Image source: GenBank Submission Handbook.
Quality Control - methods
- Data formatting guidelines for repositories ensure that the submitted data will be compatible with the database design.
- Automatic, computerized data retrieval from other databases ensures that the same procedure is used each time data are downloaded. Data quality concerns are minimized, but some possible concerns include:
- Factors that affect data transmission and import into the receiving database.
- Incorrect entry of a record or an identifier in the collaborating database may create an incorrect record in the receiving database. Databases have different levels of funding and technical support so this may be a concern with some smaller databases but not in others, like those nationally supported institutions.
- Manual curation procedures should be documented to minimize the variability in subjective decisions by curators. For example:
- Curators read the scientific literature and extract information to be input into a database. They prioritize and select papers to review, then they must distinguish between experimentally supported information vs inferential assertions. Automated, electronic curation may capture inferential assertions unintetionally. (Hirschman et al. 2010).
- One gene name may have multiple literature-based synonyms and a synonym may be associated with different genes (Howe et al., 2008).