Quality controls in integrative approaches to detect errors and inconsistencies in biological databases

Area of Science:

• Bioinformatics
• Genomics
• Proteomics

Background:

• Biomolecular data are fragmented across databases, often containing errors.
• Effective data integration and quality control are essential for deriving new scientific knowledge.
• Existing data integration methods require enhancement for comprehensive analysis and quality assurance.

Purpose of the Study:

• To develop and implement automatic procedures for maintaining and ensuring the quality of integrated biomolecular data.
• To enhance the GFINDer web system's data warehouse (GPDW) with robust quality control mechanisms.
• To identify and rectify errors and inconsistencies in publicly available biomolecular datasets.

Main Methods:

• Developed automatic procedures for updating and quality control of the genomic and proteomic data warehouse (GPDW).
• Implemented diverse data quality control techniques, including checks for structure, completeness, ontological consistency, ID evolution, quantification, and source consistency.
• Integrated controlled terminologies and ontologies to describe gene and gene product functions and phenotypes.

Main Results:

• Identified various types of errors and inconsistencies in data from multiple biological databases.
• Successfully ensured a high quality of integrated data within the GFINDer data warehouse.
• Reported identified data issues to original database curators, leading to corrections and improved data quality.

Conclusions:

• Automatic data integration and quality control are vital for reliable biomolecular data analysis.
• The GFINDer system effectively manages and validates integrated data, supporting scientific research.
• Collaborative efforts in reporting data errors significantly improve the quality of shared biological information.