gwas-intro-lombard_11111111111111111.pdf

mm

1. GENOME-WIDE
ASSOCIATION STUDIES ~
GWAS
Zané Lombard; Wits Bioinformatics

2. Identifying disease genes
 Identifying genes that contribute to disease risk is one
of the main objectives of molecular research
 Such findings have contributed to improvements in
diagnosis, prognosis and therapy
 With the successful identification of disease genes for
many single-gene disorders, the focus has shifted to
diseases with a complex, multifactorial aetiology
 What are the approaches available?

3. How our genome can influence disease
 Chromosomal abnormalities
 Mutation (direct cause and effect)
Normal red
blood cells
Sickle red blood
cells

4. How our genome can influence disease
 Multifactorial inheritance

5. Identifying disease genes – association
 Association refers to the co-occurrence of a
genetic variant with a disease trait, more
frequently than can be readily explained by
chance.
 Two approaches to association studies:
 Candidate gene approach
 Genome-wide approach

7. © Francis Collins, 2008 © Francis Collins, 2008

8. © Francis Collins, 2008
© Francis Collins, 2008

9. How is GWAS possible?
 Decide to genotype 12 million common SNPs
 Collect 1,000 cases and 1,000 controls
 Genotype all DNAs for all SNPs
 That adds up to 24 billion genotypes
 Imagine, this approach cost 50 cents a genotype.
 That’s R12 billion for each disease/trait –
completely out of the question!!

10. IMPORTANT CONCEPTS
FACILITATING GWAS

11. SNP discovery
 Most common type of genetic variation
 In latest build of dbSNP (138) there are 62,676,337 SNPs
 SNPs are scattered throughout the genome
 The abundance of SNPs and the ease with which they
can be measured make these genetic variations
significant
locus Allele: A Allele: C
Genotype: AC

12. Common Disease – Common Variant
Hypothesis
 Common disorders are likely influenced by genetic
variation that is also common in the population.
 Common variants have small effects (low penetrance)
 Multiple variants contribute to disease susceptibility

13. HapMap and Linkage Disequilibrium

14. tagSNPs & Indirect Association
 Based on analysis of data from the HapMap project,
>80% of commonly occurring SNPs in European descent
populations can be captured using a subset of 500,000
to 1,000,000 SNPs scattered across the genome.

15. Technology advances
Affymetrix
Illumina

16. Designing a GWAS
 Phenotype measures
 Case-control vs Quantitative design
 Sample size & Power
 Population considerations

17.     
   
   
   
    
   
   
   
   
   
   
   
CASES CONTROLS
Population stratification

18.     
   
   
   
   
   
CASES CONTROLS
Population Group 1
Population Stratification

19.     
   
   
   
   
   
CASES
CONTROLS
Population group 2
Population stratification

20.     
   
   
   
    
   
   
   
   
   
   
   
CASES CONTROLS
Population stratification

21. ANALYZING GWAS
DATA

22.  Image files (±400Mb/individual)
 Data files (±600Mb)
What does the data look like?

23. Analysis of a GWAS
 Quality control
 Statistical testing
 Covariates
 Multiple testing

24. Quality Control
 Batch effects
 SNP quality control
 Missingness
 HWE
 MAF
 Sample quality control
 Missingness
 Relatedness!
 Population structure

25. Statistical testing - Association
 Quantitative traits
 Generalized linear model (GLM) approaches
 Case-control
 Logistic regression

26. Manhattan plots & Regional plots

27. Multi-locus Analysis: Allele Scoring

28. Multiple testing
 The cumulative likelihood of finding one or more
false positives over the entire GWAS analysis is
high.
 Bonferroni correction:
 Adjust = 0.05 to = (0.05/k) where k is the
number of statistical tests conducted.
 False-discovery rate
 Permutation testing

29. Finishing your GWAS work
 Replication
 Meta Analysis
 Functional Studies