Bioinformatics statistics · BEGINNER LESSON
Enrichment
Enrichment analysis is a statistical way to ask whether a set of genes contains more or fewer genes with a particular annotation than expected from a background set. In biology, the annotation is often a Gene Ontology, or GO, term for a process, function, or cellular location.
What it is and what is measured
Enrichment analysis compares two counts: how many genes in a selected set have a given annotation, and how many annotated genes are present in the full set of genes measured in the experiment. The selected set may be genes that changed strongly or were otherwise flagged as interesting. The output is a p-value, which estimates how likely it is to see that many annotated genes by chance given the background. In this setting, the background is the reference set of all measured genes, not all genes in biology. [0,2]
[1] Chapter 9: Analyses Using Disease Ontologies [3] Bioinformatics enrichment tools: paths toward the ... - PMC - NIH
How it works
The usual idea is that if a biological process is truly associated with the experiment, genes linked to that process should appear more often in the interesting set than in the background. Tools test each term separately using methods such as Fisher exact, chi-square, hypergeometric, or binomial calculations. Some methods also use gene-set scores, such as fold change, instead of a simple selected-versus-not-selected list. Other approaches use ontology structure to relate parent and child terms or to combine network information. [0,2]
[1] Chapter 9: Analyses Using Disease Ontologies [3] Bioinformatics enrichment tools: paths toward the ... - PMC - NIH
How to read the result
A term is often called enriched when its p-value is below a chosen cutoff such as 0.05, but that does not by itself prove a mechanism. The result is evidence of statistical overrepresentation, not a complete biological explanation. Enrichment can be useful for summarizing large gene lists, generating hypotheses, or comparing broad patterns, but the interpretation should stay close to the data and the biological context. [0,1]
[1] Chapter 9: Analyses Using Disease Ontologies [2] Problems With GO Enrichment as an Endpoint for Ecological ...
Limits and common sources of error
Several limits can distort the result. Some genes have no annotation, so the analysis may miss relevant biology. Annotation coverage can be uneven, which creates bias. Testing many terms increases false positives, so multiple-testing correction is needed. Hierarchical ontologies also create correlated tests because parent and child terms share genes, and different terms may sit at similar depths but still represent very different levels of detail. [0,1,2]
[1] Chapter 9: Analyses Using Disease Ontologies [2] Problems With GO Enrichment as an Endpoint for Ecological ... [3] Bioinformatics enrichment tools: paths toward the ... - PMC - NIH
Why this matters
Enrichment is best treated as a summarizing tool that organizes a gene list into named biological categories while keeping the statistical uncertainty visible. It can point to processes worth checking next, but it cannot by itself establish causality, environmental specificity, or a full mechanism. Those questions need additional experiments, better annotation, or other contextual evidence. [0,1]
[1] Chapter 9: Analyses Using Disease Ontologies [2] Problems With GO Enrichment as an Endpoint for Ecological ...
SEE THE IDEA
Basic enrichment workflow
Genes from an experiment are separated into a set of interest and a background set.
Each gene is linked, when known, to GO terms for process, function, or location.
The method compares observed annotation counts with counts expected from the background.
A low p-value suggests overrepresentation, but not a causal mechanism.
The gene list is compared with annotations, the test measures overrepresentation, and the result is interpreted cautiously.
Enrichment links a gene set to ontology terms through statistics, then asks what the pattern may summarize rather than prove.
TRY IT YOURSELF
Test your understanding
Choose an answer, then check the explanation.
PUT IT TOGETHER
Match each concept to its meaning
Drag a label onto its description, or select a label and then select a description. This also works with a keyboard.
4 pairs to match
Sources and further reading
This general lesson was generated from public source material. It explains a concept; it does not interpret a private paper or provide medical advice.