Cross-population Transcriptome Prediction Models for Transcriptome-Wide Association Studies in Diverse Populations

Genome-wide association studies (GWAS) have been utilized to find associations between single nucleotide polymorphisms (SNPs) and traits. However, most GWAS are conducted on populations of European ancestries. Further analysis involving the transcriptome has to lead to new biological discoveries. Transcriptome-wide association studies (TWAS) have utilized prediction models to estimate transcriptome data and then find gene-trait associations. However, most of these prediction models are based on populations of European ancestries with different genetic structures, making cross-population predictions unreliable. Therefore, this study focuses on building and testing transcriptome prediction models with better cross-population performance.