Enrichment Disequilibrium: a novel approach for measuring the degree of enrichment after gene enrichment test

Abstract

Motivation: Commonly used gene enrichment analysis methods, such as Hypergeometric distribution, play an important role in the functional analysis of interesting gene lists. But the statistical significance obtained by these methods only represents the probability of error that is involved in accepting enrichment, and is not suitable to evaluate the degree of enrichment. Although there have been some methods to measure the enrichment degrees, such as relative enrichment factor, new methods are still needed to meet the requirements for comparing the degree of enrichment.

Results: We developed a novel method, Enrichment Disequilibrium (ED), to measure the degree of enrichment. Enrichment equilibrium means that the interesting gene set and the known functional gene set (such as a KEGG pathway) are independent (i.e. random association). ED is defined as the degree of non-independence. Compared with the relative enrichment factor, ED has a clearer biological meaning, is a standardized indicator, and has a symmetrical interval (range from -1 to +1). It is more suitable to measure the enrichment degree. For an interesting gene set, researchers can obtain some significant functional gene sets by traditional enrichment test. Then using ED, they can compare the degree of enrichment among these significant gene sets, and prioritize them.