NsRRR

NsRRR models associations between DNA methylation and gene expression by integrating gene interaction networks within a network-sparse reduced-rank regression framework.

Key Features:

• Multivariate regression framework: Employs a reduced-rank multivariate regression model to jointly relate high-dimensional DNA methylation and gene expression data.
• Utilization of gene interaction networks: Incorporates prior biological knowledge in the form of gene interaction networks to guide model estimation.
• Focus on epigenetic regulation: Specifically targets DNA methylation and its influence on transcriptional profiles, with relevance to cancer-associated epigenetic alterations.
• Statistical rigor and variable selection: Emphasizes accurate variable selection and demonstrates improved variable selection accuracy in simulation studies compared to models that omit network information.

Scientific Applications:

• Cancer research: Applied to datasets such as The Cancer Genome Atlas, including primary ovarian tumors, to investigate relationships between methylation and expression.
• Epigenetic studies: Used to elucidate how DNA methylation influences gene expression across biological contexts.

Methodology:

Implements a network-sparse reduced-rank regression framework that integrates gene interaction networks with DNA methylation and gene expression data and emphasizes variable selection, with validation by simulation studies.