EvidenceAggregatedDriverRanking

EvidenceAggregatedDriverRanking: Multi-omic driver gene prioritization for late-onset Alzheimer’s disease

EvidenceAggregatedDriverRanking identifies and ranks candidate driver genes associated with late-onset Alzheimer’s disease by aggregating multi-omic analytic outcomes and integrating external validation using Genome Wide Association Study (GWAS) summary statistics.

Key Features:

• Multi-Omic Data Integration: Aggregates multiple genomic data types and analytic outcomes to combine evidence across modalities and capture complex genetic interactions relevant to Alzheimer’s disease.
• Machine Learning Framework: Learns feature representations from known driver genes across diverse feature sets and predicts additional candidate driver genes with similar representations.
• Flexible Ranking Scheme: Prioritizes candidate driver genes using an evidence-aggregated ranking that incorporates GWAS summary statistics for external validation.
• Generalizability: Applies to RNA-Seq data and is designed to adapt to other data modalities and analysis types.

Scientific Applications:

• Missing-Label Prediction in Multiview Datasets: Predicts missing labels in benchmark multiview datasets and supports discovery of novel candidate genetic drivers of late-onset Alzheimer’s disease.
• Pathway Enrichment Analysis: Produces ranked gene sets enriched for Alzheimer’s-associated single nucleotide polymorphisms (SNPs) and pathways previously linked to the disease.

Methodology:

Trains a machine learning model on known driver genes using multiple feature sets, identifies genes with similar feature representations, and ranks candidates by aggregated evidence; the ranking is refined using GWAS summary statistics for external validation.