Polygenic and spatial insights into the genetic uniqueness of essential tremor using common variants.
Using spatial transcriptomic mapping and PRS conditioning, the study shows essential tremor shares substantial common-variant genetic signals with Parkinson's disease and cognition, and that removing these shared signals reduces PRS-based classification performance, implying common variants alone…
What the AI sees
Using spatial transcriptomic mapping and PRS conditioning, the study shows essential tremor shares substantial common-variant genetic signals with Parkinson's disease and cognition, and that removing these shared signals reduces PRS-based classification performance, implying common variants alone…
Research significance
Although it provides limited direct therapeutic targets for Parkinson's, the paper highlights shared genetic architecture with PD and argues that rare variants, nonstandard variant types, and spatially resolved expression data may be more informative for discovering disease-specific mechanisms…
Source abstract
INTRO: The heterogeneity of Essential Tremor (ET) complicates how it is diagnosed and studied. ET is often misdiagnosed as other neurological disorders and vice versa. We aimed to better understand ET by comparing brain maps informed by ET common variants to those derived from related phenotypes. A further goal was to enhance the diagnostic precision of ET by accounting for shared genetic signals between ET and genetically overlapping phenotypes. METHODS: Phenotype variant association mapping to the brain was done for ET, Parkinson's disease (PD), dystonia, and cognition across adult mouse whole brain and cerebellum spatial transcriptomic data through gsMap. Separately, ET genome-wide association study (GWAS) summary statistics were conditioned on PD and cognition to account for shared genetic signals. Using both raw and conditioned GWASes, ET polygenic risk scores (PRS) were calculated across patient cohorts and controls, and their respective ability to classify ET at the 90th percentile was compared using McNemar's test. RESULTS: Spatial mappings of GWAS signals revealed many shared associations between phenotypes. The raw ET PRS model preformed 1.33% (95% CI: [0.30% - 2.35%]; p = 0.0129) better than the conditioned ET PRS model. CONCLUSION: We lack the ability to decern ET from related phenotypes using existing common variant disease associations. Efforts to isolate a core genetic signal for ET by de-noising shared associations reduced the accuracy of patient classification. Rare variants and lesser explored variant types may be key to unlocking informative variants for ET.