AlphaFold Database

Overview

AlphaFold DB is a public repository of AI-predicted 3D protein structures for over 200 million proteins, maintained by DeepMind and EMBL-EBI. Access structure predictions with confidence metrics, download coordinate files, retrieve bulk datasets, and integrate predictions into computational workflows.

When to Use This Skill

This skill should be used when working with AI-predicted protein structures in scenarios such as:

• Retrieving protein structure predictions by UniProt ID or protein name
• Downloading PDB/mmCIF coordinate files for structural analysis
• Analyzing prediction confidence metrics (pLDDT, PAE) to assess reliability
• Accessing bulk proteome datasets via Google Cloud Platform
• Comparing predicted structures with experimental data
• Performing structure-based drug discovery or protein engineering
• Building structural models for proteins lacking experimental structures
• Integrating AlphaFold predictions into computational pipelines

Core Capabilities

1. Searching and Retrieving Predictions

Using Biopython (Recommended):

The Biopython library provides the simplest interface for retrieving AlphaFold structures:

from Bio.PDB import alphafold_db

# Get all predictions for a UniProt accession
predictions = list(alphafold_db.get_predictions("P00520"))

# Download structure file (mmCIF format)
for prediction in predictions:
    cif_file = alphafold_db.download_cif_for(prediction, directory="./structures")
    print(f"Downloaded: {cif_file}")

# Get Structure objects directly
from Bio.PDB import MMCIFParser
structures = list(alphafold_db.get_structural_models_for("P00520"))

Direct API Access:

Query predictions using REST endpoints:

import requests

# Get prediction metadata for a UniProt accession
uniprot_id = "P00520"
api_url = f"https://alphafold.ebi.ac.uk/api/prediction/{uniprot_id}"
response = requests.get(api_url)
prediction_data = response.json()

# Extract AlphaFold ID
alphafold_id = prediction_data[0]['entryId']
print(f"AlphaFold ID: {alphafold_id}")

Using UniProt to Find Accessions:

Search UniProt to find protein accessions first:

import urllib.parse, urllib.request

def get_uniprot_ids(query, query_type='PDB_ID'):
    """Query UniProt to get accession IDs"""
    url = 'https://www.uniprot.org/uploadlists/'
    params = {
        'from': query_type,
        'to': 'ACC',
        'format': 'txt',
        'query': query
    }
    data = urllib.parse.urlencode(params).encode('ascii')
    with urllib.request.urlopen(urllib.request.Request(url, data)) as response:
        return response.read().decode('utf-8').splitlines()

# Example: Find UniProt IDs for a protein name
protein_ids = get_uniprot_ids("hemoglobin", query_type="GENE_NAME")

2. Downloading Structure Files

AlphaFold provides multiple file formats for each prediction:

File Types Available:

• Model coordinates (model_v4.cif): Atomic coordinates in mmCIF/PDBx format
• Confidence scores (confidence_v4.json): Per-residue pLDDT scores (0-100)
• Predicted Aligned Error (predicted_aligned_error_v4.json): PAE matrix for residue pair confidence

Download URLs:

import requests

alphafold_id = "AF-P00520-F1"
version = "v4"

# Model coordinates (mmCIF)
model_url = f"https://alphafold.ebi.ac.uk/files/{alphafold_id}-model_{version}.cif"
response = requests.get(model_url)
with open(f"{alphafold_id}.cif", "w") as f:
    f.write(response.text)

# Confidence scores (JSON)
confidence_url = f"https://alphafold.ebi.ac.uk/files/{alphafold_id}-confidence_{version}.json"
response = requests.get(confidence_url)
confidence_data = response.json()

# Predicted Aligned Error (JSON)
pae_url = f"https://alphafold.ebi.ac.uk/files/{alphafold_id}-predicted_aligned_error_{version}.json"
response = requests.get(pae_url)
pae_data = response.json()

PDB Format (Alternative):

# Download as PDB format instead of mmCIF
pdb_url = f"https://alphafold.ebi.ac.uk/files/{alphafold_id}-model_{version}.pdb"
response = requests.get(pdb_url)
with open(f"{alphafold_id}.pdb", "wb") as f:
    f.write(response.content)

