Rare coding variants reveal 260 genes linked to sleep — including 24 never before reported
The largest rare-variant analysis of sleep genetics to date has identified 260 genes associated with human sleep phenotypes, including 29 genes linked specifically to sleep medication use — 24 of which have no prior association with any sleep trait. The findings, posted as a medRxiv preprint, suggest that different ways of measuring sleep capture biologically distinct genetic mechanisms.
The study, led by Yingzhe Zhang and colleagues from Harvard Medical School, Massachusetts General Hospital, and the Broad Institute, analyzed whole-exome sequences from 397,065 UK Biobank participants and whole-genome sequences from 171,536 All of Us participants of diverse ancestries, with replication in 31,275 individuals from the Mass General Brigham Biobank.
What they found
Across 36 sleep phenotypes spanning self-report, clinical diagnoses, medication use, and accelerometry, the team mapped:
- 260 gene associations with sleep traits
- 29 genes associated with sleep medication use
- 24 of those 29 are novel — never previously reported with any sleep phenotype
The study found modest but statistically significant heritability for rare coding variants and strong genetic correlations between sleep medication use, insomnia, and fatigue, suggesting shared biological pathways underlying these related complaints.
A temporal gene expression analysis added a striking layer: genes associated with self-reported sleep traits (e.g., chronotype, sleep duration, insomnia symptoms) showed constant high expression throughout prenatal development. In contrast, genes linked to sleep medication phenotypes peaked sharply in the late prenatal period, pointing to distinct developmental origins and biological mechanisms.
“This highlights that the measurement source itself — what we ask patients versus what we observe — captures fundamentally different biology,” the authors note.
Why it matters
Most genome-wide association studies (GWAS) of sleep have focused on common variants. This study shifts attention to rare coding variants, which have larger individual effect sizes and are more directly interpretable at the protein level. The 24 newly identified genes represent high-priority targets for therapeutic development, particularly for sleep medication pathways that may be distinct from those driving sleep timing or duration in healthy individuals.
The inclusion of the All of Us cohort, with its substantial non-European ancestry, also addresses a longstanding limitation in sleep genetics: most prior work has been done in populations of European descent. Replication across ancestrally diverse samples strengthens the generalizability of the findings.
Limits
As a preprint, this work has not yet undergone peer review. The UK Biobank cohort is healthier and less socioeconomically diverse than the general population, which may influence the distribution of sleep phenotypes. The analysis also captures coding variants only — non-coding regulatory variants, which may play important roles in sleep regulation, were not assessed.
Bottom line
Rare coding variant analysis across nearly 600,000 participants reveals hundreds of genes linked to sleep, including two dozen entirely novel associations tied to sleep medication use. The findings underscore that how we measure sleep may determine which biology we discover, and open new avenues for sleep-targeted therapeutics informed by rare variation.
Source: Zhang Y, Lu W, Kunorozva L, et al. Rare Coding Variants Reveal Distinct Genetic Architectures Across Multidimensional Sleep Phenotypes. medRxiv Preprint]. 2026 Jun 18:2026.06.16.26355625. DOI: [10.64898/2026.06.16.26355625