Artificial intelligence and multi-omics nominate TAZ as an insomnia-related diagnostic and druggable target for...

Insomnia is one of the most common non-motor comorbidities of Parkinson's disease (PD) and often before the onset of motor symptoms. Identifying the molecular mechanisms of insomnia may facilitate the early diagnosis of PD and contribute to therapeutic development.

Five human PD substantia nigra (SN) bulk-seq datasets (GSE20141, GSE7621, GSE20164, GSE20163, and GSE20333), with an insomnia-related gene list, were acquired from GEO and Genecard databases. First, the integration of GSE20141 and GSE7621 was analyzed to identify insomnia-related DEGs using limma and the WGCNA framework. GSE20164 and GSE20163 combination were used as a training set for insomnia-related hub gene recognition. Furthermore, the aforementioned four datasets, along with an independent validation set (GSE20333), were cross-validated for insomnia-related diagnostic model construction. The human PD-SN single-cell profile (GSE140231) was utilized for exploring the mechanisms underlying the heterogeneity of insomnia-related hub genes in spatial and temporal contexts. Furthermore, a cutting-edge artificial intelligence (AI)-driven framework (DrugRefLector) and molecular docking techniques was used to identify an optimal agent for the treatment of PD based on the GSE20164 and GSE20163 integrated dataset. Finally, an in vitro q-RT-PCR experiment was conducted to estimate the targeted gene expression.

TAZ (WWTR1) is associated with the increased expression of insomnia-related diagnostic markers linked to PD pathogenesis, mainly in neurons, and has excellent predictive performance for PD diagnosis. Furthermore, BRD-K97481123 can be considered as a potential therapeutic agent for the treatment of PD by targeting TAZ.

By integrating AI pipelines and multi-omics, our study first traced TAZ mechanisms in PD pathogenesis and elaborated on TAZ's predictive and druggable potential for PD patients.