GOexpress: an R/Bioconductor package for the identification and visualisation of robust gene ontology signatures...

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Identifying gene expression profiles is crucial for understanding molecular pathways and finding biomarkers.
• Current methods often focus on simple two-group comparisons, limiting analysis of complex experimental designs.
• Advanced methods like functional class scoring and machine learning offer powerful alternatives for pathway analysis.

Purpose of the Study:

• To introduce GOexpress, a software package for scoring gene ontology features to classify samples from multiple experimental groups.
• To enable the identification of gene panels and pathways relevant to sample classification.
• To provide a tool for analyzing complex experimental data, including continuous and categorical factors.

Main Methods:

• GOexpress integrates normalized gene expression data (microarray, RNA-seq) with sample phenotypic information and gene ontology annotations.
• It employs a supervised learning approach, with a default random forest algorithm.
• The method competitively scores genes to determine their importance in classifying predefined sample groups.

Main Results:

• GOexpress effectively scores and summarizes the capacity of gene ontology features to classify samples across multiple experimental groups.
• It generates rankings of genes and gene ontology terms based on their classification power.
• The software facilitates the evaluation of gene importance in distinguishing experimental conditions.

Conclusions:

• GOexpress enables rapid identification and visualization of gene panels that robustly classify sample groups.
• It supports both categorical and continuous experimental factors, enhancing analytical flexibility.
• The package integrates seamlessly with existing computational biology workflows via standard Bioconductor packages and public gene ontology annotations.