ENCODE Database

Overview

The ENCODE (Encyclopedia of DNA Elements) Project has generated thousands of functional genomics experiments — TF ChIP-seq, ATAC-seq, DNase-seq, histone ChIP-seq, and RNA-seq — across 1000+ human and mouse cell types and tissues. The ENCODE Portal REST API provides structured JSON access to experiment metadata, file download URLs, and SCREEN cCRE (candidate cis-Regulatory Elements) annotations. All data is freely accessible without authentication for most endpoints.

When to Use

• Downloading TF ChIP-seq peak files (BED) for a specific transcription factor and cell type to annotate regulatory regions
• Finding ATAC-seq or DNase-seq peaks in a cell type to identify open chromatin regions near a gene of interest
• Retrieving cCREs (candidate cis-Regulatory Elements) overlapping a genomic region from ENCODE SCREEN
• Building reference regulatory tracks for variant annotation pipelines (e.g., annotating VCF variants against ENCODE peak sets)
• Exploring which experiments are available for a biosample (cell line, tissue, developmental stage) before planning a wet-lab experiment
• Querying all ChIP-seq experiments for a transcription factor across multiple cell types for comparative regulatory analysis
• Use regulomedb-database instead when you want pre-computed regulatory scores for specific SNPs — RegulomeDB integrates ENCODE data with eQTL and motif evidence into a single score
• Use deeptools-ngs-analysis instead when you have your own BAM files and need to generate bigWig coverage tracks; ENCODE database is for retrieving existing deposited data

Prerequisites

• Python packages: requests, pandas, matplotlib
• Data requirements: experiment accessions (e.g., ENCSR000AKC), biosample names (e.g., K562), TF target names (e.g., CTCF, TP53), or genomic regions (chr7:117548628-117748628)
• Environment: internet connection; no authentication required for public data; add Authorization: Bearer {api_key} header for submitter access
• Rate limits: no published hard limit; add time.sleep(0.5) for large batch queries to avoid connection resets

pip install requests pandas matplotlib

Quick Start

import requests

BASE = "https://www.encodeproject.org"

def search_experiments(assay="TF ChIP-seq", target="CTCF", biosample="K562", limit=5):
    """Find ENCODE experiments matching assay type, target, and biosample."""
    params = {
        "type": "Experiment",
        "assay_title": assay,
        "target.label": target,
        "biosample_ontology.term_name": biosample,   # `biosample_summary` is a verbose freetext string; filter by ontology term name
        "status": "released",
        "format": "json",
        "limit": limit,
    }
    r = requests.get(f"{BASE}/search/", params=params, timeout=30)
    r.raise_for_status()
    data = r.json()
    experiments = data.get("@graph", [])
    print(f"Found {data.get('total', 0)} experiments for {target} ChIP-seq in {biosample}")
    for exp in experiments:
        print(f"  {exp['accession']}  {exp.get('biosample_summary', '')}  {exp.get('lab', {}).get('title', '')}")
    return experiments

exps = search_experiments(assay="TF ChIP-seq", target="CTCF", biosample="K562")

Core API

Query 1: Experiment Search — Find Experiments by Assay, Biosample, Target

Search the ENCODE Portal for experiments matching structured criteria.

import requests, pandas as pd

BASE = "https://www.encodeproject.org"

def search_experiments(assay_title=None, target=None, biosample=None,
                       organism="Homo sapiens", status="released", limit=50):
    """
    Search ENCODE experiments with flexible filters.
    Returns: pd.DataFrame of matching experiments.
    """
    params = {
        "type": "Experiment",
        "status": status,
        "replicates.library.biosample.donor.organism.scientific_name": organism,
        "format": "json",
        "limit": limit,
    }
    if assay_title:
        params["assay_title"] = assay_title
    if target:
        params["target.label"] = target
    if biosample:
        params["biosample_ontology.term_name"] = biosample   # filter by ontology term, not the freetext `biosample_summary`

    r = requests.get(f"{BASE}/search/", params=params, timeout=30)
    r.raise_for_status()
    data = r.json()
    total = data.get("total", 0)
    print(f"Total matching experiments: {total} (showing {min(limit, total)})")

    records = []
    for exp in data.get("@graph", []):
        records.append({
            "accession": exp.get("accession"),
            "assay": exp.get("assay_title"),
            "biosample": exp.get("biosample_summary"),
            "target": exp.get("target", {}).get("label", ""),
            "lab": exp.get("lab", {}).get("title", ""),
            "date_released": exp.get("date_released", ""),
        })
    df = pd.DataFrame(records)
    print(df.to_string(index=False))
    return df

# CTCF ChIP-seq in HCT116 colon cancer cells
df = search_experiments(assay_title="TF ChIP-seq", target="CTCF", biosample="HCT116")

# ATAC-seq experiments in multiple cell types
df_atac = search_experiments(assay_title="ATAC-seq", limit=20)
print(f"\nUnique cell types: {df_atac['biosample'].nunique()}")

Query 2: File Download — Get Metadata and Download URLs for BED/bigWig Files

Retrieve file metadata for a specific experiment and obtain download URLs.

import requests, pandas as pd

BASE = "https://www.encodeproject.org"

def get_experiment_files(accession, file_format="bed", output_type="peaks",
                         assembly="GRCh38"):
    """
    Get file download URLs for a specific ENCODE experiment.
    accession: experiment accession, e.g. 'ENCSR000AKC'
    file_format: 'bed', 'bigWig', 'fastq', 'bam'
    output_type: 'peaks', 'signal', 'alignments', 'reads'
    Returns: pd.DataFrame of matching files with download URLs.
    """
    params = {
        "type": "File",
        "dataset": f"/experiments/{accession}/",
        "file_format": file_format,
        "output_type": output_type,
        "assembly": assembly,
        "status": "released",
        "format": "json",
        "limit": 50,
    }
    r = requests.get(f"{BASE}/search/", params=params, timeout=30)
    r.raise_for_status()
    data = r.json()
    files = data.get("@graph", [])
    print(f"Found {len(files)} {file_format} files ({output_type}) for {accession}")

    records = []
    for f in files:
        records.append({
            "file_accession": f.get("accession"),
            "file_format": f.get("file_format"),
            "output_type": f.get("output_type"),
            "assembly": f.get("assembly"),
            "file_size_mb": round(f.get("file_size", 0) / 1e6, 2),
            "download_url": BASE + f.get("href", ""),
            "biological_replicate": str(f.get("biological_replicates", [])),
        })
    df = pd.DataFrame(records)
    print(df[["file_accession", "output_type", "assembly", "file_size_mb"]].to_string(index=False))
    return df

# Get peak BED files for a CTCF ChIP-seq experiment
df_files = get_experiment_files("ENCSR000AKC", file_format="bed", output_type="peaks")

import urllib.request

def download_encode_file(download_url, output_path):
    """Download an ENCODE file from its href URL."""
    print(f"Downloading {download_url} → {output_path}")
    urllib.request.urlretrieve(download_url, output_path)
    import os
    size_mb = os.path.getsize(output_path) / 1e6
    print(f"Downloaded {size_mb:.1f} MB → {output_path}")

# Download the first peak file
if len(df_files) > 0:
    url = df_files.iloc[0]["download_url"]
    download_encode_file(url, "CTCF_peaks.bed.gz")

Query 3: cCRE Region Query — ENCODE SCREEN Candidate cis-Regulatory Elements

Query ENCODE SCREEN for candidate cis-Regulatory Elements in a genomic region using the SCREEN API.

import requests, pandas as pd

SCREEN_BASE = "https://a