Archival data

• repository of information
• redundant; might have many sequence records for the same gene, each from a different lab
• submitters maintain editorial control over their records:
  what goes in is what comes out
• no controlled vocabulary
• variation in annotation of biological features

Curated data

• non-redundant; one record for each gene, or each splice variant
• each record is intended to present an encapsulation of the current understanding of a gene or protein, similar to a review article
• records contain value-added information that have been added by an expert(s)