GROOLS: reactive graph reasoning for genome annotation through biological processes.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• High-quality functional annotation is crucial for understanding genomic phenotypic consequences.
• Millions of genomic sequences lack reliable functional assignments despite bioinformatics advancements.
• Curation of protein functions within biological processes aids annotation evaluation.

Purpose of the Study:

• To develop an expert system for evaluating the completeness and consistency of predicted protein functions.
• To improve the accuracy of functional annotation in genomic databanks.
• To assist biocurators in identifying and resolving annotation discrepancies.

Main Methods:

• Developed GROOLS (Genomic Rule Object-Oriented Logic System), an expert system employing paraconsistent logic.
• Modeled biological processes, such as metabolic pathways, using a hierarchical knowledge representation and graph structure.
• Propagated observations (predictions, expectations) via rules to identify uncertainties and inconsistencies.

Main Results:

• Applied GROOLS to 14 microbial organisms, evaluating functional annotation completeness and consistency.
• The system successfully highlighted uncertainties and inconsistencies in predicted protein functions.
• Demonstrated the utility of GROOLS in assessing functional annotation accuracy.

Conclusions:

• GROOLS software evaluates the accuracy of functional unit and pathway predictions using experimental data like growth phenotypes.
• The system aids biocurators by focusing on missing or contradictory observations in protein functional annotation.
• Enhances the reliability of genomic functional annotation through logical reasoning and data integration.