Longevity of Biological Databases

TLDR

Public web‑based databases are essential for biological research by storing past results, guiding future work, and providing broad access, yet their longevity remains uncertain as many new databases are created each year. This study examined the 18‑year survival of 326 biological databases. Over 60 % of databases were dead within 18 years, 14 % archived, and those that survived tended to be institutionally important with core support; longevity depends on infrastructure and long‑term financial strategies, underscoring the need for researchers and funders to consider data security and investment ramifications when creating independent databases.

Abstract

Public Web-based databases are essential for present-day biological research: they i) store the results of past laboratory experiments; ii) guide the focus of future ones; and, iii) allow all to benefit from the wealth of information they contain. Many new databases are born each year; but how long do they live? This study looked at the 18-year survival of 326 databases. Over 60% were dead within that time period, and a further 14% were archived, no longer updated. Those that survived were, for the most part, important to their institution’s main focus, and had core institutional support. Database longevity depends on the existence of infrastructures that are underpinned by long-term financial strategies. Researchers and funders need to consider the ramifications for the security of their data, and of the financial investments in them, if they choose to create new databases independently of core infrastructures.