Principles of Genetic Epidemiology Kirsten Ohm Kyvik Genetic

Principles of Genetic Epidemiology Kirsten Ohm Kyvik

Genetic epidemiology deals with the etiology, distribution, and control of disease (epidemiology) in groups of relatives and with inherited causes of disease (genetics) in populations (adapted from Morton and Chung 1978)

Steps in genetic epidemiology • • • § Evidence for familial aggregation Is familial aggregation due to genes or environment? Specific genetic mechanisms Taking advantage of designs involving § Families § Twins § Adoptees and their families

Fundamentals § Definition of phenotype § Classification of phenotype § Natural history of phenotype

Adaptation of concept of causation § Family status changes risk profile § Observations on family members not independent § Boundary between cohort and case-control studies is blurred

Multifactorial inheritance Monogenic Mød en forsker Quantitativ

T R E S H O L D M O D E L

Family studies

Design of familiestudies n Identify probands – ”ascertainment probability” n Information on phenotype in relatives (1. degree, 2. degree etc. ) n Compare groups of relatives n Compare with background population

Familial aggregation = genetic aetiology? Against: § Effect of:

Groups of relatives Risk of siblings compared to risk in parent-offspring • RR(sib) = RR(par) • RR(sib) >> RR(par • RR(sib) and RR(par) small, but bigger than population risk

Expected risk pattern

Parkinson’s disease in Iceland (Sveinbjørnsdottir et al. NEJM, 2000) Relatives Risk ratio p (family vs population) Sibling 6. 3 <0. 001 Children 3. 0 0. 001 Nephew/niece 2. 4 <0. 001 Cousin 2. 4 0. 1 Spouse 1. 9 0. 16

Genetic epidemiology of infantile hypertrophic pyloric stenosis The IHPS register • • • Funen based Cases from 1950 to 2004 A total of 892 cases, 870 identified in CPR Questionnaire send to all cases Reply from 65%

Smoothed prevalence

Recurrence risk in relatives Recurrence risk % (95% Confidence Interval) Group Female Population 1. degree 0. 11 Male (0. 060. 15) 0. 43 All (0. 400. 46) 0. 27 (0. 240. 30) 5. 7 (3. 9 -9. 5) 4. 4 (3. 4 -6. 1) 4. 8 (4. 1 -7. 0) Parent 4. 5 (1. 4 -7. 4) 3. 9 (2. 4 -5. 7) 4. 0 (2. 9 -6. 2) Offspring 4. 5 (0. 145. 3) 4. 5 (0. 108. 3) 4. 5 (0. 248. 3) Siblings 11. 4 (4. 017. 5) 5. 1 (3. 010. 8) 6. 6 (4. 7 -9. 8) 2. degree Grandpa rents 0. 76 (-0. 131. 5) 0. 51 (0. 101. 1) 0. 57 (0. 201. 0)

Twin studies

Aims • • • What is the risk/recurrence risk in twins Is a phenotype genetically determined Aetiological models Size of genetic variation / heritability Genes, markers, chromosomal regions Environmental determinants

DESIGNS n Classical twin study with separated MZ twins n Twin family studies n Twin-control studies

Classical twin study MZ pairs: DZ pairs:

DESIGNS n Classical twin study with separated MZ twins n Twin family studies n Twin-control studies

Is a phenotype genetically determined? • Categorical data • Continous data

Types of concordance Pairwise: Probability that both in a pair is affected: Casewise/probandwise: Probability that a twin is diseased given that the twin partner is diseased:

Probandwise concordance Estimate of the casewise probability by the proband method. 2 C 1 + C 2 --------- 2 C 1 + C 2 + D

Concordance CMZ = CDZ CMZ > CDZ CMZ <1. 0 (100%)

Solutions to problems with age at diagnosis n Survival analysis Actuarial/Kaplan Meier methodology Frailty models Newer models n Others? Correction methods

Concordance type 1 diabetes Zygosity Pairs (probands) Conc Disc Concordance Pairwise* Probandwise MZ 10(18) 16 0. 38 [0. 20 -0. 59] 0. 53 [0. 33 -0. 73] 0. 70 [0. 45 -0. 95) DZ 4 (8) 65 0. 06 [0. 02 -0. 14] 0. 11 [0. 05 -0. 21] 0. 13 [0. 04 -0. 21] ( ) Number of probands; [ ] 95% confidence limits. * Chi 21 d. f. = 10. 93, p < 0. 001 Cumulated

Cumulative concordance type 1 diabetes Interpretable as cumulative risk from birth % 0 -100 MZ 0. 70 DZ 0. 13 Age 0 -40

Correlations Twin-twin correlations r. MZ = r. DZ r. MZ > r. DZ r. MZ < 1. 0 (100%)

