The article introduces a gene-based association method called PrediXcan, which aims to identify the molecular mechanisms underlying genetic variation and phenotype associations. PrediXcan estimates the genetically regulated component of gene expression using whole-genome tissue-dependent prediction models trained with reference transcriptome datasets. The method correlates this "imputed" gene expression with the phenotype to identify genes involved in the etiology of the phenotype. The authors demonstrate that PrediXcan can detect known and novel genes associated with disease traits and provide insights into the mechanisms of these associations. The method is particularly focused on gene expression regulation, which has been established as a contributor to common diseases. PrediXcan offers advantages such as reduced multiple testing burden and a principled approach to follow-up experiments. The authors apply PrediXcan to the Wellcome Trust Case Control Consortium (WTCCC) data, identifying significant associations with seven complex disease phenotypes, including novel genome-wide significant genes. The method provides biological insights into the mechanisms of disease susceptibility and drug response, and can be applied to existing GWAS datasets through the use of publicly available prediction models.The article introduces a gene-based association method called PrediXcan, which aims to identify the molecular mechanisms underlying genetic variation and phenotype associations. PrediXcan estimates the genetically regulated component of gene expression using whole-genome tissue-dependent prediction models trained with reference transcriptome datasets. The method correlates this "imputed" gene expression with the phenotype to identify genes involved in the etiology of the phenotype. The authors demonstrate that PrediXcan can detect known and novel genes associated with disease traits and provide insights into the mechanisms of these associations. The method is particularly focused on gene expression regulation, which has been established as a contributor to common diseases. PrediXcan offers advantages such as reduced multiple testing burden and a principled approach to follow-up experiments. The authors apply PrediXcan to the Wellcome Trust Case Control Consortium (WTCCC) data, identifying significant associations with seven complex disease phenotypes, including novel genome-wide significant genes. The method provides biological insights into the mechanisms of disease susceptibility and drug response, and can be applied to existing GWAS datasets through the use of publicly available prediction models.