3. Working with Confidence Metrics

AlphaFold predictions include confidence estimates critical for interpretation:

pLDDT (per-residue confidence):

import json
import requests

# Load confidence scores
alphafold_id = "AF-P00520-F1"
confidence_url = f"https://alphafold.ebi.ac.uk/files/{alphafold_id}-confidence_v4.json"
confidence = requests.get(confidence_url).json()

# Extract pLDDT scores
plddt_scores = confidence['confidenceScore']

# Interpret confidence levels
# pLDDT > 90: Very high confidence
# pLDDT 70-90: High confidence
# pLDDT 50-70: Low confidence
# pLDDT < 50: Very low confidence

high_confidence_residues = [i for i, score in enumerate(plddt_scores) if score > 90]
print(f"High confidence residues: {len(high_confidence_residues)}/{len(plddt_scores)}")

PAE (Predicted Aligned Error):

PAE indicates confidence in relative domain positions:

import numpy as np
import matplotlib.pyplot as plt

# Load PAE matrix
pae_url = f"https://alphafold.ebi.ac.uk/files/{alphafold_id}-predicted_aligned_error_v4.json"
pae = requests.get(pae_url).json()

# Visualize PAE matrix
pae_matrix = np.array(pae['distance'])
plt.figure(figsize=(10, 8))
plt.imshow(pae_matrix, cmap='viridis_r', vmin=0, vmax=30)
plt.colorbar(label='PAE (Å)')
plt.title(f'Predicted Aligned Error: {alphafold_id}')
plt.xlabel('Residue')
plt.ylabel('Residue')
plt.savefig(f'{alphafold_id}_pae.png', dpi=300, bbox_inches='tight')

# Low PAE values (<5 Å) indicate confident relative positioning
# High PAE values (>15 Å) suggest uncertain domain arrangements

4. Bulk Data Access via Google Cloud

For large-scale analyses, use Google Cloud datasets:

Google Cloud Storage:

# Install gsutil
uv pip install gsutil

# List available data
gsutil ls gs://public-datasets-deepmind-alphafold-v4/

# Download entire proteomes (by taxonomy ID)
gsutil -m cp gs://public-datasets-deepmind-alphafold-v4/proteomes/proteome-tax_id-9606-*.tar .

# Download specific files
gsutil cp gs://public-datasets-deepmind-alphafold-v4/accession_ids.csv .

BigQuery Metadata Access:

from google.cloud import bigquery

# Initialize client
client = bigquery.Client()

# Query metadata
query = """
SELECT
  entryId,
  uniprotAccession,
  organismScientificName,
  globalMetricValue,
  fractionPlddtVeryHigh
FROM `bigquery-public-data.deepmind_alphafold.metadata`
WHERE organismScientificName = 'Homo sapiens'
  AND fractionPlddtVeryHigh > 0.8
LIMIT 100
"""

results = client.query(query).to_dataframe()
print(f"Found {len(results)} high-confidence human proteins")

Download by Species:

import subprocess

def download_proteome(taxonomy_id, output_dir="./proteomes"):
    """Download all AlphaFold predictions for a species"""
    pattern = f"gs://public-datasets-deepmind-alphafold-v4/proteomes/proteome-tax_id-{taxonomy_id}-*_v4.tar"
    cmd = f"gsutil -m cp {pattern} {output_dir}/"
    subprocess.run(cmd, shell=True, check=True)

# Download E. coli proteome (tax ID: 83333)
download_proteome(83333)

# Download human proteome (tax ID: 9606)
download_proteome(9606)

5. Parsing and Analyzing Structures

Work with downloaded AlphaFold structures using BioPython:

from Bio.PDB import MMCIFParser, PDBIO
import numpy as np

# Parse mmCIF file
parser = MMCIFParser(QUIET=True)
structure = parser.get_structure("protein", "AF-P00520-F1-model_v4.cif")

# Extract coordinates
coords = []
for model in structure:
    for chain in model:
        for residue in chain:
            if 'CA' in residue:  # Alpha carbons only
                coords.append(residue['CA'].get_coord())

coords = np.array(coords)
print(f"Structure has {len(coords)} residues")

# Calculate distances
from scipy.spatial.distance import pdist, squareform
distance_matrix = s