Module 2 - content Flashcards
(59 cards)
1
Q
biallelic marker
A
3 possible genotypes
2
Q
Pleitropy
A
one gene influences more than one or unrelated phenotypes
3
Q
Penetrance
A
percentage of people who carry allele and express phenotype
4
Q
Linkage disequilibrium
A
non-random assortment of alleles at 2 or more loci, stronger = decreased recombination
5
Q
Haplotype
A
specific combination of alleles occurring on the same chromosomes
6
Q
Heritability
A
proportion of phenotypic variance in a population that is explained by genetic variations
7
Q
Genetic variation associated with disease
A
can be the sole cause of disease or it can increase the risk of common diseases
8
Q
GWAS aim
A
detect associations between genetic variants and phenotypes in a population
9
Q
GWAS goal
A
- better understanding of biology
- develop treatments or interventions
9
Q
GWAS reality
A
often identifies associated marker rather than causal variants
10
Q
Steps of a GWAS
A
- cohort selection
- data collection
- genotyping
- quality control
- imputation
- association testing
- meta-analysis
- replication
- post-GWAS analysis
11
Q
GWAS cohort selection factors
A
- population based GWAS
- Family based GWAS
- need to consider age, sex, ethnicity, country, school etc
12
Q
Population based GWAS selection
A
- genotyping and phenotyping done from individuals which are randomly chosen from the population
- usually case-control studies = presence or absence of a certain phenotype
- active recruitment of cases and controls can increase statistical power
- case and control should be genotyped together on the same chip to increase quality and statistical power
13
Q
Family based GWAS selection
A
- needs a greater sample size to achieve the same statistical power of unrelated individuals
- avoids population stratification
14
Q
Data collection
A
- can be from cohorts or biobanks
- usually need a large sample size for good statistical power
- biobanks = data from thousands of genotyped individuals who have been phenotyped via questionnaires etc
15
Q
UK biobank
A
- exome and whole genome sequencing
- investigates the contributions of genetic variation and environmental exposure to the development of disease
- participants are healthier, healthier and more educated than the general population
- European ethnicity
16
Q
China kadorie biobank
A
- investigates the genetic and enviro causes of common chronic diseases in the chine population
- have a custom genotyping platform
17
Q
Genotyping techniques
A
- microarrays
- next generation sequencing for whole genome or whole exome sequencing
- custom SNP arrays
- commercial SNP arrays
18
Q
Custom SNP arrays
A
- expensive
- allows increased genomic coverages for populations other than European
19
Q
Commercial SNP arrays
A
- cheaper as they are already made
- usually only European so doesnt include other ethnicities
20
Q
Quality control
A
- removal of rare variants
- ensure phenotypes are well matched with genetic data (sex vs chromosomes)
21
Q
Imputation
A
- allows the evaluation genetic markers which are not directly genotyped
- increases the power of GWAS as it includes SNPs which may be poorly targeted on chips thus may initially be mised
- done using a sequenced haplotype reference panel such as the 1000 genomes project
22
Q
Limitation of sequenced haplotype reference panels
A
excludes indigenous populations
23
Q
Association testing
A
- linear or logistic regression models
- includes covariates such as age, sex and ancestry
- include a random effect term to increase statistical power
- consider linkage disequilibrium
- account for false discovery
24
Meta-analysis
- in consortium = data from multiple cohorts are analysed together
- performed on the summary statistics of these consortia
- combining contributions of all cohorts all for a more precise estimation of effect sizes and the significance effects in GWAS by weighting individual cohorts by their sample size
25
Challenges for consortia
- different analysis strategies and phenotype modelling between studies
- sample size = increased size can cause noise
- modifiable effects or interactions not accounted for such as enviro, age and sex
26
Post-GWAS analyses - PRS
- polyrisk score = predict the risk of disease in a target cohort
- calculated as a weighted sum scores of risks with weights based on effect sizes from GWAS
- SNPs selected based on P values from GWAS
27
PRS limitations
- scores have been largely calculated from European DNA sequences = decreased accuracy in populations distinct from Europeans
- amount of variance in a phenotypic trait explained by a PRS is limited by the heritability of the trait
- same size
28
Statistical power to detect associations depends on
- experimental sample size - larger = more power
- distribution of effect sizes of unknown causal genetic variants
- frequency of causal variant
- effect size
29
GWAS limitations
- population stratification
- polygenicity
- ethics
30
Population stratification
- differences in allele frequencies between groups due to differences in ancestry rather than a genetic association
- many disease/trait associated SNPs do not replicate in studies of populations from different ancestries
31
Polygenicity
- most diseases are polygenic so it can be hard to uncover the underlying biological mechanism
- thousands of variants can each have a small effect on a trait
- can be helped using WES and WGS to find rare variants of larger effect sizes
32
