6 - Introduction to Genetic Epidemiology

Question 1

Modes of Inheritance

Answer 1

Mendelial

Non-Mendelian

Question 2

Mendelial

Answer 2

Autosomal Dominant
Autosomal Recessive
X-Linked Recessive

Question 3

Non-Mendelian (or Complex)

Answer 3

High Heritability (Birth defects, Cleft palate, Neural tube defects)
Low Heritability (CAD, Parkinson's)

Question 4

Twin Studies

Answer 4

Are monozygotic twins more similar than dizygotic twins?

Assumption: They have the same exposures

Question 5

Candidate Gene Study

Answer 5

Single nucleotide polymorphism (SNP)
or haplotype (characteristic pattern of
SNPs) in gene that codes for protein in
known pathway
Tests association of SNP or haplotype
with presence of disease or phenotype

Question 6

Threats to Validity - Candidate Gene Studies

Answer 6

Study Design Problems
Cohort study = confounding
Case control = confounding +
selection bias
Misclassification
Endpoint; genotyping
Multiple comparisons
Lack of replication

Question 7

Threats to Validity - Confounding

Answer 7

Genetic
SNP of interest is correlated with
causal SNP
In gene or neighboring regions
(linkage disequilibrium)
More generally (population
stratification)
Non-genetic

Question 8

Population Stratification - Strategies to address

Answer 8

Family-based approaches
Restriction
Genomic control/ancestral
informative markers

Question 9

Non-Genetic Confounding - Strategies to address

Answer 9

Measure and adjust for non-genetic

confounder

Question 10

Misclassification - Types

Answer 10

Disease misclassification
Genotyping errors
Quality control
Call rates
Hardy-Weinberg Equilibrium
χ2 of observed vs. expected allele
distribution

Question 11

Multiple Comparisons

Answer 11

Since two alleles in every locus, multiple
possible combinations
Additive GG to GA to AA
Dominant GG vs. GA or AA
Recessive GG or GA vs. AA
Multiple SNPs per gene and multiple
genes

Question 12

Multiple Comparisons - Approaches

Answer 12

Bonferroni correction, False Discovery Rate

Replication / meta-analysis

Question 13

Genome Wide Association Study

Answer 13

Purpose: “discovery” of new pathways
Common disease, common variants
Measures SNPs dispersed across the
genome to allow genome-wide scan
– ~1 million SNPs
– Randomly or purposefully selected
Unconstrained by prior hypotheses
Requires large sample sizes

Question 14

Dealing with False Positives (GWAS)

Answer 14

1 million tests at P

Question 15

Threats to Validity - GWAS

Answer 15

Underlying Study Design
False Positive Results
Genotyping/data errors
Confounding by Population
Stratification

Question 16

Limitations of GWAS

Answer 16

Require large samples sizes, dense
genotyping
False negatives
Early GWAS had limited information on
environmental factors, confounders
Modest associations
Identify regions of interest, not
necessarily “causal” SNP

Question 17

Mendelian Randomization

Answer 17

Assumption: random assortment of gene
variants from parents to offspring during
gamete formation and conception
AG x AG =
AA (25%) AG (50%) GG (25%)

Uses “latent” variable to test if pathway
“causes” disease

Question 18

Instrumental Variable

Answer 18

Affects disease only through gene product

Question 19

Limitations of Mendelian Randomization

Answer 19

“Failure” of randomization
GG x AG = 
GG (75%)
Confounding (linkage disequilibrium
population stratification, non-genetic)
Requires bigger sample sizes
Causal SNPs often not known; weak
correlations

Question 20

Gene x Environment Interaction

Answer 20

Genes may cause disease and/or
modify risk of an environmental stimulus
Smoking, alcohol…
Medications…
Test for interaction term for gene X
environmental factor