GenBank won biology's data by getting journals to refuse papers without its ID

By the late 1970s, gene sequencing was producing more data than journals could print. A 1979 meeting at Rockefeller University asked how to manage the flood; back at Los Alamos, physicist Walter Goad began collecting sequences and writing analysis software. In 1982 the NIH funded his proposal, and GenBank was born. By the end of 1983 it held over 2,000 sequences — about two million base pairs, Los Alamos' obituary for Goad records.

The decisive design was institutional, not technical. Depositors received a unique accession number, and journals enforced the habit: as NCBI's own Nucleic Acids Research paper puts it, 'many journals require authors with sequence data to submit the data to a public sequence database as a condition of publication.' No number, no paper — so every researcher deposited before citing.

Openness removed every rival's reason to exist. GenBank's data was free and unrestricted, and through the INSDC partnership with Europe's ENA and Japan's DDBJ, the three archives exchanged data daily — one deposit, anywhere, served the whole world.

The flywheel ran for decades. Los Alamos counted about 9 million sequences (10 billion bases) by 2000; GenBank's own release notes (release 273.0) now count 267 million traditional records holding 8.2 trillion bases, plus 6.4 billion set-based records holding 51.8 trillion bases.

Make the database ID a ticket to publication

Journals did the enforcing: no accession number, no publication, so compliance cost editors and researchers almost nothing.

The accession number doubled as a permanent citation handle, making papers that used GenBank more useful, not merely compliant.

Free, unrestricted redistribution removed the price and licensing friction a commercial archive would have faced.

Daily exchange with ENA and DDBJ through INSDC meant one deposit reached the whole world, leaving regional rivals no niche.

Embed your artifact in a workflow people cannot skip. A standard that becomes a citation requirement enforces itself — no sales team, no lock-in contract needed.

GenBank became the default memory of molecular biology: sequencers, pipelines and papers assume it exists. Release 273.0 (2026) counts 267 million traditional records with 8.2 trillion bases, and the journal mandate it pioneered became the template for data deposition across science — from protein structures to clinical trials.