Ethics
- issues regarding consent for future use of samples and data
- lack of diversity and inclusion
- PRS proposed as method for emryo selections
33
Strategies to uncover new susceptibility loci
- better phenotypes to decrease noise
- better statistical approaches
- larger sample sizes
- increased diversity
34
Depression
a mood disorder that is defined as chronic and persistent feelings of sadness, lack of pleasure in things they used to enjoy and behavioural changes
35
Measuring depression
- self-report questionnaires
- clinical interviews
- behavioural tasks
- neuroimaging
- pharmacological challenges
36
Self-report questionnaires benefits
- low cost
- easy
- allows larger sample size = increased statistical power
- many scales are validated and standardised
37
Self-report questionnaires limitations
- low specificity due to restriction on the number and type of questions
- some cant be used for diagnosis
- potential response bias
- may not be suitable for all populations
38
Clinical interview benefits
- higher specificity
- can be tailored to individuals
- high validity
39
Clinical interview limitations
- high cost
- time consuming
- might not be suitable for all populations = limited verbal or cognitive abilities
40
Behavioural tasks
computer based tasks designed to assess cognitive processes associated with depression
41
Neuroimaging
brain imaging techniques to examine changes in brain activity and connectivity associated with different phenotypes
42
Pharmacological challenges
- pharmacological agents used to provoke or alleviate symptoms of depression phenotypes
- allowing investigation of biological mechanisms
43
Future research for depression
need to look broadly and specifically at depression types as we have a poor understanding
44
How do we distinguish which genetic loci are associated with which depression phenotype
stratify datasets by relevant clinical symptoms and assess genetic profiles
45
Heritability differences in genders
- women = more likely to internalise
- men = more likely to externalise
- shown in twin studies
46
Factors associated with depression
- early life
- risk behaviours such as smoking and substance abuse
- location
- co-morbidities
- demographics such as age or gender
47
sample size relationship with power
- larger sample size = more data points
- more data points = increased chance of detecting those of smaller effects = more power
- smaller samples and low effect sizes = increases the risk of false negatives = reduced power
- too high of a sample size wastes time and resources
- too small of a sample size you are unlikely to detect any difference even if there is one
48
relationship between power, sample size and effect size
- if you have a small effect size you would need a large sample size to detect it > makes false positives less likely = more power
49
Relationship between allele frequency and statistical power
- rare or uncommon alleles are hard to detect as they usually have a small effect size = reduced statistical power
- rare variants would require a large sample size to obtain the small statistical power of a common variant with a large effect size
50
The importance of considering covariates and environmental factors when undertaking a genetic association analysis
- Covariates = age, gender ancestry
- Environmental = diet, life events etc
- These factors can interact with genetic factors leading to an increased risk of a condition such as depression
- Crouse et al. 2024 > Correlation between PRS for MDD in women and experiencing SLE such as childhood abuse
- Nelemans et al. 2020 > high parental criticism amplified depressive symptoms in dutch adolescents with high PRS for MDD
- Silveira et al. 2023 > found MDD females had a significant association with C-reactive protein and metabolic traits, while males had associations with dysregulated histone modifications
- Lin et al. 2023 > lowest risk of depression at 1.5hours of outdoor light exposure per day - higher risk when below or above this
51
The difference between a genetic marker in linkage and a causal genetic marker in the context of non-European populations
- Genetic marker in linkage = marker which is associated with a causal variant but is not the causal variant > measured via LD
- causal marker = marker which is the causal variant
- most GWAS data is based on Euro populations which is problematic as LD varies between Euro and Non-Euro populations
- Vergara-Lope et al. 2019 found that LD maps of African ancestries are longer than that of Euro populations
52
The similarities and differences between ancestry and ethnicity
- ancestry = genetic background, doesnt change
- Ethnicity = individually determined cultural association, can be changed
- Both can influence identity and resemble heritage
53
How is PRS derived
Genotype > GWAS > select variants > calculate PRS via summing the number of risk alleles carried by an individual by GWAS derived weight > quality control
54
What makes a good PRS
- same phenotype
- large number of variants = increased statistical power
- same ancestry
55
Depression heritability
30-40% > suggests moderate genetic contribution but is polygenic = many variants with small effect sizes
56
Disease prevalence of depression in NZ
- 15-20%
- higher rates in Maori and pacific populations due to socioeconomic disparities, historical trauma and systemic inequities
57
Example of a clinical diagnosis for depression
DSM-5 criteria
58
