Onto-CC: a web server for identifying Gene Ontology conceptual clusters

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Gene Ontology (GO) is widely used for analyzing coexpressed genes, but faces uncertainty regarding ontology choice and specificity.
• GO annotations can be incomplete or biased by existing knowledge, leading to potential inaccuracies in gene grouping hypotheses.
• Current methods often overlook these limitations, using GO terms for direct gene expression profiling instead of independent validation.

Purpose of the Study:

• To introduce Onto-CC, a web tool designed for the independent explanation and validation of gene grouping hypotheses using GO clusters.
• To reduce query uncertainty by identifying optimal conceptual clusters that integrate terms from different GO ontologies and specificity levels.
• To provide alternative, optimal explanations for biological queries, generating novel insights.

Main Methods:

• Implemented the EMO-CC methodology, inspired by Conceptual Clustering algorithms, to find clusters within the GO Directed Acyclic Graph (DAG) tree.
• Utilized multiobjective/multimodal optimization techniques to manage optimal cluster sets as parallel hypotheses.
• Developed Onto-CC as an automatic method for GO-based gene grouping hypothesis validation (expression versus GO).

Main Results:

• Onto-CC successfully reduces uncertainty in gene function analysis by combining GO terms across ontologies and specificity levels.
• The tool generates optimal cluster sets, offering parallel hypotheses and alternative explanations for biological queries.
• Demonstrated successful application in testing medical and biological hypotheses, including predicting gene expression profiles in inflammatory responses.

Conclusions:

• Onto-CC provides a robust approach for validating gene grouping hypotheses, mitigating inherent uncertainties in GO annotations.
• The tool facilitates deeper biological insights by offering novel explanations and predictions of gene expression.
• Onto-CC is available in two versions: Ready2GO for precalculated genomes and Advanced Onto-CC for custom annotations.