GWAS Flashcards
Define Genetic Epidemiology
Study of the role of genetic factors regulating physiological traits (e.g. LFT) and modifying disease risk
Either alone or through interaction with environmental factors
What 4 study designs are used in genetic epidemiology? Indicate which uses DNA and which does not.
No DNA
- Familial Aggregation
- Seggregation Analysis
DNA
- Linkage
- Association
Explain what familial aggregation investigates
Looks at whether there is a genetic component to a study
- Are relatives of diseased indiviuals more likely to be diseased than the general population?*
- Is the clustering of disease in families different from what you’d expect based on the prevelance in the general population?*
Explain what Seggregation analysis investigates
Looks at the mode of inheritence of the trait
Dominant, recessive, additive, X-linked
Explain what Linkage studies investigates
Identifies the chromosomal region where the causal gene is located
Explain what Association studies investigates
Identifies which genetic variant of which gene is asociated with the trait
What is the difference between a variant and a mutation
Varient is a ‘common’ change in the DNA
Mutations are rare variations (<1%)
Which mutation or varitation occurs frequently and is commonly studied and explain what it means
Single Nucleotide Polymorphism (SNP)
Single base-pair change
With regards to Gene Association Studies, what are the two main approaches to study design?
- Hypothesis-driven
- Candidate-gene Association Studies
- Hypothesis-free
- Genome-wide Assosication Studies (GWAS)
Explain the features of a hypothesis-driven approach.
- Investigation of a known variant in genes and whether is is related to a trait
- Priori hypothesis
- Gene variant chosen based on
- Biological knowledge (in vitro or animal models)
- Previous knowledge of association
What is a limitation of a candidate-gene association study (hypothesis-driven)
Can only investigate the chosen gene variant
Explain the features of a hypothesis-free approach.
- Investigates genetic variation across the whole genome
- No priori hypothesis
What is a benefit of conducting a GWAS (hypothesis-free)
- Can identify gene variatns that are no suspected to be associated with trait
- New genes could be associated to trait
- More power to detect genetic associations
Whats are the limitations for GWAS (hypothesis-free)
- Requires multiple testing (hypothesis testing in the same data set)
- Needs huge sample size to reach low significance levels
- Can only detect common variatns that has small effects
- Natural selection selects to decrease disadvantageous alleles
- Only investigates SNP variants
- CNV: Copy number variation
- SNP assessed are tags and not causal
- Variants near each other are often inherited together
- SNP acts a a surrogate for the causal gene variant using correlation knowdlege between SNP and causal variant
- Loss of power if this correlation is not strong
- Difficult to evaluate gene-environment interactions
What can be done to increase the power to identify new genes in GWA studies (3)
- Increase coverage of genotyping
- Next Generation Sequencing
- No jumps from SNP to another
- Too expensive for large smaple size
- Improve phenotyipng and increase its specificty
- Asthma phenotypes
- Incorporate prior knowledge about SNPs and genes
What is the significance of multiple testing, it’s implication and how can it be corrected?
Significance
- Normal p-value of significance (5%) has a 1 in 20 chance of false positives
- Therefore, need a lower p-value to reduce false positives
Implication
- To reach significance on low p-value, you need a huge sample size (↑↑n)
Correction
-
Bonferroni Correction
- Divide p-value by numver of tests you perform
- Significance threshold set at n=1,000,000 (i.e. p < 5 x10-8)
- Found no difference over 1,000,000
- Also addtressed through meta-analysis (international consortia)
How have GWAS improved knowledge on chronic disease
- Identification of new biological pathways to disease
- ORMDL3 - childhood asthma - Moffat et al Nature 2007
- ATGI61 - Crohn’s Disease - Rioux et al
- Discovery of links between diseases thought of as unrelated
- ORMDL3 - childhood asthma and Crohn’s disease
Describe a study that found a new gene variant associated with Severe childhood asthma, by looking at a specific phenotype in the GWAS
Bonnelykke et al Nat Genet 2014
Identified CDHR3 as an asscoiated gene variant
Only possible because they looked through a specific asthma phenotype
Define penetrance in the context of GWAS
Proportion of subjects with variant who have the disease
With regards to penetrance and allele frequency, describe:
- Mendelian disease
- Hard to identify genetically
- Highly unusal for common disease
- Variants identified by GWAS
- Intermediate
- Mendelian disease
- ↑Penetrance
- ↓Frequency
- Hard to identify genetically
- ↓Penetrance
- ↓Frequency
- Highly unusal for common disease
- ↑Penetrance
- ↑Frequency
- Variants identified by GWAS
- ↓Penetrance
- ↑Frequency
- Intermediate
- ↔Penetrance
- ↔Frequency
What is the main difference with classical and genetic epidemiology
Classical = observatrional
Genetic = Genetic association
What is the main limitation of classical epiddemiology
Difficult to establish causality due to:
- Confounding
- Reverse Causation
- Presence of diseaseoutcome that modified the exposrues not vv
What is the (only) source of confounding in Genetic epidemiology studies
Population stratification
Selecting for specific subjects
