ASpediaFI

ASpediaFI integrates alternative splicing (AS) events, gene expression, and pathway information to identify differential alternative splicing (DAS) events, co-regulated genes, and functionally relevant pathways.

Key Features:

• Heterogeneous Interaction Network Construction: Constructs an interaction network that integrates genes, AS events, and pathways by connecting feature nodes through co-expression relationships or pathway gene set knowledge.
• Random Walk with Restart Algorithm: Employs the random walk with restart algorithm to interrogate proximity from a query gene set and prioritize AS events, genes, and pathways within the network.
• DAS and Co-regulated Gene/Pathway Identification: Systematically identifies differential alternative splicing (DAS) events, co-regulated genes, and pathways relevant to biological processes using the integrated network.
• Performance Evaluation: Performance was evaluated using simulated RNA sequencing (RNA-seq) datasets across varying sequencing depth and sample size and compared with other methods, and the approach was applied to three public datasets from cancer patients or cell lines.
• Integration with Global Co-expression Networks: Leverages global co-expression networks and pathway information to assess the functional importance of spliced genes within subnetworks.

Scientific Applications:

• Cancer Research: Applied to public datasets from cancer patients or cell lines to uncover AS events and pathways implicated in tumorigenesis and progression.
• Phenotype and Disease Studies: Delineates biological processes through DAS events to investigate genetic underpinnings of phenotypes and diseases.

Methodology:

Constructs a heterogeneous network integrating genes, AS events, and pathways via co-expression relationships or pathway gene set knowledge and applies the random walk with restart algorithm to assess proximity from a query gene set; performance was assessed using simulated RNA-seq datasets across sequencing depth and sample size and validated on public cancer datasets.