GWAS studies

Question 1

What are Genome-wide association studies asking?

Answer 1

Is that variant more common in people with a disease or trait, or people without?

Question 2

What do GWAS look for?

Answer 2

Genome-wide association studies (GWAS) involve testing genetic variants across the genomes of many individuals to identify genotype–phenotype associations.

Question 3

Describe basic outline of GWAS

Answer 3

First assemble a large set of study participants

Genotype them at all the genome locations youre interested in.

Calculate frequency of each genotype in cases and in controls (thus the Odds ratio)

The calculate astatistical significance for the Odds ratio for each variant to decide if it is involved or not.

Question 4

Whats another term for Common variants

Answer 4

SIngle Nucleotide Polymorphisms

Question 5

100 genomes project, gnomAD and TopMed projects

Answer 5

Large population studies.
Have surveyed large numbers of individuals to identify all the places they vary.

They have settled on there being around 15 million common SNPs in the genome.

Question 6

Why is there a problem with there being 15 million common SNPs?

What is the solution?

Answer 6

How do you genotype 15 million SNPs?
GZenome sequencing is too expensive.
Microarray technology is cheap, but common commercially avaliable arrays can only type 1 million variants.

Linakge Disequilibrium

Question 7

What is linkage disequilibruim?

Answer 7

Phenomenon that alleles don’t assort randomly between generation.

Question 8

What is a haplotype?

Answer 8

A haplotype is a group of alleles in an organism that are inherited together from a single parent.

Alleles at variants close together on the same chromosome tend to occur together more often than is expected by chance. These blocks of alleles are called haplotypes.

Question 9

What are haplotype blocks and what are they used for?

Answer 9

a haplotype block is a region of an organism’s genome in which there is little evidence of a history of genetic recombination

haplotype blocks have been used to increase the power of QTL (quantitative trait loci) detection in genome-wide association studies (GWAS) and the prediction accuracy with genomic selection

Question 10

What are tag-SNP’s?

Answer 10

A single nucleotide polymorphism, or SNP, that is used to “tag” a particular haplotype in a region of the genome.

Question 11

What are recombination hotspots and what does this mean?

Answer 11

Recombination isnt random in humans, but happens mostly at recombination hotspots, so as alleles arise, they are rarely seprated from the alleles they are associated with.

Question 12

What did the HapMap project do?
Why was this important?

Answer 12

Mapped these haplotype blocks across a range of human populations.

It’s now possible to genotype a million SNPs that could tag most of the SNPs in the human genome.

Commercial arrays created that allowed researchers to quickly and cheaply genotype these SNPs, effectively genotyping all SNPs present in more than 1% of the population.

Question 13

What is genotype imputation?

Answer 13

Genotype imputation is a process of estimating missing genotypes from the haplotype or genotype reference panel. It can effectively boost the power of detecting single nucleotide polymorphisms (SNPs) in genome-wide association studies, integrate multi-studies for meta-analysis, and be applied in fine-mapping studies

Question 14

What problem does this reveal with tag-SNPs?

Answer 14

If it’s found that a tag-SNP is associated with a disease, but anyone with the risk allele at the tag SNP, also has variants at all the other positions in LD with the tag-SNP, it can’t be known which of those SNPs is actually associated with the disease.

This is a big problem as haplotype blocks may cover many kb. They can contain multiple genes and potential regulatory regions.

Question 15

Once we have genotyped individuals at all common SNPs - what is the next problem?

Answer 15

It must be decided if an SNP is statistically associated with a trait of disease.

Question 16

What was a problem in early human genetic studies?

Answer 16

• Using statistical test to analyse counts of genotypes in cases and controls
• Normally 0.05 threshold used.
• 5% chance of happening by chance.
• If looking at 1 million SNPs, expect 5% to happen by chance.
• This means 50,000 flase positives.

Question 17

What is used now to reduce false positives in statisical studies?

