Incorporating Deep Learning Methods to Complex Trait Genetics

Genetic prediction of molecular traits help bridge the gap between genotype and phenotype especially by integrating them with GWAS studies. These prediction models are trained using reference datasets where hundreds of individuals with genotype and omics data are collected. Emerging methods in deep learning using state of the art architectures such as transformers promise to improve prediction approaches and reduced the necessary sample sizes. I will discuss current opportunities and challenges of deep learning approaches to inform the biology of complex diseases.