AgentSkillsCN

bio-entrez-fetch

使用 Biopython Bio.Entrez 从 NCBI 数据库中获取记录。在下载序列、获取 GenBank 记录、获取文献摘要,或将 NCBI 数据解析为 Biopython 对象时,可选用此功能。

SKILL.md
--- frontmatter
name: bio-entrez-fetch
description: Retrieve records from NCBI databases using Biopython Bio.Entrez. Use when downloading sequences, fetching GenBank records, getting document summaries, or parsing NCBI data into Biopython objects.
tool_type: python
primary_tool: Bio.Entrez

Entrez Fetch

Retrieve records from NCBI databases using Biopython's Entrez module (EFetch, ESummary utilities).

Required Setup

python
from Bio import Entrez

Entrez.email = 'your.email@example.com'  # Required by NCBI
Entrez.api_key = 'your_api_key'          # Optional, raises rate limit 3->10 req/sec

Core Functions

Entrez.efetch() - Retrieve Full Records

Fetch complete records in various formats from any NCBI database.

python
# Fetch GenBank record by ID
handle = Entrez.efetch(db='nucleotide', id='NM_007294', rettype='gb', retmode='text')
genbank_text = handle.read()
handle.close()

# Fetch FASTA sequence
handle = Entrez.efetch(db='nucleotide', id='NM_007294', rettype='fasta', retmode='text')
fasta_text = handle.read()
handle.close()

# Fetch multiple records
handle = Entrez.efetch(db='nucleotide', id='NM_007294,NM_000059', rettype='fasta', retmode='text')

Key Parameters:

ParameterDescriptionExample
dbDatabase name'nucleotide', 'protein', 'pubmed'
idRecord ID(s)'NM_007294' or '123,456,789'
rettypeReturn type'fasta', 'gb', 'abstract'
retmodeReturn mode'text', 'xml'
retstartStart index0
retmaxMax records20
WebEnvHistory server sessionFrom esearch
query_keyHistory server queryFrom esearch

Common Return Types by Database

Nucleotide/Protein:

rettyperetmodeDescription
'fasta''text'FASTA sequence
'gb''text'GenBank flat file
'gp''text'GenPept flat file (protein)
'gbwithparts''text'GenBank with contig sequences
'seqid''text'Seq-id only
'acc''text'Accession only

PubMed:

rettyperetmodeDescription
'abstract''text'Abstract text
'medline''text'MEDLINE format
'xml''xml'Full PubMed XML

Gene:

rettyperetmodeDescription
'gene_table''text'Gene table format
'xml''xml'Full gene XML

Entrez.esummary() - Document Summaries

Get brief summaries without downloading full records. Faster than efetch.

python
# Get summary for nucleotide record
handle = Entrez.esummary(db='nucleotide', id='NM_007294')
record = Entrez.read(handle)
handle.close()

summary = record[0]  # First (only) record
print(f"Title: {summary['Title']}")
print(f"Length: {summary['Length']}")
print(f"Organism: {summary['Organism']}")

Common Summary Fields:

python
# Nucleotide/Protein
summary['Title']          # Record title/description
summary['Caption']        # Short identifier
summary['Length']         # Sequence length
summary['Organism']       # Source organism
summary['TaxId']          # Taxonomy ID
summary['AccessionVersion']  # Full accession.version

# PubMed
summary['Title']          # Article title
summary['AuthorList']     # Authors
summary['Source']         # Journal
summary['PubDate']        # Publication date
summary['DOI']            # Digital Object Identifier

Parsing with Biopython

Parse into SeqRecord Objects

python
from Bio import Entrez, SeqIO

Entrez.email = 'your.email@example.com'

# Parse GenBank into SeqRecord
handle = Entrez.efetch(db='nucleotide', id='NM_007294', rettype='gb', retmode='text')
record = SeqIO.read(handle, 'genbank')
handle.close()

print(f"ID: {record.id}")
print(f"Length: {len(record.seq)}")
print(f"Features: {len(record.features)}")

# Parse FASTA into SeqRecord
handle = Entrez.efetch(db='nucleotide', id='NM_007294', rettype='fasta', retmode='text')
record = SeqIO.read(handle, 'fasta')
handle.close()

Parse Multiple Records

python
# Fetch multiple as FASTA
handle = Entrez.efetch(db='nucleotide', id='NM_007294,NM_000059,NM_000546', rettype='fasta', retmode='text')
records = list(SeqIO.parse(handle, 'fasta'))
handle.close()

for record in records:
    print(f"{record.id}: {len(record.seq)} bp")

Parse XML with Entrez.read()

python
# For structured data, use XML mode
handle = Entrez.efetch(db='gene', id='672', retmode='xml')
records = Entrez.read(handle)
handle.close()

# Navigate nested structure
gene = records[0]
print(f"Gene: {gene['Entrezgene_gene']['Gene-ref']['Gene-ref_locus']}")

Code Patterns

Fetch Sequence by Accession

python
from Bio import Entrez, SeqIO

Entrez.email = 'your.email@example.com'

def fetch_sequence(accession, db='nucleotide'):
    handle = Entrez.efetch(db=db, id=accession, rettype='fasta', retmode='text')
    record = SeqIO.read(handle, 'fasta')
    handle.close()
    return record

seq = fetch_sequence('NM_007294')
print(f"{seq.id}: {seq.seq[:50]}...")

