Colocalisation results from "Ultra-fast genetic colocalisation across millions of traits"

This repository contains the data used in the article "Ultra-fast genetic colocalisation across millions of traits". The datasets support the analysis and code available at: https://github.com/AlasooLab/gpu-coloc-manuscript. For both coloc.bf_bf and gpu-coloc, we used prior probabilities p1 = p2 = 1e-4. Contents: tabix_ftp_paths.tsv.zipMetadata table listing tabix-indexed QTL datasets from the eQTL Catalogue. ensemble.tsv.zipA table mapping each molecular trait ID in the eQTL Catalogue to its corresponding gene. UKBB_and_eQTL_Catalogue_credible_sets.npy.zipFinnGen_and_eQTL_Catalogue_credible_sets.npy.zipNumPy arrays containing credible set IDs (dataset, trait, signal ID) for colocalisation with the eQTL Catalogue, UK Biobank, FinnGen, and Rahu et al. metabolic traits. 56_metabolites_finemapping_credible_sets.tsv.zipFinemapped credible sets for 56 metabolites from Rahu et al. (Zenodo link). ge_ge_results_5e-6_0.8.tsv.zipge_ge_results_1e-6_0.8.tsv.zipge_ge_results_1e-7_0.8.tsv.zipgpu-coloc results for colocalisation within the eQTL Catalogue for datasets where quant_method is ge, microarray, or aptamer. Each file corresponds to a different prior p12. All results shown have PP.H4 ≥ 0.8. ge_ge_clpp.tsv.zipCLPP results for the same eQTL Catalogue datasets (ge, microarray, aptamer). coloc_R_benchmark_results.tsv.zipBenchmark results from coloc.bf_bf and gpu-coloc using p12 = 1e-6. gpu-coloc_benchmark_results.tsv.zipgpu-coloc benchmark output using p12 = 1e-6. finngen_lbf_results_0.8_5e-6.tsv.zipfinngen_lbf_results_0.8_1e-6.tsv.zipfinngen_lbf_results_0.8_1e-7.tsv.zipgpu-coloc results for colocalisation between the eQTL Catalogue and FinnGen (release r12), with varying p12priors. All results have PP.H4 ≥ 0.8. finngen_clpp_results.tsv.zipCLPP results for colocalisation between eQTL Catalogue and FinnGen r12. metabolic_lbf_results_0.8_1e-6.tsv.zipgpu-coloc results for colocalisation between the eQTL Catalogue and Rahu et al. (metabolic traits), using p12 = 1e-6 and PP.H4 ≥ 0.8. met_ge_clpp.tsv.zipCLPP results for colocalisation between the eQTL Catalogue and Rahu et al.