The UK Biobank –The Scientific Challenge

John BellRegius Professor of Medicine, Oxford University

Chairman, UK Biobank Scientific CommitteeChairman, Steering Committee, MOLPAGE

Biobank UK: The Context

• A landmark epidemiology/genetic/genomic prospective cohort for UK science

• Modern, efficient methodology in patient recruitment, statistics, IT, genotyping, LIMS and genomics

• The resource to develop and expand UK expertise in the integrated field of genetic epidemiology

• To identify the major genetic and environmental factors that contribute to common disease pathogenesis

Recruitment500,000 participants

aged 40-69

Data & Sample storage

• Environmental exposure

• Physiological variables

• Neuropsychiatricevaluation

• Biochemical markers

• DNA, plasma, urine

Follow-up• Record linkage• Registries• GP records• Hospital

admissions• Cancer registries• Prescriptions• Death

Biobank Objectives – April 20051. Study genetic and environmental factors in disease.*2. Study genes and environment in a variety of settings

(disease subtype, age, sex, social settings)3. Use prevalence cases to validate or establish genetic

association in a range of disease4. Characterise impact of socio-economic factors on

disease*5. Characterise in detail genetically or other risk factor

defined subpopulations to understand natural history of disease and study biological effects of these events (intensively phenotyped subpopulations)

Biobank Objectives – April 2005*6. NHS utilization*7. Utilise biomarkers (protein, small molecules) to define

disease subtypes and as markers of environmental exposures

*8. Identify disease related risk factors using genomicallyderived prediction and progression testing

9. Pharmacogenetics (?)

NHSutilisation

Social factors

in disease

Biomarkers& disease surrogates

Environ-mental

determinantsof disease Disease

genes

Intensive Phenotypingof genotypepopulations

Psycho-logical factors

in disease

Diagnostics

GeneEnvironmentinteraction

Pharmaco-genetics

BioBank500,000

participants

Experimental Medicine: Genotype-Phenotype correlations and disease

β2 Adrenoreceptor (BAR)-2 gene PPAR Pro12/Ala

Data from OBB

Haplotype Glucose(mmol/l)

BAR 6/6 (n=25) 5.32 (0.08)BAR 2/2 (n=170) 5.18 (0.03)BAR 4/4 (n=100) 5.04 (0.04)

ANOVA, P<0.01

Timetable

• 2003: 6 RCCs selected. • 2004/5: Protocol incl questionaire,

physiological measurements, recruitment strategy developed

• 2004/5: Storage facilities established• 2005: Phase 1 Pilots began Feb to

evaluate assessment visit• Phase 2 Pilots begin February to assess

recruitment strategies

Assessment visit

• 120 minutes• Consent: paperless• 140 questions on touch screen for direct

data entry• Direct interview for cognitive function• Measurements• Blood, urine collected

Measurements

• Height• Weight• Impedence• BP• Waist-hip• Spirometry• Hand Grip• ECG

• Ankle-brachial index• Pulse-wave velocity• Carotid IMT• Axial DEXA scan• Calcaneal BMD• Timed Shuttle test• Spirometry

reversibility

REMEASUREMENT AT 1 MONTH IN 10%

Questionnaire

• 140 core questions with 80 stems• Separate consideration of psychological

status, cognitive function, environment and diet

• Diet: 24 hour recall by touch screen and subsequent remeasurement

Why Prospective Study?• Most accurate approach to defining

environmental exposures• Essential for understanding impact of premorbid

conditions on disease natural history (cognitive function, anxiety, biochemical and metabolic variation)

• Evaluation of predictive biomarkers in disease• Proven value of prospective studies

(Framingham, Nurses Health Study, Doctors Study, MRFIT, EPIC)

0 20 40 60 80 100 120

OXCHECK

Scottish Heart Study

Paisley & Renfew

Oxford/FPA

Whitehall

National Child Development

ALSPAC

BUPA Cohort Study

EPIC-Norfolk

SEARCH

British Doctors' Study

RCGP Oral Contraception Study

EPIC-Oxford

Heart Protection Study

Million Women Study

University of Oxford

Smaller UK studies: Regional Heart Study 8000; 1946 Birth Cohort 5000; Caerphilly & Speedwell 4000; Ely study 2000.

Cohort size (thousands)

UK COHORT STUDIES (>10,000 participants)

Phenotyping• Extensive at recruitment

– Full biochemical screen– Toxin profile– Physiologic BP, ECG, cardiovascular, dopplers– Psychometric testing: anxiety, cognitive function– Socio-economic parameters

• Opportunities for new phenotyping tools on stored samples (biomarkers)

• Rephenotyping in defined subsets for validation and specific studies of subpopulations defined by risk factors

After 10 years of follow-up

After 5 years of follow-up

ASCERTAINABLE THROUGH ROUTINE PRACTICE AND OTHER FOLLOW-UP

115005800Diabetes millitus (incidence)82504100Myocardial infarction (incidence)45002250Stroke (incidence)20001000Dementia1700850Rheumatoid arthritis (incidence)1300650Parkinson’s disease (incidence)1200600Hip fracture (incidence)

