Identification of shared blood transcriptomic features and candidate sedative targets in chronic insomnia with comorbid anxiety: an in-silico study

Abstract Background Chronic insomnia and anxiety co-occur frequently and reciprocally worsen each other's clinical trajectory, yet the molecular features shared by the two conditions remain poorly defined. We aimed to extract a peripheral blood transcriptomic signature linking the two disorders and to predict sedative/anxiolytic intervention targets. Methods PBMC profiles from GSE208668 (chronic insomnia, 17 vs 25) and whole-blood profiles from GSE61672 (generalized anxiety disorder, 157 vs 179) were analysed with limma. Because no anxiety gene survived Benjamini–Hochberg correction, a directional concordance screen combining nominal anxiety signals with insomnia significance produced 80 comorbidity candidate genes. Immune landscapes were inferred by CIBERSORT and ssGSEA. Hub genes were selected by LASSO, SVM-RFE and random forest; a logistic regression signature was trained in GSE61672. Sensitivity analyses probed screening thresholds and missing sex information. A curated 49-compound sedative/anxiolytic library was used for target overlap, Connectivity-Map-style connectivity scoring and AutoDock Vina docking of 14 hub proteins, with inert-ligand negative controls. Two independent sleep-loss cohorts (GSE39445, GSE56931) served as external validation. Results Insomnia PBMCs showed expansion of resting memory CD4 T cells and contraction of γδ T, Th1 and activated dendritic cell fractions. The pink WGCNA module was strongly enriched with the comorbidity genes (P = 5.6 × 10⁻⁹), and insomnia-cohort WGCNA independently recovered a trait-associated module (r = 0.73). Three machine-learning methods converged on 13 hub genes (training AUC 0.764); the cross-cohort AUC of 1.000 is uninterpretable because of circularity. The gene set was stable across threshold and sex-covariate sensitivity analyses. C3AR1 was the only hub gene overlapping a known drug target (P = 0.0024); docking predicted favourable C3AR1 binding for zolpidem (− 8.53 kcal/mol), zaleplon (− 8.21) and midazolam (− 8.10), beyond the inert-ligand baseline. Directional consistency in acute sleep-loss cohorts was limited (48.7% and 46.8%), with GRINA and TRIM58 significant in both. Conclusion A candidate immune-oriented comorbidity signature links chronic insomnia and generalized anxiety, with C3AR1-centred complement signalling a candidate bridge to sedative pharmacology. The findings are exploratory, hypothesis-generating and entirely computational, and require validation in tissue-uniform, age-matched chronic-insomnia cohorts.