Transcriptome-wide association studies (TWAS) have been successful in identifying disease susceptibility genes by integrating cis-variants predicted gene expression with genome-wide association studies (GWAS) data, while trans-located variants for predicting gene expression remain largely unexplored. Here, we introduce transTF-TWAS, which incorporates transcription factor (TF)-linked trans-located variants to enhance model building. Using data from the Genotype-Tissue Expression project and extensive disease GWAS datasets, we demonstrate that transTF-TWAS approach significantly contributes to the discovery of disease risk genes.