The term “switch-like gene expression” refers to a binary pattern in which a gene is distinctly “on” in some individuals and “off” in others, as opposed to genes whose expression fluctuates over time. While the lac operon in E. coli is an example of a prokaryotic phenomenon, little is known about the prevalence and biological implications of such patterns in humans. Some genes clearly show bimodal expression, with most human genes expected to have graded expression resembling a dimmer switch.
This study highlighted a knowledge gap by examining the transcriptomes, methylomes, and genomes of 943 people in 27 human tissues to identify switch-like genes. A DNA methylation study suggests that the tissue-specific bimodal expression patterns are likely driven by hormone-mediated epigenetic changes. The global switch-like pattern is explained by genetically driven epigenetic silencing. For example, in vaginal tissues, experimental studies observed that low estrogen levels are associated with decreased ALOX12 expression and reduced epithelial thickness.
Bimodal gene expression has been extensively studied in cancer biology, where on and off states are linked to various disease phenotypes and their prognoses. The goal of the genotype-tissue expression (GTEx) project is to create a comprehensive public resource for understanding gene expression and regulation across various tissues. The predictors in the regression model are not correlated by converting a set of intercorrelated confounders into an orthogonal set of principal components (PCs) using principal component analysis (PCA).
A total of 516,267 gene-tissue pairings were examined for bimodal expression by using Hartigan’s dip test for unimodality in various individuals. Multiple hypothesis correction and recalibration were used to account for confounders and false positives. To identify gene-tissue combinations with bimodal expression across individuals, researchers employed a two-round dip test procedure, consisting of one round on raw expression data and another on confounder-corrected expression data. To further investigate the disease association, the R package “ontology Index” was used to analyse Disease Ontology files. In this file, diseases are arranged in a hierarchical pattern from primary to secondary categories.
To determine the genetic basis of bimodality for switch-like genes in cluster 2A (a group of 32 genes exhibiting consistent bimodal expression across all tissues), Ensembl BioMart was utilized to obtain the coordinates of the genes in both hg19 and hg38 builds using Ensembl IDs. Researchers conducted analysis to determine whether cluster-1 vagina-specific switch-like genes are overrepresented among the top 40 genes differentially expressed in vaginal atrophy. Cluster 1, comprising 432 genes, represents genes such as KRT17 that exhibit bimodal expression in a limited subset of tissues, as determined by manual inspection. GPX1P1 is one of 32 genes in Cluster 2A that show bimodal expression across all tissues.
Since hormonal variations between sexes can influence gene expression, the study examined sex-biased and hormone-mediated patterns of gene expression. The expression of ALOX12 in the basal and parabasal layers did not differ significantly between healthy and atrophic vaginal tissues, as determined by immunohistochemical staining.
Gene-environment association studies are more personalized if the gene-environment interactions influence switch-like gene expression patterns. This multi-tissue analysis provides a comprehensive resource on switch-like gene expression, underscoring the interplay between genetic, epigenetic, and environmental factors. It also offers new insights into tissue specialization and disease risk.
Reference: Aqil A, Li Y, Wang Z, et al. Switch-like gene expression modulates disease risk. Nat Commun. 2025;16:5323. doi:10.1038/s41467-025-60513-x
