-line tool:

```bash
# Install
curl -o datasets "https://ftp.ncbi.nlm.nih.gov/pub/data

SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.

What the source does

Filesystem

none-observed

Network

none-observed

Shell

none-observed

Dependencies

pinned

Secrets in source

none-found

Findings (0)

No findings outside the package's declared scope.

Gates applied: no_behavioural_pass.