GOexpress

GOexpress classifies samples and ranks genes and Gene Ontology (GO) terms by integrating normalized microarray and RNA-seq gene expression data with supervised learning (random forest) to identify ontology-related molecular signatures.

Key Features:

• Integration with Gene Expression Data: Integrates normalized gene expression data from microarray and RNA-seq experiments for downstream analysis.
• Supervised Learning (Random Forest): Uses a supervised learning framework, primarily the random forest algorithm, to evaluate interactions among experimental factors for sample classification.
• Functional Class Scoring: Applies functional class scoring to assess and rank genes and Gene Ontology terms by their ability to classify samples across multiple predefined groups.
• Competitive Gene Scoring: Performs competitive scoring of expressed genes to determine their relative importance in distinguishing sample groups.
• Support for Diverse Experimental Factors: Handles categorical factors (e.g., infection status, treatment) and continuous variables (e.g., time-series, drug concentration) in analyses.
• Visualization Capabilities: Provides visualization of gene expression profiles to aid interpretation of classification results.
• Bioconductor Integration and GO Annotations: Leverages Bioconductor extension packages and publicly available Gene Ontology annotations for analyses.

Scientific Applications:

• Pathway and Ontology Analysis: Performs pathway-level and Gene Ontology-based analyses to interpret molecular measurements.
• Complex Experimental Designs: Enables analysis of studies with multi-level categorical factors and continuous variables.
• Biomarker Discovery: Identifies ontology-related gene panels and candidate diagnostic or prognostic biomarkers that classify sample groups.
• Differential Expression Interpretation: Enhances interpretation of differential expression data by ranking genes and GO terms according to classification performance.

Methodology:

Integrates normalized microarray and RNA-seq expression data, applies supervised learning with random forest to evaluate factor interactions, uses functional class scoring and competitive gene scoring to rank genes and GO terms, and employs Bioconductor packages with public Gene Ontology annotations.