2024-03-06

Presentation by Chirag Lakhani

Localizing AD polygenic signal using deep learning

GWAS of AD shows it is a polygenic disease. Functional annotations aid in finding molecular features that best explain a polygenic trait.

Previously, people used DL delta scores to predict the functional effect of variants.

More than 50% of AD heritability is captured by SNPs which have cis-regulatory regions in microglia. CRE elements \(\rightarrow\) enhancers, promoters, open chromatin regions. There is 10x enrichment of average heritability explained by these variants with respect to the global average.

DL delta scores from Enformer can be used as a prior probability in finemapping.