INTRACLASS CORRELATIONS ln. TSH (Pia Skov Hansen) MZ n=284 pairs DZ n=285 pairs r. MZ=0. 64 (CI 0. 56 -0. 70) r. DZ=0. 29 (CI 0. 18 -0. 39) p<0. 00005

INTRACLASS CORRELATIONS ln. TSH

Aetiological components § Additive genetic variance § Dominant genetic variance/epistasis § Common environmental variance § Unique environmental variance

Inheritance Models in Single Gene Trait Genotype Group Model A is Dominant A is Recessive A is Co -Dominant AA Aa aa

Inheritance Models in Quantitative Trait Model -x A is Completely Dominant aa A is Partially Dominant aa A is Not Dominant aa Population Mean 0 +x AA Aa Aa Aa AA AA

Heritability V (total) = VG + VE § V (total) = VA + VD + VI + VC + VE § h 2 narrow = VA/VA + VD + VI + VC + VE § h 2 broad = VA + VD + VI/VA + VD + VI + VC + VE §

Heritability § Function of population, NOT a constant § Does not apply to individuals § Biased if mean and variance not the same in MZ and DZ § Greater MZ covariance will inflate h 2

Correlations and aetiological models r. MZ < 1 r. MZ = r. DZ = 0 r. MZ = r. DZ > 0 r. MZ = 2 r. DZ > 0 r. MZ > 2 r. DZ r. MZ < 2 r. DZ

Aetiological models and genetic variation § Variance analysis § Regression analysis § Structural equation modelling

Path model for twin analysis

Pleiotrophy

RESULTS TSH-LEVEL Unique Environmental effect 0. 36 Genetic effect 0. 64 The genetic effects account for 64% of the variation

Multivariate ACE Model BMI Waist Gluc 12 0 Ins 0 SBP DBP HDL TG 0. 86 (0. 01) -0. 13 (0. 06) 0. 48 (0. 04) 0. 29 (0. 04) 0. 27 (0. 04) -0. 18 (0. 05) 0. 20 (0. 06) -0. 16 (0. 06) 0. 51 (0. 05) 0. 30 (0. 05) 0. 26 (0. 05) -0. 19 (0. 06) 0. 26 (0. 06) 0. 09 (0. 08) 0. 12 (0. 07) 0. 11 (0. 07) -0. 02 (0. 08) 0. 23 (0. 08) 0. 31 (0. 06) 0. 29 (0. 06) -0. 17 (0. 07) 0. 31 (0. 07) 0. 71 (0. 03) -0. 09 (0. 06) 0. 28 (0. 06) -0. 01 (0. 06) 0. 27 (0. 06) Waist 0. 85 (0. 01) Gluc 12 0 0. 02 0. 03 (0. 03) Ins 0 0. 46 (0. 02) 0. 46 (0. 02 ) 0. 13 (0. 03) SBP 0. 28 (0. 03) 0. 26 (0. 03) 0. 14 (0. 03) 0. 23 (0. 03) DBP 0. 26 (0. 03) 0. 23 (0. 03) 0. 13 (0. 03) 0. 23 (0. 03) 0. 69 (0. 02) HDL -0. 17 (0. 03) -0. 19 (0. 03) -0. 04 (0. 03) -0. 14 (0. 03) -0. 01 (0. 03) -0. 03 (0. 03) TG 0. 22 (0. 03) 0. 27 (0. 03) 0. 20 (0. 03) 0. 35 (0. 02) 0. 20 (0. 03) -0. 24 (0. 07) -0. 22 (0. 03)

Important assumptions • Biology of twinning • ”True” zygosity • Equal environment assumption • true or not true? • Generalisability

Adoption studies

Adoption design Adoptees are expected to

Early death in adoptees Cause of death Parent dead < 50 yrs Parent dead < 70 yrs Natural Bio Ado 1. 98* 0. 96 1. 49 0. 8 Infection Bio Ado 5. 81* 0. 73 5* 1 Vasculær Bio Ado 4. 52* 3. 02 1. 92 1. 5 Cancer Bio Ado 1. 19 5. 16* 0. 87 1. 49

Assumptions and problems § Early adoption § Non-familial adoption § Comparable environment in biological and adoptive family § Contact to biological family § Intra-uterine environment § Transcultural adoptions

Comparison of correlations Correlation Twin studies MZ DZ MZA 0. 7 0. 36 0. 7 Family studies PO Sib 0. 27 0. 25 Adoption studies Bio Ado 0. 17 0. 1

Comparison heritability Heritability Twin studies MZA 50 -90% 60 -70% Family studies 20 -80% Adoption studies 20 -60%