This README is a description of the files available to download as part of the Neale Lab GWAS of UK Biobank phenotypes. For a description of the project and details of the analysis, please see http://www.nealelab.is/uk-biobank. To download GWAS results, see the links in the manifest tab below. At the top of each column in the manifest is a triangle. Click the triangle and search options become available for that column. Once you've found the code you are looking for, refer to the "wget command" column for the corresponding wget command to download the relevant results file. The code used to generate the files described here is publicly available: https://github.com/Nealelab/UK_Biobank_GWAS. Questions or concerns not addressed by this README, the project website, or the Github repository can be directed to nealelab.ukb@gmail.com. variants.tsv.bgz ================ This file contains annotations on each variant in the GWAS, calculated across the analysis subset of 361,194 samples. NOTE: The order of variants in this file matches the order of variants in the results files described below. To join these annotations with a results file, either match on the "variant" field or simply paste the columns together (e.g. "paste variants.tsv K50.gwas.imputed_v3.both_sexes.tsv"). variant string Variant identifier in the form "chr:pos:ref:alt", where "ref" is aligned to the forward strand of GRCh37 and "alt" is the effect allele (use this to join with results files). chr string Chromosome of the variant. pos int Position of the variant in GRCh37 coordinates. ref string Reference allele on the forward strand. alt string Alternate allele (not necessarily minor allele). rsid string rsid (not guaranteed to be unique). varid string Unique variant identifier included in imputed BGEN files. consequence string Consequence annotated using VEP version 85. consequence_category string Category of VEP-annotated consequence ("ptv", "missense", "synonymous", "non_coding"). info float Imputation INFO score as provided by UK Biobank. call_rate float Call rate (calculated using hardcall genotypes). AC int Allele count (calculated using hardcall genotypes). AF float Allele frequency (calculated using hardcall genotypes). minor_allele string Minor allele (equal to ref allele when AF > 0.5, otherwise equal to alt allele). minor_AF float Minor allele frequency (calculated using hardcall genotypes). p_hwe float Hardy-Weinberg p-value. n_called int Number of samples with defined genotype at this variant. n_not_called int Number of samples without a defined genotype at this variant. n_hom_ref int Number of samples with homozygous reference genotype at this variant. n_het int Number of samples with heterozygous genotype at this variant. n_hom_var int Number of samples with homozygous alternate genotype at this variant. n_non_ref int Number of samples with non-homozygous reference genotype at this variant (n_het + n_hom_var) r_heterozygosity float Proportion of samples with heterozygous genotype at this variant. r_het_hom_var float Ratio of samples with heterozygous genotype to samples with homozygous alternate genotype at this variant. r_expected_het_frequency float Expected r_heterozygosity based on Hardy-Weinberg equilibrium. samples.{both_sexes,female,male}.tsv.bgz ======================================== These files contain the plate name and well of each sample included in our GWAS. You can use these to subset the generic sample QC file provided by UK Biobank, and match the samples to the IDs specific to your application. plate_name string Plate on which the sample was processed. well string Well in which the sample was processed. european_samples.tsv.bgz ======================== This file contains the plate name and well of each sample remaining in the analysis after removing ancestry outlier samples based on a custom principal components analysis. plate_name string Plate on which the sample was processed. well string Well in which in the sample was processed. phenotypes.{both_sexes,female,male}.tsv.bgz =========================================== These files contain a description and summary of each phenotype included in the analysis. phenotype string Unique phenotype identifier. Format differs depending on the source of the phenotype. description string Free text description of the phenotype. variable_type string {"categorical", "ordinal", "continuous_irnt", "continuous_raw"} Variable type. Each continuous variable has two versions: an untransformed version ("continuous_raw") and a version where values have been inverse rank normalized ("continuous_irnt"). source string {"icd10", "finngen", "phesant"} Source of the phenotype. See notes below. n_non_missing int Number of samples within the analysis subset defined for this phenotype. n_missing int Number of samples within the analysis subset not defined for this phenotype. n_controls int For case/control phenotypes, number of control samples within the analysis subset. n_cases int For case/control phenotypes, number of case samples within the analysis subset. PHESANT_transformation string This field describes the transformations performed by PHESANT for the applicable phenotypes. notes string Any additional notes. Analysis subset sizes: - both_sexes: 361,194 samples - female: 194,174 samples - male: 167,020 samples Phenotype sources: - icd10: These phenotypes were generated from UK Biobank fields 41202-0.0 - 41202-0.379. For each sample, the set of ICD10 codes (truncated to the first three characters, e.g. "K50") included in these fields was collected. The ICD10 phenotypes are booleans indicating whether the ICD10 code is included in that set of codes for each sample. - finngen: These phenotypes were manually curated by collaborators in the FinnGen research project. Many are combinations of different ICD10 codes. - phesant: These phenotypes were automatically processed using a modified version of the software PHESANT (https://www.ncbi.nlm.nih.gov/pubmed/29040602). .gwas.imputed_v3.{both_sexes,female,male}.tsv.bgz ================================================================= These are the GWAS results files (e.g., "K50.gwas.imputed_v3.both_sexes.tsv.bgz"). variant string Variant identifier in the form "chr:pos:ref:alt", where "ref" is aligned to the forward strand of GRCh37 and "alt" is the effect allele (use this to join with variant annotation file). minor_allele string The minor allele (alt allele is not always minor). minor_AF float Frequency of the minor allele in the n_complete_samples defined for this phenotype. expected_case_minor_AC float (Optional) For case/control phenotypes, calculated as (2 * minor_AF * n_cases). expected_min_category_minor_AC float (Optional) For categorical phenotypes with less than 5 categories, calculated as (2 * minor_AF * number of samples in smallest category). low_confidence_variant boolean Flag indicating low confidence results based on the following heuristics: Case/control phenotypes: expected_case_minor_AC < 25 or minor_AF < 0.001. Categorical phenotypes with less than 5 categories: expected_min_category_minor_AC < 25 or minor_AF < 0.001 Quantitative phenotypes: minor_AF < 0.001. n_complete_samples int Number of samples defined for this phenotype. AC float Allele count of alt allele calculated on dosages within n_complete_samples. ytx float Dot product of phenotype vector y and genotype vector x (alt allele count in cases for case/control phenotypes). beta float Estimated effect size of alt allele. se float Estimated standard error of beta. tstat float t-statistic of beta estimate (= beta/se). pval float p-value of beta significance test.