MORE

MORE constructs phenotype-specific regulatory networks by integrating multi-omic data and identifying significant regulators using regression and regularization methods.

Key Features:

• Advanced Regression Techniques: Implements Partial Least Squares (PLS) or Multiple Linear Regression (MLR) augmented with Elastic Net or Iterative Sparse Group Lasso (ISGL) regularizations to model target omic expressions.
• Versatility Across Omics Layers: Supports integration of any number and type of omics layers for multi-omic analyses.
• Integration of Prior Knowledge: Incorporates prior regulatory knowledge to inform and refine regulator identification.
• Phenotype-Specific Network Construction: Constructs condition- or phenotype-specific regulatory networks to capture regulatory mechanisms linked to diseases or treatments.

Scientific Applications:

• Disease Mechanism Elucidation: Identifies phenotype-specific regulatory mechanisms relevant to diseases such as cancer.
• Biomarker Discovery and Validation: Pinpoints significant regulators associated with specific conditions or treatments to support biomarker discovery and validation.
• Personalized Medicine: Links regulatory mechanisms to distinct survival outcomes, as demonstrated in an ovarian cancer dataset, to support personalized therapeutic strategies.

Methodology:

Performs regression-based modeling using PLS or MLR with Elastic Net or ISGL regularization and variable selection to identify significant regulators by modeling target omic expressions as functions of experimental variables (e.g., diseases or treatments) and potential regulators; evaluated on simulated datasets and benchmarked against state-of-the-art tools for regulator identification accuracy, model goodness-of-fit, and computational efficiency.