2540980Lung cancer (incidence)1410610Non-Hodgkin’s lymphoma (incidence)1130430Bladder cancer1120510Ovarian cancer (incidence)855325Stomach cancer (incidence)

2540980Lung cancer (incidence)1410610Non-Hodgkin’s lymphoma (incidence)1130430Bladder cancer1120510Ovarian cancer (incidence)855325Stomach cancer (incidence)

After 10 years of follow-up

After 5 years of follow-upASCERTAINABLE FROM ROUTINE DATA

Total number of cases expectedCondition

Prospective Studiesage range

suggestions to include younger age groups

45-69 35-69 35-44EXPECTED DEATHSischaemic heart disease 4300 3200 130cerebrovascular disease 1100 830 40

EXPECTED INCIDENCEbreast cancer 6270 5140 640colorectal cancer 5410 4000 150prostate cancer 3290 2360 6

myocardial infarction 8250 6270 380*stroke 4500 3370 170

*based on rates in participants 35-49

Canada

USA ChinaMexico

UK Europe

Large Prospective Cohorts with Biological samples

What is novel about Biobank?

1. Size of study2. Breadth of biological resource3. Consideration of biomarkers and

genes4. Recall of subsets for intensive

phenotyping5. Novel Storage Facility6. Broad use of NHS records

MOLPAGE

European Consortium

• 10 Universities/Research Institutes• 2 Pharma• 6 Biotech/SME

• 4 Year Programme• European Union FP6 • 12 million Euros

Molecular Phenotyping to Accelerate Genomic Epidemiology

MolPAGE Goals

Molecular Phenotyping to Accelerate Genomic Epidemiology

“The application of genomic, metabonomic and proteomic tools to develop new technical, data

analysis and integration protocols that will enable large-scale biomarker typing

and discovery projects”

Clinical Focus

The consortium will initially focus on;

(1) Type 2 diabetes – to identify biomarkers indicative of a pre-clinical status and increased risk for developing diabetes

(2) Cardiovasuclar – to identify biomarkers indicative of an increased risk of developing vascular disease in diabetics

Later, the protocols will be applied to;

(3) Cancer

MolPAGE Programme

The programme is divided into three parts:

• The evaluation of sample collection and storage methodology; to optimally reduce sample variation and maximise analyte stability

• The development of genomic, metabonomic and proteomic tools;for molecular phenotyping on an epidemiologic scalefrom readily obtainable clinical samples

• The development of bioinformatic tools for data warehousing, data interrogation and statistical analysis in large sample sets

Approach

The consortium will adopt two separate approaches;

(1) Genome wide, systematic methodology (Metabonomics, mass spec based proteomics)

(2) Limited analysis, using sets of candidate biomarkers (Transcript profiling, DNA methylation studies, affinity arrays, protein arrays, tissue arrays)

Platforms For theanalysis of large sets ofbiomolecules

Sample collection& standardisation

of processing,storage anddescription

DNAsequence

Epigenomics

Transcriptprofiling

Mass Spec proteomics

Meta-bonomics

databasing Stat.analysis

Peptidomics

Affinityproteomics

Platforms for the analysis of focused setsof bio-molecules

WP1

WP10

WP4

WP5

WP5

WP2

WP6,7

WP3

WP9WP8

WP10

WP4

WP5

WP5

WP2

WP6,7

WP3

Training

WP11

Management

WP12

Work Package Overview

Consortium Partners & Projects (1)

Work Package 1: Sample Collection, Processing & Storage

• University of Oxford• Guys and St Thomas Hospital NHS Trust • Obesity Research Unit, INSERM• Charles University, Prague

• Provision of blood, serum, plasma, RNA, urine, fat and metabolic tissue samples to all technology development work packages

• Collection of twin samples to measure variability

• Establish optimal collection and storage parameters

Sample Collections

MUSCLE, LIVER, FAT

~100Individuals undergoing surgerySurgical samples (Sx)

BLOODURINE

~1000Families with early onset T2DYoung Onset (YO)

BLOODURINE, FAT

>4800 twin pairs

Caucasian twin pairs aged 18-80 (50% MZ)

St Thomas’ UK Adult Twin Registry(TWINS)

BLOODURINE, FAT

~100Obese individuals taking part in a weight reduction program

PraToulOb study(PTO)

BLOODURINE, FAT

>1000 indivsPopulation based sample of UK middle aged individuals

Oxford Biobank(OXBB)

BLOODURINE

~290 families~1450 indivs

UK sibships asc’d for proband with BP >5th centile

HTO(HTO)

BLOODURINE

~600 families~2500 indivs

UK sibships ascertained on basis of single T2D and several unaffected sibs

Progene / Diabetes in Families(Prog: DIF)

BLOODURINE

CELL LINES

~827 families~2600 indivs

UK pedigrees segregating T2D and unaffected relatives

Warren 2 families (W2)

