Association Analysis

Question 1

What is genetic association?

Answer 1

Genetic association is the presence of a variant allele at a higher frequency in unrelated subjects with a particular disease (cases), compared to those that do not have the disease (controls).

For disease, we could use the broader term ‘trait’, for example, height is not a disease.

Question 2

What are case control studies?

Answer 2

Cases are subjects with the disease of interest, (e.g. obesity, schizophrenia, hypertension).

The definition of the disease must be applied in a rigorous and consistent way.

The controls must be as well-matched as possible for non-disease traits. This could be for sex, age, ethnicity, location, etc. Thus, they must be identical to the cases, but just not have the disease. This is, of course, not easy because there are many outlying factors.

Question 3

How would you use case control studies for genetic association?

Answer 3

1. You would match the affected cases and unaffected controls for all the other risk factors.
2. You would measure the genetic loci of interest.
3. You would perform a statistical analysis to determine which genetic loci correlate with each disease.
4. You would identify the genomic region associated with the disease.

Question 4

What are some features of the best case control studies?

Answer 4

• large numbers of well-defined cases (1000s)
• equal numbers of matched controls
• reliable genotyping technology (SNPs are used)
• standard statistical analysis (PLINK is used)
• positive associations should be replicated

Question 5

Why do we need so many genetic markers?

Answer 5

Individuals in a population are genetically far more diverse that individuals in a single family.

To capture this genetic diversity, we need to use 100,00s or millions of genetic markers.

Question 6

What would be some features of the ideal genetic marker?

Answer 6

• polymorphic
• randomly distributed across the genome
• fixed location in genome
• frequent in genome
• frequent in population
• stable with time
• easy to assay (genotype)

Question 7

What is a single nucleotide polymorphism (SNP)?

Answer 7

It is common in the genome (~1/300 nucleotides). There have been around 12 million common SNPs identified in the human genome.

They are generated by mismatch repair during mitosis.

Question 8

Where can you find SNPs?

Answer 8

SNPs may be in a:
GENE (CODING REGION):
- no amino acid change (synonymous)
- amino acid change (non-synonymous)
- new stop codon (nonsense)

GENE (NON-CODING REGION):

• promoter: mRNA and protein level changed
• terminator: mRNA and protein level changed
• splice site: altered mRNA and altered protein level

INTERGENIC REGION (98% of SNPs)

Question 9

What is the purpose of minor allele frequency (MAF)?

Answer 9

To check if an SNP is frequent or not, you will use minor allele frequency (the less common allele is called the ‘minor’ allele).

Question 10

What is genetic association?

Answer 10

Genetic association is the presence of a variant allele at a higher frequency in unrelated subjects with a particular disease (cases), compared to those that do not have the disease.

Question 11

How would we use genome-wide association studies (GWAS)?

Answer 11

We would use markers across the whole genome (such as SNP microarrays). We would then look for associations between the disease and each marker (we use a chi-squared test for statistical significance).

This has resulted in the detection of large numbers of disease-associated genes.

Question 12

How are GWAS results presented?

Answer 12

GWAS data is presented a single graph called a Manhattan plot.

The x-axis is the position of the SNP on the chromosome (showing position from chromosome 1 to 22). The y-axis is the -log10(p-value) of the association. This makes the results easier to compare.

The colours look solid for the most part of the graph because most SNPs are not associated with a disease.

Question 13

What do the results of the GWAS tell us?

Answer 13

The strong peaks show lots of associated SNPs, all in one genomic locus.

The peak does not identify the gene as causing the disease. The peak only identifies the genomic region associated with the disease.

Question 14

What is a major benefit of meta-analyses?

Answer 14

It would be difficult to do very large studies (>10K cases).

It would instead be easier to combine smaller studies’ results:

• pre-experiment: consortium
• post-experiment: meta-analysis

A meta-analysis allows the statistical combination of results from multiple studies.

Question 15

List some secondary characteristics of obesity.

Answer 15

• pulmonary disease
• non-alcoholic fatty liver disease
• gall bladder disease
• gynaecological abnormalities
• osteoarthritis
• skin problems
• gout
• idiopathic intracranial pressure
• cataracts
• coronary heart disease
• severe pancreatitis
• cancer
• phlebitis

Question 16

What proof is there that obesity is highly genetic?

Answer 16

From conducting twin studies, it’s been found that nearly 70-80% of our body shape is genetically determined.
Adoption studies found it to be 30-40%, while family studies found it to be 40-60%.

Question 17

What gene has been found in relation with obesity?

Answer 17

The FTO gene has been found as of significance in the obesity GWAS.

Studies all have their unique ways of defining obesity; however, all of them have found the FTO gene to be involved.

Question 18

What are some problems with GWAS?

Answer 18

GWAS has identifies associations that are statistically stong and reproducible. However, their contribution to th genetic component of disease is estimated to be low (>5%).

Some possible answers as to why:

• many common SNPS are of small effect
• not looking for rare SNPs
• not looking at copy number variation
• not looking at epigenetic variation
• heritability is overestimated