RefBool

RefBool: Reference-based three-state gene expression discretization

RefBool classifies gene expression measurements into three discrete states (active, intermediate, inactive) using reference-guided thresholds and statistical validation to improve discretization accuracy.

Key Features:

• Three-State Classification: Assigns gene expression data to active, intermediate, or inactive states, capturing intermediate expression levels beyond binary discretization.
• Statistical Validation: Computes p-values and q-values for each gene classification to assess statistical significance and control false discoveries.
• Reference-Based Discretization: Utilizes reference datasets to define biologically grounded classification thresholds.

Scientific Applications:

• Gene Expression Analysis: Improves identification of molecular regulatory mechanisms, supports analysis of complex biological processes, and enhances clustering performance, including in neuroepithelial differentiation studies.

Methodology:

RefBool applies reference datasets to guide discretization of gene expression measurements into three states, calculates p-values and q-values to validate each classification, and integrates statistically supported assignments to produce granular, biologically informed expression states.