Every hour of lost sleep leaves a molecular trace in the brain. For years, scientists have tried to catalog those traces by measuring which genes turn on or off when animals are kept awake. But individual studies, limited by small sample sizes and biological noise, have produced conflicting lists. The result is a field rich in data but poor in consensus.
A new meta-analysis published in Neurobiology of Sleep and Circadian Rhythms changes that. By pooling six independent microarray studies into a single statistical framework, an international team has produced the largest transcriptomic meta-analysis of sleep deprivation ever conducted in the mouse brain. The work identifies 498 genes that reliably change expression with sleep loss, 402 of which were not previously linked to sleep at all.
The meta-analysis
The researchers began by scouring the literature for microarray studies that met strict criteria: mice aged 8 to 52 weeks, sleep deprivation lasting 3 to 12 hours, conducted entirely during the light phase using gentle handling or novel object exposure, and housed on a standard 12-hour light/dark cycle with food available at all times. Six studies passed the filter, yielding 23 pairwise comparisons between sleep and sleep-deprived brain tissue samples.
In total, the analysis drew on 173 microarrays representing 245 mice, 123 of which had slept normally and 122 of which had been sleep deprived. The investigators applied a random effects model to account for variation between studies and corrected for multiple testing using a false discovery rate threshold of q < 0.01, a stringent cutoff designed to minimize false positives when scanning thousands of genes simultaneously.
Findings
The meta-analysis returned 498 genes whose expression changed significantly with sleep deprivation. Of these, only 96 had been flagged as sleep-related in the original individual studies, meaning that roughly 80 percent of the genes the analysis identified had gone undetected in prior work. The remaining 402 genes represent novel candidates that may play roles in the brain’s response to sleep loss.
The team did not stop at the mouse genome. Using a method called PEGS, they compared their list of sleep deprivation-responsive genes against human genome-wide association study data for four sleep traits: chronotype, insomnia, daytime sleepiness, and sleep duration. Fourteen of the novel mouse genes also showed associations with at least one of these traits in humans, raising the possibility that the same molecular pathways are at work across species.
Three additional novel genes carried sleep-related phenotypes in knockout mice from the International Mouse Phenotyping Consortium, a large-scale effort to systematically characterize the function of every gene in the mouse genome. Those findings suggest that disrupting these genes alters sleep behavior in ways that parallel the molecular changes seen after acute sleep deprivation.
Functional analysis of the 498-gene set pointed to biological processes including synaptic signaling, protein phosphorylation, and the stress response. The team also flagged three genes with known relevance to human disease: Maoa, linked to neurobehavioral disorders; Naglu, associated with the lysosomal storage disorder mucopolysaccharidosis type IIIB, which often presents with sleep disturbances; and Tipin, involved in DNA replication stress.
Rasd1 validation
Among the most consistently upregulated genes across all six studies was Rasd1, also known as Dexras1, a gene previously implicated in circadian rhythm regulation and glucocorticoid signaling. To confirm the meta-analysis result, the researchers performed quantitative PCR on independent brain tissue samples from sleep-deprived mice. The upregulation held.
They then turned to a knockout mouse line lacking the Rasd1 gene and characterized its sleep using electroencephalography. Compared with wild-type littermates, Rasd1 knockout animals showed significant changes in the total amount of behavioral sleep, its distribution across the light/dark cycle, and the structure of individual sleep bouts. The results suggest that Rasd1 is not merely a marker of sleep loss but a functional player in how the brain regulates sleep under both baseline and deprived conditions. The finding also demonstrates the value of the meta-analysis as a hypothesis-generating tool: the computational screen flagged a gene that subsequent experimentation confirmed as biologically meaningful.
Why it matters
The sleep field has long struggled with reproducibility. Individual transcriptomic studies of sleep deprivation are expensive and technically demanding, and they often yield results that do not replicate across laboratories. The problem is compounded by small sample sizes and differences in mouse strains, deprivation protocols, and tissue dissection methods.
This meta-analysis addresses that fragmentation head-on. By integrating data from multiple laboratories using a disciplined statistical framework, the authors have produced a consensus gene list that is far more robust than anything a single study could provide. The 402 novel genes represent a significant expansion of the known molecular landscape of sleep loss, and the 14 genes with human trait associations offer a direct bridge between mouse biology and human sleep health.
For researchers studying the mechanisms of sleep homeostasis or the health consequences of chronic sleep restriction, the 498-gene set functions as a roadmap. It identifies specific molecular pathways, candidate genes for functional validation, and potential targets for therapeutic intervention. The study is also the largest meta-analysis of its kind, setting a methodological precedent for how the field should handle its growing volumes of transcriptomic data.
Limits
The analysis is restricted to microarray data, which captures only a pre-selected set of transcripts and may miss splice variants, non-coding RNAs, and low-abundance messages that RNA sequencing would detect. The studies included come exclusively from mouse brain tissue, raising questions about how well the findings generalize to humans or to peripheral tissues. The sleep deprivation protocols, while standardized in the inclusion criteria, still varied between studies in duration, exact method, and brain region sampled. Finally, the human genetic associations are correlational and do not establish whether the identified genes cause sleep disturbances or respond to them.
Bottom line
A meta-analysis of 245 mice across six independent studies has identified 498 genes that change expression with sleep deprivation, 402 of which were previously unknown. Fourteen of the novel genes also show links to human sleep traits, and one candidate, Rasd1, was experimentally validated as a functional regulator of sleep. The work provides the field with its largest and most reliable transcriptomic resource to date and demonstrates the power of meta-analysis to extract signal from noisy biological data.
Source: Abdalla OHMH, Dunlop E, Wilson TS, Reardon PK, Iqbal M, Tam SKE, Vyazovskiy VV, Ray DW, Brown LA, Cheng HYM, Peirson SN. A meta-analysis of sleep deprivation transcriptomic studies in mouse brain reveals hundreds of novel candidate sleep genes. Neurobiology of Sleep and Circadian Rhythms 21 (2026): 100149. DOI: 10.1016/j.nbscr.2026.100149. Open access (CC BY).