This dataset contains all model weights and corresponding datasets generated by Betti et al. in the manuscript Genetically regulated enhancer RNA expression predicts enhancer-promoter contact frequency and reveals genetic mechanisms at complex trait-associated loci. The following are the contents of the sub-directories in this dataset: coloc: Colocalization results for genome-wide significant (p < 5 x 10-8) GWAS associations in the UK Biobank with eRNA and canonical gene eQTLs (Supplementary Tables 14 and 15). contact_model_training: Input datasets from whole blood and brain, respectively, that were used to train the neural network-based models of contact frequency. eqtl_mapping: eQTLs mapped across 49 cell and tissue types for both eRNAs and canonical genes. scz_mr: Inputs and results for Mendelian randomization analysis of eRNA and canonical gene-based TWAS of schizophrenia. scz_twas: eRNA and canonical gene-based TWAS results of schizophrenia. trained_models: Model weights and SNP covariance matrices for genetically regulated eRNA expression (GReX) across 49 cell and tissue types. uk_biobank_twas: eRNA-based TWAS summary statistics for 4,671 UK Biobank traits across 49 cell and tissue types. Please cite: Betti, M.J., Aldrich, M.C., Lin, P., & Gamazon, E.R. (2024). Genetically regulated enhancer RNA expression predicts enhancer-promoter contact frequency and reveals genetic mechanisms at complex trait-associated loci. Preprint. Betti, M.J., Aldrich, M.C., Lin, P., & Gamazon, E.R. (2024). eRNA GReX (Version 2.0). Zenodo. 10.5281/zenodo.14027849