GO annotation in InterPro: why stability does not indicate accuracy in a sea of changing annotations

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Annotation stability in biological databases is often equated with annotation reliability.
• However, various data-driven events can impact annotation stability in protein sequence and family databases.

Purpose of the Study:

• To investigate the factors influencing annotation stability within biological databases.
• To analyze the consequences of these events on the InterPro database.
• To demonstrate that annotation removal or reassignment is not solely indicative of erroneous curation.

Main Methods:

• Analysis of annotation changes within the InterPro database.
• Identification and categorization of data-driven events affecting annotation stability.
• Case studies illustrating the impact of these events on annotation reassignment.

Main Results:

• Annotation removal or reassignment in InterPro is not always due to curator error.
• Data-driven updates and re-analyses can lead to legitimate annotation changes.
• Annotation stability is a complex metric influenced by multiple factors beyond initial curation accuracy.

Conclusions:

• Annotation stability should be interpreted cautiously, considering the dynamic nature of biological data.
• The InterPro database provides a case study for understanding annotation dynamics.
• Rethinking the direct correlation between annotation removal and annotation inaccuracy is necessary.