AvailableNumbers DescriptionName

Twin Studies

MZ

40 pairs

MZ female Representative for BMI

DZ

20 pairs

MZD

10 pairs

DZ female Representative for BMI

MZ same sexDiscordant for BMI*

(two visits)+ extra biopsies

From each twin collect;

Blood For RNA

Urine

Blood For EBV

Serum Plasma

Adipose

*Repeat for 10 pairsDiscordant for T2D

Consortium Partners & Projects (2)

Work Package 2: Metabonomics

• Imperial College & Novo Nordisk

• Optimising bio fluid Nuclear Magnetic Resonance to identify biomarkers

• To produce an NMR bio-fluid database in man•

Work Package 3: Transcript Profiling

• Novo Nordisk

• RNA analysis using Affymetrix array technology to identify diagnostic transcription patterns in PBL, EBV transformed lymphocytes and fat

Early Biomarker promise…..

“separation” in metabonomicspace of samples from individualswith normal coronary arteries and triple vessel disease

Transcript profiling defines breast cancer prognosis

Van de Vijver et al NEJM 2002

Schadt et al. Genetics of gene expression surveyed in maize, mouse and man. Nature 422, 297 – 302, 2003

Consortium Partners & Projects (3)

Work Package 4: DNA methylation studies

• Epigenomics AG, Berlin• Centre National de Genotypage

• To characterise the DNA methylation profile of candidate genes in disease using E SMETM software

• To identify de novo differentially methylated genes in DNA from fat and PBMC’s

Consortium Partners & Projects (4)

• Centre National de Genotypage• Oxford Gene Technologies• Roche, Basel, Switzerland

• Clinical Proteomics (ClinProt) •• Differential peptide Display

• 2D gel peptide mass fingerprinting

Work Package 5: Mass Spectrometry Proteomics

• BioVisioN AG, Hannover• Centre National de Genotypage

Consortium Partners & Projects (5)Work Packages 6 & 7:Affinity Ligand & Affinity Array Proteomics

• Royal Institute of Technology, Stockholm• Affibody AB, Bromma• Centre National de Genotypage

• Generation of 200 antibodies to human serum proteins – including known and de novo MolPAGE biomarkers

• Tissue microarrays using control and diseased metabolic tissue•• First generation protein array based on affibodies

• First generation ‘Compact Disc’ for liquid handling of serum and mass spec detection

• University of Uppsala• Gyros AB, Uppsala

Status of chromosome 21 proteomics project

Done

No protein

No RNA

January 25, 2002

Work Package 8:• University of Oxford Data analysis & statistical• Guys and St. Thomas Hospital Trust methods development

Work Package 9:• EBI, Cambridge Sample Database &• IMCS, Riga Data warehousing

Work Package 10:• Centre National de Genotypage Genetic Analysis

Work Package 11:• University of Pavia Training & Mobility

Consortium Partners & Projects (6)

Sample Collections

Prog

Tissue arrays

eQTL study

Biomarker discovery

Sample processing

Heritability, variability etc

Methods developmentothergdmYOSxObbW2TwinsSamples for:

Progene(&DIF)

OxBB(and PTO)

Young-onset

DNARNAProteinSamples for metabonomicsEpigenomics

Plasma/serumUrineAdipocyte biopsiesSurgical samples: liver, fat, muscle

TECHNOLOGY COLLECTIONS

GENOMIC EPIDEMIOLOGY

....subsequent doubts

Sample Collections

MUSCLE, LIVER, FAT

Large biopsies of muscle, fat, liver possible

~100Individuals undergoing surgery

Surgical samples (Sx)

BLOODURINE

Individuals with known genetic basis for their diabetes

~1000Families with early onset T2D

Young Onset (YO)

BLOODURINE

FAT

Extensive bioclinical data for majority including DNA; genome scan complete on 1500 DZ pairs

>4800 twin pairs

Caucasian twin pairs aged 18-80 (50% MZ)

St Thomas’ UK Adult Twin Registry(TWINS)

BLOODURINE

FAT

Fat biopsies available before and after weight reduction program; extensively phenotyped

~100obese indivs taking part in a weight reduction program

PraToulOb study(PTO)

BLOODURINE

FAT

Extensively phenotyped for metabolic traits; consented to further contact for detailed physiological studies

>1000 indivsPopulation based sample of UK middle aged individuals

Oxford Biobank(OXBB)

BLOODURINE

Extensively phenotyped for cardiovascular and metabolic traits; genome scan completed;

~290 families~1450 indivs

UK sibships asc’d for proband with BP >5th centile

HTO(HTO)

BLOODURINE

Extensive intermediate phenotypes;Proportion fo these currently being resampled in 2005

~600 families~2500 indivs

UK sibships ascertained on basis of single T2D and several unaffected sibs

Progene / Diabetes in Families(Prog: DIF)

BLOODURINE

CELL LINES

Cell lines; extensive phenotypes; genome scan for linkage completed; longitudinal study in unaffected relatives (=Progene – see next item)

~827 families~2600 indivs

UK pedigrees segregating T2D and unaffected relatives

Warren 2 families (W2; Progene)

AvailableAdded scientific value Numbers DescriptionName