General update from UKB Meeting for scientific community, 26 June 2014, London Access: >700 researchers, >100 applications, >50 approved, >40 have data Web-based questionnaires 1 st repeat of baseline assessment in 20,000 – completed 2012/13 Wrist-worn accelerometers in 100,000 – well underway Assays on samples from all participants – to complete mid 2015 Multimodal imaging in 100,000 - brain, cardiac and abdo MRI, whole body DEXA, & 3D carotid ultrasound – piloting in 2014 Genome wide genotyping of all 500,000 – to complete mid 2015 MRC UK Dementia Platform Proposal for prolonged cardiac rhythm monitoring in 100,000

Strategy for follow-up Key cohort-wide sources for data linkage: –Death registrations –Cancer registrations –Hospital episode data –Primary care data Questionnaires (web +/- mail) Other linkages sought or under discussion: –mental health minimum dataset; screening databases; disease registers/audits; imaging; histopathology reports/specimens; NHS dental records Future possible linkages: –occupational records; DWP ; HMRC; private medical records… Close liaison with wide range of data providers Working in collaboration with Farr Institute

UK-wide linked health care data

Expert-led adjudication of outcomes Outcomes working group Developing methods for ascertainment, confirmation, and sub-classification of disease outcomes: Cancer Diabetes Cardiac outcomes Stroke Mental health outcomes Ocular outcomes Neurodegenerative outcomes Respiratory outcomes Musculoskeletal outcomes Pilots progressing well; preparing for scaled up algorithms and web adjudication Pilots commencing Pilots being developed

Cardiovascular: Cholesterol Direct LDL-Cholesterol HDL-Cholesterol Triglyceride Apolipoprotein A Apolipoprotein B C-reactive Protein Fibrinogen D-dimer [NT-proBNP] Lipoprotein A Bone and joint: Vitamin D Alkaline Phosphatase Calcium Rheumatoid factor Cancer: IGF-1 Testosterone Oestradiol Sex hormone binding globulin (SHBG) Diabetes: HbA1c Glucose Renal: Enzymatic creatinine Cystatin C Urea Phosphate Total protein Sodium (Urine) Potassium (Urine) Creatinine (Urine) Albumin (Urine) Liver: Albumin Alanine aminotransferase Aspartate aminotransferase Direct Bilirubin Total Bilirubin Gamma Glutamyltransferase Others: Urate Current UKB assay list

Genotyping Chip Design Expert group (chair P Donnelly) considered issues of chip design, in a platform-agnostic way, before procurement. Specific sets of SNPs chosen for inclusion, with further advice from experts in particular areas. Affymetrix used their imputation-aware algorithms to choose additional SNPs to provide good coverage of the genome. For common variants, accuracy of imputed genotypes typically > 95%. Imputation accuracy within UK Biobank will be higher than with current approaches.

18.8K eQTLs Markers : 23K Variants, from HGMD and other databases  80K non-synonymous SNPs and InDels  33K LoF variants Chosen based on MAF in UK UK Biobank Array Content Summary Axiom UK BioBank Array 821K Axiom UK BioBank Array 821K Exome Content Putatively Pathogenic Variants Imputation GWAS grid ApoE ADME Content Miscellaneous eQTLs Neanderthal ancestors 346K CEU-based Grid. MAF>5% in EUR. Impute v2-based Selection 2K variants : Pharmaadme.org PharmaGKB.org 40 K variants: Neurodegenerative disease Cancer Metabolic Phenotypes Blood Phenotypes Lung Function 275K EUR Grid Booster. 1% < MAF< 5% in EUR. Impute v2-based Selection 8K HLA/KIR markers; 8.2K NHGRI GWAS catalogue markers; 1K Y chromosome markers 200 mtDNA markers. ~20K SNPs tagging haplotypes inherited from Neanderthal ancestors ~20K SNPs tagging haplotypes inherited from Neanderthal ancestors