# Multicondition expression profiling reveals limitations of canonical housekeeping genes

## TL;DR

This paper introduces a new method to evaluate housekeeping genes, showing that many commonly used ones are affected by drugs, which challenges their reliability as reference genes.

## Contribution

A novel algorithm scores genes based on expression ubiquity, revealing limitations of traditional housekeeping genes under drug treatments.

## Key findings

- Many traditional housekeeping genes are strongly influenced by commonly used drugs.
- Some genes maintain strong invariance under chemical stimuli, making them better reference genes.
- An interactive web interface helps researchers select robust reference genes and explore gene rankings.

## Abstract

Housekeeping genes are commonly used as reference genes for analyzing expression data. We introduce an algorithm that scores and ranks genes based on their expression ubiquity using three criteria: the proportion of samples with high expression, overall expression levels, and expression invariance. The resulting continuous and configurable measure of gene ubiquity overcomes the limitations of current categorical classifications.

Using this measure, we systematically evaluate how genes traditionally assumed to be ubiquitously expressed (so-called stable housekeeping genes) respond to drugs. By integrating expression data from the Genotype-Tissue Expression portal (physiological conditions) and the Connectivity Map (drug effects), we identify subsets of genes that maintain strong invariance under chemical stimuli, rendering them particularly suitable as reference genes. At the same time, we show that many seemingly ubiquitous genes are strongly influenced by commonly used drugs. Our findings open up new conceptual perspectives on drug mechanisms and cellular regulatory responses and they call into question the reliability of several classic housekeeping genes.

To facilitate broader application of these results, we present an interactive web interface (https://ubigen.uni-rostock.de) that allows researchers to explore gene rankings, adjust scoring criteria and analyze specific gene sets. This platform facilitates the selection of robust, ubiquitously expressed reference genes for experimental applications and deepens our understanding of transcriptional regulation under physiological and chemically perturbed conditions.

The online version contains supplementary material available at 10.1186/s12864-026-12563-8.