ENA: the European Nucleotide Archive

By the end of this lesson you will be able to find the exact same BRCA1 sequence in Europe that you already know from GenBank, read its European accession, and pull thousands of records in bulk over FTP or Aspera — and, importantly, you will stop thinking of ENA and GenBank as rivals. They are two windows onto one shared archive.

ENA is GenBank's mirror, not its competitor

The GenBank lesson introduced the INSDC — the International Nucleotide Sequence Database Collaboration — the daily-exchange machinery that keeps three archives holding identical sequence records. ENA, the European Nucleotide Archive, is the European partner in that collaboration, run by EMBL-EBI. GenBank (at NCBI) and DDBJ (in Japan) are the other two.

"Partner" is the whole point. When a lab deposits a sequence at any one of the three, INSDC propagates it to the other two within about a day. So the 1994 BRCA1 mRNA you met as GenBank U14680 is not copied to ENA as some second-class duplicate — it is in ENA, as the same record, carrying the same sequence and the same core annotation. Asking "should I use GenBank or ENA for this sequence?" is like asking which mirror shows the real you. Pick whichever hub's tools you prefer; the DNA is identical.

What ENA holds that a single GenBank page does not surface

ENA is not only a mirror of annotated sequences. It organises the data into three broad layers, and the second and third are where it earns a lesson of its own:

That bottom layer is the same raw-read world you met as NCBI's SRA in the SRA lesson. ENA mirrors it under the name it uses in Europe, and it also holds the shared project and sample accessions you have already met — a PRJEB/PRJNA study, an SAMEA/SAMN sample. One BRCA1 sequencing study can therefore be walked top to bottom in ENA: from the project, to each biosample, to the raw run reads, to any assembled sequence that came out.

A practical difference worth knowing: ENA is often the friendlier place to download raw reads. It serves FASTQ directly, whereas SRA hands you its own .sra container that you unpack with fasterq-dump. If you have wrestled with that, the downloading from SRA and ENA walkthrough covers both routes.

The browser: reading one record

ENA's web browser is built around the accession-as-URL pattern you just used. Every INSDC record lives at a predictable address — .../ena/browser/view/<accession> — which makes ENA links easy to share and easy to script. The record view shows the same anatomy you learned to read in GenBank: the description, the source organism, the FEATURES table with its gene and CDS entries, and the cross-references (db_xref) that point out to other databases.

For a machine-readable copy, ENA exposes each record through a plain URL you can hit with any HTTP client:

# Fetch the BRCA1 mRNA as an EMBL-format flat file
curl "https://www.ebi.ac.uk/ena/browser/api/embl/U14680" -o brca1.embl

# Or the same record as FASTA
curl "https://www.ebi.ac.uk/ena/browser/api/fasta/U14680" -o brca1.fasta

Note the embl in that first URL. ENA's native flat file is the EMBL format — the same information as a GenBank flat file, reorganised into EMBL's two-letter line codes (ID, DE, FT, SQ) instead of GenBank's keywords. Biopython reads both, so the format difference almost never touches your code:

from Bio import SeqIO
import urllib.request

url = "https://www.ebi.ac.uk/ena/browser/api/embl/U14680"
urllib.request.urlretrieve(url, "brca1.embl")

rec = SeqIO.read("brca1.embl", "embl")     # note: "embl", not "genbank"
print(rec.id, len(rec.seq), "bp")          # U14680.1 5711 bp

The sequence length and content match what GenBank gives you base for base, because it is the same INSDC record.

Bulk download: FTP and Aspera

The browser is for one record at a time. When you need thousands — every sequence in a project, or a whole taxonomic slice — ENA's real strength is bulk transfer. It offers two channels:

• FTP / HTTPS — the universal option. Every ENA data file has a stable path under ftp.sra.ebi.ac.uk, so you can drive downloads with wget, curl, or a plain browser. Simple, portable, no special software.
• Aspera — a commercial high-speed protocol (client ascp) that saturates fast links far better than FTP for very large transfers, such as whole raw-read datasets. It needs the Aspera client installed, but for tens of gigabytes it is dramatically quicker.

To discover which files to pull, ENA provides a Portal API that returns a report — often as TSV — listing every run, sample, or file under a query, with its download URL. The usual workflow is: query the Portal API for a project's file list, then feed those URLs to wget or ascp.

# Ask the Portal API for every read run in a project, with FASTQ URLs
curl "https://www.ebi.ac.uk/ena/portal/api/filereport?accession=PRJEB1234&result=read_run&fields=run_accession,fastq_ftp&format=tsv"

That two-step pattern — list, then fetch — scales from a handful of records to an entire study without any clicking.

Where this leaves you

Carry one idea forward: for sequence data, Europe and NCBI are mirrors, and choosing between them is a matter of tooling and taste, not truth. The same BRCA1 mRNA, U14680, sits identically in both archives, reachable by the same accession, describable in either flat-file dialect. That symmetry does not carry over to everything at EMBL-EBI, though. Some European resources have no NCBI twin at all — they are genuinely distinct databases with their own model of the data, built around ideas GenBank never had to represent.