The Biomolecular Interaction Network Database in PSI-MI 2.5

Database (Oxford). 2011 Jan 12:2011:baq037. doi: 10.1093/database/baq037. Print 2011.


The Biomolecular Interaction Network Database (BIND) is a major source of curated biomolecular interactions, which has been unmaintained for the last few years, a trend which will eventually result in the loss of a significant amount of unique biomolecular interaction information, mostly as database identifiers become out of date. To help reverse this trend, we converted BIND to a standard format, Proteomics Standard Initiative-Molecular Interaction 2.5, starting from the last curated data release (from 2005) available in a custom XML format and made the core components (interactions and complexes) plus additional valuable curated information available for download ( Major work during the conversion process was required to update out of date molecule identifiers resulting in a more comprehensive conversion of BIND, by measures including number of species and interactor types covered, than what is currently accessible elsewhere. This work also highlights issues of data modeling, controlled vocabulary adoption and data cleaning that can serve as a general case study on the future compatibility of interaction databases. Database URL:

Publication types

  • Research Support, N.I.H., Extramural
  • Research Support, Non-U.S. Gov't

MeSH terms

  • Animals
  • Databases, Protein*
  • Humans
  • Internet
  • Protein Binding
  • Proteomics / standards*
  • Rats
  • Reproducibility of Results
  • Vocabulary, Controlled