Answer 17

A stricter threshold is used.
p< 10^-8.
Any SNP with a p value of less than this has achieved ‘genome wide significance’.

Question 18

What is the problem with this stricter threshold?

Answer 18

Just because we think that any SNP that makes the ‘genome-wide significance’ threshold is definitely associated with the disease, that doesn’t mean we think that all those that don’t definitelt aren’t associated.

For smaller P-values, many even most of them are real - we don’t know how many there are or which are real.

Question 19

What do we know?

Answer 19

Larger the effect, the smaller the p value. The larger the sample size, the smaller the p-value. This means GWAS studies require very large sample sizes to be able to detect genes.

Question 20

What are the two ways you can design a SWAS study?

Answer 20

Case-control

Cohort design

Question 21

Describe a case-control study

Answer 21

• Collection of peoople with thw trait or disease and select equal number of people matched by age, gender, ethnicity, genetic ancestry
• Make participants as homogeneous a possible, so that only differees are whether they have the disease or not.

Question 22

Describe the cohort study

Answer 22

• Recruit very large, general-purpose study population and collect many different phenotypes about them as well as their genotype.
• SUch populations can be prespective - that is you recruit participants before you even know if they will develop the disease or not.
• Such a population can be used to study many different diseases and traits, but the frequency of any given trait or disease is probably low.

Question 23

Given an example of a Cohort study

Answer 23

BioBank - has half a million participants, all genotyped and with detailed phenotypic information recoreded.

Question 24

Whats a problem in cohort study?

Answer 24

How do you know that hose with the disease are well matched in other ways to those without?

Question 25

Example to illustate problem with cohort studies

Answer 25

• Use of chopsticks
• Two replicated studies found a strong, reproducible variant that affects a students ability to use chopsticks.
• What was not accounted for was that some students were from east asian background where chopstick is culturally normal, while other from background where people do not use chopsticks regularly from young age.
• A gene was found that distinguished east-asian genetic heritage from toher background, but had nothing mechanistically to do with chopstick usage, which is a culturally determined trait.

Question 26

What is population stratification?

Answer 26

Where the average genetic background og cases is different to those of controls and/or correlates with the trait being studied.

Question 27

What is PCA?

Answer 27

Tries to reduce genotype at 1 million loci to a smaller number of important dimensions.

Question 28

What does actual size of sample required depend on?

Answer 28

How big an effect the SNPs you are looking for have.

Question 29

What is meta-analysis?

Answer 29

a statistical process that combines the data of multiple studies to find common results and to identify overall trends.

Meta-analysis is a quantitative, formal, epidemiological study design used to systematically assess the results of previous research to derive conclusions about that body of research.

Question 30

What is DIAgram consortium

Answer 30

Consortium of groups studying the genetics basis of Type II Diabetes.

Question 31

What are haplotype blocks used for?

Answer 31

Haplotype blocks have been used to increase the power of QTL (quantitative trait loci) detection in genome-wide association studies (GWAS) and the prediction accuracy with genomic selection

Question 32

type 2 diabetes

Answer 32

Associated with resistance to the action of insulin and is associated with poor diet, BMI and obesity/ Heritability estimates for diabetes range from 30% to about 70% and may differ between age groups and countries.

WHile we know that there are very strong environmental contribustions to diabetes, it is also clear that it runs in families.

Question 33

What sis DIAgram find?

Answer 33

Identified 13 loci that reached genome wide significance.
These SNPS had odds ratios between 1.06 and 1.14 (that is, risk allele carriers had between 6% and 14% higher odds of having T2D than non-carriers).

Combined with other SNPs known to affect T2D at the time, accounted for only about 10% of the estimated heritability in diabetes/

Question 34

What are eQTLs

Answer 34

expression of quantitative trait loci

SNPs that change the expression of a gene

Question 35

Where are lead SNPs found?

Answer 35

found more often than expected by chance in enhancer regions that are active in all sorts of cell types you might expect to be associated with the disease - these are statistical patterns rather than hard rules.