Gene Ontology (GO)

Describes the larger biological programs accomplished by multiple molecular activities. BP terms capture pathways and processes rather than single reactions.

• Examples: 'DNA repair', 'signal transduction', 'inflammatory response'
• Key distinction: A biological process is not equivalent to a pathway; it represents a recognized series of events
• Scope: Includes developmental processes, response to stimuli, and metabolic programs
• Annotation logic: A gene product is annotated to a BP if it contributes to that process

Captures the elemental activities of a gene product at the molecular level. MF terms describe what a molecule does, not where or when it acts.

• Examples: 'catalytic activity', 'transporter activity', 'binding'
• Key distinction: MF describes single-step activities, not multi-step processes
• Scope: Includes enzymatic reactions, receptor binding, and structural molecule activity
• Annotation logic: A gene product is annotated to an MF based on its direct biochemical activity

Specifies the subcellular locations and macromolecular complexes where a gene product is active. CC terms anchor function and process to physical space.

• Examples: 'mitochondrion', 'nucleoplasm', 'ribosome'
• Key distinction: CC describes where a gene product executes its function, not the process itself
• Scope: Includes organelles, protein complexes, and membrane domains
• Annotation logic: A gene product is annotated to a CC if it is physically located there

Every GO annotation is accompanied by an evidence code that indicates the type of supporting data, enabling users to filter by confidence level.

• Experimental evidence codes: IDA (Inferred from Direct Assay), IPI (Inferred from Physical Interaction), IMP (Inferred from Mutant Phenotype)
• Computational evidence codes: IEA (Inferred from Electronic Annotation), ISS (Inferred from Sequence or Structural Similarity)
• Curator-assigned codes: TAS (Traceable Author Statement), NAS (Non-traceable Author Statement)
• ND (No biological Data): Explicitly marks genes with no known function

A GO domain describing the larger biological programs accomplished by multiple molecular activities. Biological Process terms capture pathways and transformations, such as 'signal transduction' or 'DNA repair', rather than the individual molecular steps that execute them. A single process can involve numerous gene products collaborating across time and space. Unlike Molecular Function, BP terms are defined by their outcome, not their mechanism.

The elemental, persistent activities of a gene product at the molecular level. Molecular Function terms describe what a protein or RNA does without specifying where or when it acts. Examples include 'catalytic activity', 'transporter activity', and 'binding'. MF terms represent single-step actions, not multi-step pathways, and are independent of the larger biological context in which they operate.

The subcellular structures and macromolecular complexes where a gene product is active. Cellular Component terms define locations relative to cellular anatomy, such as 'mitochondrial inner membrane', 'nucleolus', or 'proteasome complex'. CC annotation does not describe processes or functions but provides the spatial context essential for understanding where a gene product executes its role.

A controlled vocabulary documenting the type of experimental or computational support backing each annotation. Key evidence codes include:

• IDA (Inferred from Direct Assay): Direct experimental validation
• IMP (Inferred from Mutant Phenotype): Observed variation in a process due to genetic alteration
• IEA (Inferred from Electronic Annotation): Automated computational prediction, requiring manual review for high-confidence use
• IPI (Inferred from Physical Interaction): Protein-protein or protein-DNA binding evidence

The foundational logical constraint of the GO directed acyclic graph (DAG). The True Path Rule states that if a gene product is annotated to a child term, it is automatically annotated to all parent terms along the path to the root. This ensures transitive closure and enables consistent querying across granularity levels. Violating this rule breaks the ontology's inferential reasoning capabilities.

A structured record linking a specific gene product to a GO term with supporting metadata. Each GO Annotation captures:

• The gene product identifier (UniProtKB, MGI, etc.)
• The GO term accession
• An evidence code and supporting reference (PMID or DOI)
• The assigned date and contributing database Annotations are the atomic units enabling large-scale functional enrichment analysis.