Fetch GenBank with Features

python
def fetch_genbank(accession):
    handle = Entrez.efetch(db='nucleotide', id=accession, rettype='gb', retmode='text')
    record = SeqIO.read(handle, 'genbank')
    handle.close()
    return record

gb = fetch_genbank('NM_007294')
for feature in gb.features:
    if feature.type == 'CDS':
        print(f"CDS: {feature.location}")
        print(f"Product: {feature.qualifiers.get('product', ['?'])[0]}")

Fetch PubMed Abstract

python
def fetch_abstract(pmid):
    handle = Entrez.efetch(db='pubmed', id=pmid, rettype='abstract', retmode='text')
    abstract = handle.read()
    handle.close()
    return abstract

abstract = fetch_abstract('35412348')
print(abstract)

Get Record Summaries

python
def get_summaries(db, ids):
    if isinstance(ids, list):
        ids = ','.join(ids)
    handle = Entrez.esummary(db=db, id=ids)
    records = Entrez.read(handle)
    handle.close()
    return records

summaries = get_summaries('nucleotide', ['NM_007294', 'NM_000059'])
for s in summaries:
    print(f"{s['Caption']}: {s['Title'][:50]}... ({s['Length']} bp)")

Search Then Fetch

python
# Search for records
handle = Entrez.esearch(db='nucleotide', term='human[orgn] AND insulin[gene] AND mRNA[fkey]', retmax=5)
search_results = Entrez.read(handle)
handle.close()

ids = search_results['IdList']

# Fetch the sequences
handle = Entrez.efetch(db='nucleotide', id=','.join(ids), rettype='fasta', retmode='text')
records = list(SeqIO.parse(handle, 'fasta'))
handle.close()

for record in records:
    print(f"{record.id}: {len(record.seq)} bp")

Fetch Protein by Gene ID

python
# Search gene database
handle = Entrez.esearch(db='gene', term='BRCA1[sym] AND human[orgn]')
result = Entrez.read(handle)
handle.close()
gene_id = result['IdList'][0]

# Get linked protein IDs
handle = Entrez.elink(dbfrom='gene', db='protein', id=gene_id)
links = Entrez.read(handle)
handle.close()

protein_ids = [link['Id'] for link in links[0]['LinkSetDb'][0]['Link'][:3]]

# Fetch proteins
handle = Entrez.efetch(db='protein', id=','.join(protein_ids), rettype='fasta', retmode='text')
proteins = list(SeqIO.parse(handle, 'fasta'))
handle.close()

Save Fetched Records to File

python
def download_sequences(ids, output_file, db='nucleotide', format='fasta'):
    handle = Entrez.efetch(db=db, id=','.join(ids), rettype=format, retmode='text')
    with open(output_file, 'w') as out:
        out.write(handle.read())
    handle.close()

download_sequences(['NM_007294', 'NM_000059'], 'brca_genes.fasta')

Common Errors

ErrorCauseSolution
HTTPError 400Invalid ID or parametersVerify ID exists, check rettype
HTTPError 429Rate limit exceededAdd delays or use API key
Empty resultRecord doesn't existVerify accession in web browser
ValueError in SeqIOWrong format specifiedMatch rettype with SeqIO format
ExpatErrorXML parsing errorUse retmode='text' instead

Decision Tree

code
Need to retrieve NCBI records?
├── Need full sequence?
│   └── Use efetch with rettype='fasta'
├── Need sequence + annotations?
│   └── Use efetch with rettype='gb' (GenBank)
├── Just need metadata (length, organism)?
│   └── Use esummary (faster)
├── Need PubMed abstract?
│   └── Use efetch with rettype='abstract'
├── Need structured data for parsing?
│   └── Use efetch with retmode='xml' + Entrez.read()
├── Downloading many records?
│   └── See batch-downloads skill
└── Need records from multiple databases?
    └── See entrez-link skill first

Related Skills

  • entrez-search - Find record IDs before fetching
  • entrez-link - Find related records in other databases
  • batch-downloads - Download large numbers of records efficiently
  • sequence-io/read-sequences - Parse downloaded sequences with SeqIO