RXA

RXA performs relative expression analysis in an R package to analyze gene expression profiles in high-throughput datasets by focusing on the relative ordering of gene expressions to produce biologically interpretable predictive models.

Key Features:

• Relative Expression Analysis (RXA): Emphasizes relative expression levels and the ordering of genes rather than absolute expression values.
• Three-Gene Model: Implements a three-gene RXA variant that has been systematically compared with earlier methods across cancer studies.
• Biologically Relevant Decision Rules: Uses simple decision rules based on a reference gene's expression relative to two differentially expressed genes, with examples including PPP1CB and RNF14 for BRCA1 prediction.
• Biological Interpretation: Analyzes protein-protein interactions within gene triplets to provide a plausible biological basis for the decision rules.
• Clinical Applicability: Identifies genomic marker interactions with clear biological interpretations applicable to clinical problems such as BRCA1 mutation identification and estrogen receptor (ER) status prediction.

Scientific Applications:

• Cancer Research: Validated across multiple cancer studies and used to predict clinically relevant markers including germline BRCA1 mutations and estrogen receptor (ER) status in breast cancer.
• Genetic Studies: Detects marker interactions useful for understanding underlying disease mechanisms and gene relationships.

Methodology:

Focus on relative gene expression ordering; use of a three-gene RXA model; application of simple decision rules comparing a reference gene to two other genes (e.g., PPP1CB and RNF14); systematic comparison with earlier methods across cancer studies; analysis of protein-protein interactions within gene triplets.