Abstract Introduction Sleep and circadian disturbances are highly prevalent in autism spectrum disorder (ASD), yet the mechanisms linking ASD risk genes to human sleep/circadian physiology remain unclear. Advances in single-cell expression quantitative trait loci (eQTL) datasets now enable testing of gene and cell-type-specific relationships with sleep traits. Methods We intersected the SPARK/ClinGen list of 162 ASD risk genes with cell-type–specific cis-eQTLs (N=983), yielding 111 genes with suitable instruments across seven brain cell types and 24 neuronal and glial sub-classes. These cis-eQTLs were used as expression proxies in a drug-target Mendelian randomization (DTMR) framework to estimate causal relationships with UK Biobank sleep traits (chronotype, sleep duration, ease of getting up, snoring; N≤413,343) under multiple-testing correction. To further prioritize ASD sleep–circadian genes, we triangulated DTMR findings with (i) postmortem human dorsolateral prefrontal cortex (dlPFC) time-series transcriptomics, and (ii) Drosophila RNAi circadian phenotyping. Results Across sleep outcomes, 12 ASD risk genes exhibited Bonferroni-significant, cell-type–specific relationships, primarily for chronotype (11 genes) and ease of getting up (5 genes). BCL11A showed concordant and opposite-direction chronotype associations across excitatory neurons, interneurons, and oligodendrocyte progenitors. RELN demonstrated robust chronotype associations across five neuronal and glial subclasses, strongest in interneurons. MEF2C displayed large chronotype relationships in homeostatic microglia and opposite-direction getting-up effects in activated microglia. Several Bonferroni- or FDR-significant genes (ACTB, EHMT1, DNMT3A, SETBP1, SHANK2, MAGEL2, IRF2BPL, CACNA1C, MEIS2) also displayed rhythmic or clock-regulated expression in dlPFC cell types, supporting conserved circadian roles. Further, Drosophila homologs of ASD genes such as NBEA, PHF3, PSMD12, and RERE produced canonical circadian phenotypes with complementary human MR associations. Although no ASD gene was implicated across all three systems, several showed convergent evidence across human genetics, brain circadian regulation, and Drosophila screening. Conclusion Integrating human genetic, transcriptomic, and cross-species functional evidence reveals that ASD risk genes exert cell-type–specific impact on human sleep and circadian traits, especially chronotype, highlighting conserved molecular pathways linking ASD genetics to sleep–circadian biology. Support (if any) SFARI, Eagles Autism Foundation, NINDS R35NS132223 (RA, GW, BL)
Rosoff et al. (2026) studied this question.