Enrichment or depletion of a GO category within a class of genes: which test?

By I. Rivals, L. Personnaz, Lieng Taing, Marie-Claude Potier. 2007

253 citations recorded by the source corpus.

MOTIVATION A number of available program packages determine the significant enrichments and/or depletions of GO categories among a class of genes of interest. Whereas a correct formulation of the problem leads to a single exact null distribution, these GO tools use a large variety of statistical tests whose denominations often do not clarify the underlying P-value computations. SUMMARY We review the different formulations of the problem and the tests they lead to: the binomial, chi2, equality of two probabilities, Fisher's exact and hypergeometric tests. We clarify the relationships existing between these tests, in particular the equivalence between the hypergeometric test and Fisher's exact test. We recall that the other tests are valid only for large samples, the test of equality of two probabilities and the chi2-test being equivalent. We discuss the appropriateness of one- and two-sided P-values, as well as some discreteness and conservatism issues. SUPPLEMENTARY INFORMATION Supplementary data are available at Bioinformatics online.

Subfields: Artificial Intelligence, Genetics, Geometry and Topology

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Genetics

In Geometry and Topology

Citations in the corpus

Enriching the gene set analysis of genome-wide data by incorporating directionality of gene expression and combining statistical hypotheses and methods (2013, methodology): “These methods are popular among GO-based tools (6), including BiNGO (7), DAVID (8), GOstat (9) and GOEAST (10).”
DIANA miRPath v.2.0: investigating the combinatorial effect of microRNAs in pathways (2012, methodology): “The default algorithm utilized is the one-tailed Fisher’s exact test, which is often performed in such context (22,23).”
STRING v9.1: protein-protein interaction networks, with increased coverage and integration (2012, methodology): “The terms are sorted by their enrichment P-value, which we compute using a Hypergeometric test, as explained in (53).”
Mut2Vec: distributed representation of cancerous mutations (2018, methodology): “We estimated the statistical significance based on the hypergeometric test [27] on entire driver gene mutations in the IntOGen database.”
GOEAST: a web-based software toolkit for Gene Ontology enrichment analysis (2008, methodology): “Since the Fisher’s exact test and the hypergeometric test are statistical equivalent (29), GOEAST uses hypergeometric test by default because it is much faster to be calculated than Fisher’s exact test by R packages.”
Efficient Test and Visualization of Multi-Set Intersections (2015, methodology): “FET has been widely used in evaluating the enrichment of known functional pathways in predicted gene signatures 7 .”
Gene-set approach for expression pattern analysis (2008, methodology): “[2] reviewed in detail the methods and tools for IGA.”
Enrichment Map: A Network-Based Method for Gene-Set Enrichment Visualization and Interpretation (2010, methodology): “The one-tail Fisher’s Exact Test [17], based on the hypergeometric distribution , was the first method proposed to address this problem [18], and continues to be one of the most used testing methods of this type [19].”
Meta-analysis identifies 29 additional ulcerative colitis risk loci, increasing the number of confirmed associations to 47 (2011, methodology): “To assess the statistical enrichment of functional gene sets from molecular function, biological process categories and pathways for candidate genes, P values were computed using the hypergeometric tes”
Mining breast cancer genes with a network based noise-tolerant approach (2013, methodology): “The p-value is calculated by randomly picking sets of genes from the genome and computing the probability of obtaining more genes with GO terms annotated to those in the study set [52].”
REVIGO Summarizes and Visualizes Long Lists of Gene Ontology Terms (2011, background): “In order to interpret the results of such experiments, statistical testing for over-and under-representation of gene functional categories is used [1].”
Comprehensive comparison of the interaction of the E2 master regulator with its cognate target DNA sites in 73 human papillomavirus types by sequence statistics (2007, methodology): “We use the hypergeometric probability distribution to calculate the probability of picking 10 or more high-risk types by choosing 43 types at random (75).”
Bioinformatics enrichment tools: paths toward the comprehensive functional analysis of large gene lists (2008, unknown): “number of genes in the list that hit a given biology class as compared to pure random chance, can be performed with the aid of some common and well-known statistical methods (11,12,76), including Chi-square, Fisher’s exact test, Binomial probability and Hypergeometric distribution, etc.”
Archetypal analysis of diverse Pseudomonas aeruginosa transcriptomes reveals adaptation in cystic fibrosis airways (2013, methodology): “GO classes that were over-represented within the group of up-or down-regulated genes were identified by Hypergeometric distribution test [26].”
The STRING database in 2017: quality-controlled protein–protein association networks, made broadly accessible (2016, methodology): “The enrichment tests are done for a variety of classification systems (Gene Ontology, KEGG, Pfam and InterPro), and employ a Fisher’s exact test followed by a correction for multiple testing (42,43).”