Population-scale, disease-context aware meta single-nucleus eQTLs mapping (N = 1,197) in nine cell types from cortex and substantia nigra identifies 125 risk genes and cell types for PD GWAS signals.
CREsted, a sequence-based deep learning model, analyzes genomic regulatory code, decodes enhancer grammar, and designs synthetic enhancers. It preprocesses single-cell data, models chromatin accessibility, and compares cell states across tissues.
