Cell-type-specific Transcriptomic-wide Association Studies Detected 91 Independent Risk Genes for Alzheimer's Disease Dementia

Abstract

Existing TWASs of AD dementia typically use a single statistical method to identify cell-type-specific risk genes. Here we sought to improve on existing approaches and utilized an omnibus xWAS pipeline to integrate snRNA-seq dataset (n=415) of dorsolateral prefrontal cortex (DLPFC) and the latest GWAS data of AD dementia to detect cell-type-specific risk genes. We identified 223 cell-type-specific TWAS risk genes across six major brain cell types, including 91 independent associations of which 11 are novel. Integrating proteomics data (n=716) of DLPFC and GWAS data, we identified 21 significant PWAS risk genes including 13 independent associations, overlapping with 32% independent cell-type-specific TWAS associations. Protein-protein interaction network analyses showed that our TWAS findings are functionally linked to established AD risk genes such as APOE, BIN1, and MAPT. These results underscore the value of leveraging large-scale snRNA-seq and proteomics data to uncover novel cell-type-specific mechanisms underlying AD dementia.