GOfuncR

GOfuncR performs Gene Ontology (GO) enrichment analysis in R/Bioconductor using the FUNC statistical framework to provide multiple enrichment tests and permutation-based family-wise error control while using OrganismDb/OrgDb (default Homo.sapiens) annotations and a bundled GO ontology graph.

Key Features:

• FUNC-based framework: Implements the statistical framework of the FUNC ontology enrichment software for structured GO analysis.
• Annotation sources: Retrieves GO annotations from OrganismDb or OrgDb packages with the default annotation package Homo.sapiens.
• Bundled ontology graph: Includes a GO ontology graph bundled within the package that is updated regularly.
• Hypergeometric test: Implements the standard candidate-versus-background enrichment test using the hypergeometric distribution.
• Wilcoxon rank-sum test: Provides a Wilcoxon rank-sum test for genes associated with continuous or ranked statistics.
• Binomial test: Offers a binomial test for analyses involving two counts per gene.
• Fisher’s exact and Chi-square tests: Supports Fisher’s exact or Chi-square tests for genes associated with a 2×2 count structure.
• Permutation-based multiple-testing correction: Controls family-wise error rates under dependency from the ontology hierarchy using permutation-derived empirical null distributions from randomization of gene-associated variables.
• GO graph and annotation interrogation: Provides tools for exploring the GO graph and interrogating annotation structure.
• Incorporation of biological covariates: Allows inclusion of covariates such as gene length or genomic proximity that can influence enrichment results.
• Custom resources support: Accepts custom gene coordinates, bespoke ontologies, or alternative annotation resources for specialized organisms or custom functional hierarchies.

Scientific Applications:

• High-throughput genomics studies: Statistical GO enrichment analysis for genome-scale experiments using various gene-associated statistics.
• Ranked or continuous gene metrics: Enrichment testing when genes are associated with continuous or ranked statistics via Wilcoxon rank-sum.
• Two-count gene data: Analyses where each gene has two counts using the binomial test.
• 2×2 count structures: Enrichment analyses for genes represented by 2×2 count tables using Fisher’s exact or Chi-square tests.
• Specialized organisms and custom ontologies: Functional analyses for non-standard organisms or bespoke functional hierarchies using custom annotations and ontologies.

Methodology:

Implements hypergeometric, Wilcoxon rank-sum, binomial, Fisher’s exact and Chi-square tests; retrieves annotations from OrganismDb/OrgDb (default Homo.sapiens) and uses a bundled GO ontology graph; evaluates term significance and controls family-wise error via permutation-based empirical null distributions derived from randomization of gene-associated variables.