NETPAGE

NETPAGE applies network propagation to integrate tissue-specific gene interaction networks with rare variant data to generate smoothed gene scores for gene-based association testing and disease prediction.

Key Features:

• Network Propagation Framework: Models information flow across gene interaction networks to simulate how genetic variation percolates and to produce smoothed gene scores.
• Tissue-specific Interaction Integration: Incorporates tissue-specific gene interaction networks (for example, hippocampus networks) to provide context-aware propagation of variant effects.
• Sequencing Data Integration: Integrates whole-genome sequencing (WGS) and whole-exome sequencing (WES) data for gene-level analysis.
• Smoothed Gene Scores: Produces gene scores that reflect the cumulative impact of rare variants within biological pathways and network neighborhoods.
• Sparse Regression for Prediction: Uses sparse regression to predict disease status and to identify connected genes whose mutation profiles are predictive of case-control status.

Scientific Applications:

• Alzheimer’s disease association testing: Applied to sporadic late-onset Alzheimer’s disease using WGS from the AD Neuroimaging Initiative (ADNI) cohort and WES from the AD Sequencing Project (ADSP) to identify predictive gene sets.
• Tissue-specific discovery: Identified connected genes within hippocampus interaction networks whose smoothed mutation profiles predict case-control status.
• Disease progression assessment: Employed smoothed gene scores and sparse regression to assess risk and prediction of conversion from Mild Cognitive Impairment (MCI) to Alzheimer’s disease.
• Transcriptomic and enrichment validation: Demonstrated correlations between smoothed gene scores and AD risk, revealed tissue-specific transcriptional dysregulation in independent RNA-seq datasets, and highlighted enrichments in AD-related gene sets and terms.

Methodology:

Integrates WGS and WES data with tissue-specific gene interaction networks; simulates genetic variation flow through networks via network propagation to generate smoothed gene scores; applies sparse regression models to predict disease status and assess progression.