IsoDA

IsoDA predicts isoform-disease associations (IDAs) by integrating genome, transcriptome, and proteome data to identify disease-linked isoforms.

Key Features:

• Alternative Splicing and Functional Diversity: Accounts for alternative splicing to model functional diversity among protein isoforms relevant to disease associations.
• Isoform-Disease Association Focus: Targets isoform-level associations (IDAs) rather than aggregate gene-disease associations (GDAs) to provide more granular links between isoforms and diseases.
• Data Fusion via Joint Matrix Factorization: Integrates multiomics data (genome, transcriptome, proteome) through joint matrix factorization to combine complementary molecular information.
• Dispatch and Aggregation Mechanism: Implements a dispatch and aggregation term to distribute gene-disease associations to individual isoforms and aggregate isoform associations back to host genes.
• Performance and Validation: Demonstrates improved performance over existing methods at gene and isoform levels and identifies apolipoprotein E isoforms associated with Alzheimer's disease and vascular endothelial growth factor A isoforms linked to coronary heart disease.

Scientific Applications:

• Isoform-level disease association mapping: Enables identification of specific isoforms implicated in diseases for more precise molecular characterization.
• Biomarker and therapeutic target discovery: Supports discovery of isoform-specific biomarkers and potential therapeutic targets.
• Molecular mechanism elucidation: Facilitates investigation of how alternative splicing and isoform variation contribute to disease pathology.
• Comparative gene vs. isoform analyses: Provides a framework to compare gene-level and isoform-level associations to refine disease-related hypotheses.

Methodology:

Integrates genome, transcriptome, and proteome data using joint matrix factorization and employs a dispatch-and-aggregation term to map gene-disease associations to isoforms and aggregate isoform associations to genes.