pBRIT: gene prioritization by correlating functional and phenotypic annotations through integrative data fusion.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Computational gene prioritization is crucial for identifying genes associated with diseases.
• Existing methods may not effectively handle the complexity and sparsity of biological annotation data.

Purpose of the Study:

• To introduce pBRIT (prioritization using Bayesian Ridge regression and Information Theoretic model), a novel, adaptive, and scalable tool for computational gene prioritization.
• To integrate multiple diverse biological data sources for improved gene identification accuracy.

Main Methods:

• pBRIT employs an Information-Theoretic approach to model feature dependencies and sparsity.
• Bayesian Ridge regression is used to learn mappings between functional and phenotype annotations.
• Genes are prioritized based on phenotypic concordance with known disease-associated genes.

Main Results:

• pBRIT achieved high performance with Area Under the Curve (AUC) scores ranging from 0.92 to 0.96 on benchmark datasets.
• The tool demonstrated strong performance (AUC 0.80) on time-stamped Human Phenotype Ontology (HPO) entries, indicating good sensitivity and specificity.
• pBRIT showed stable performance despite changes in annotation data and is fast and scalable.

Conclusions:

• pBRIT is an effective and robust tool for gene prioritization in disease gene identification.
• Its ability to handle data sparsity and dependencies, combined with high performance, makes it suitable for routine bioinformatics pipelines.
• The tool is publicly available for broader research application.