BioC viewer: a web-based tool for displaying and merging annotations in BioC.

Area of Science:

• Bioinformatics
• Computational Biology
• Text Mining

Background:

• BioC is an XML-based format for text mining and curation, but annotation misalignment is a common challenge.
• Different software environments (e.g., ASCII vs. Unicode) can cause text and annotation mismatches.
• The BioC Viewer was initially developed for BioCreative V to aid curators in handling BioC annotations.

Purpose of the Study:

• To describe the BioC Viewer tool and its improvements for addressing BioC annotation misalignment.
• To enhance interoperability between text mining tools and manual curation results.
• To provide a user-friendly interface for biological data curation.

Main Methods:

• Implemented a BioC Viewer with a merge process for BioC files from the same article.
• Developed an automatic offset adjustment mechanism to correct mismatches between annotated text and actual text.
• Designed a user-friendly web interface for efficient curation operations.

Main Results:

• The BioC Viewer successfully assists in curating protein-protein and genetic interaction pairs from full-text articles.
• Offset adjustment effectively resolves misalignment issues caused by different text encoding environments.
• The tool offers a merge process for consolidating BioC files from the same source.
• Positive feedback from curators highlights the web interface's usability and learnability.

Conclusions:

• The improved BioC Viewer effectively addresses BioC annotation misalignment, enhancing data interoperability.
• The tool's user-friendly interface and automatic correction features streamline biological data curation.
• BioC Viewer facilitates more accurate and efficient text mining and manual curation workflows.