SINBAD

SINBAD identifies splice variants associated with patient survival and links them to altered protein-protein interactions and mutational signatures, including impacts on DNA repair via homologous recombination.

Key Features:

• Identification of Survival-significant Splice Variants: Identifies splice variants associated with patient survival from large-scale cancer datasets such as The Cancer Genome Atlas (TCGA) using statistical analysis.
• Prediction of Rewired Protein-Protein Interactions: Predicts alterations in protein-protein interactions caused by alternative splicing through construction of Multi-Granularity Graphs to identify predicted interaction gains and losses.
• Focus on DNA Repair Mechanisms: Links identified splice variants to biological processes such as DNA repair via homologous recombination.
• Computational Validation and Hypothesis Generation: Generates experimentally testable hypotheses and validates proposed mechanisms through mutational signature analysis that correlates splice-variant expression levels with mutation patterns.
• Statistical Robustness: Applies the Null Empirically Estimated P-value (NEEP) method to assess statistical significance of survival-associated splice variants.

Scientific Applications:

• Cancer Biomarker Discovery: Supports discovery of splice-variant biomarkers associated with patient survival for cancer diagnosis and prognosis.
• Mechanistic Insights into Cancer Progression: Provides mechanistic links between alternative splicing, rewired protein-protein interactions, and pathways driving tumor initiation, progression, and invasion.
• Personalized Medicine: Enables stratification of patients by splice-variant profiles to inform personalized therapeutic strategies.

Methodology:

Analyzes large-scale genomic datasets (e.g., TCGA) for alternative splicing, applies NEEP-based statistical tests to identify survival-associated splice variants, constructs Multi-Granularity Graphs to predict protein-protein interaction gains and losses, and performs mutational signature analysis for validation.