Genetic Association Study Part II

1
Genetic Association Study Part II

• ???

2
Genetic association studies

• Genetic association studies are central to
  efforts to identify and characterize genomic
  variants underlying susceptibility to
  multifactorial disease
• However, obtaining robust replication of initial
  association findings has proved difficult.

3
Factors affect the quality of association studies

• Much of this inconsistency can be attributed to
  inadequacies in study design, implementation, and
  interpretation
• inadequately powered sample groups are a major
  concern
• appropriate sample-recruitment strategy
• logical variant selection
• minimum genotyping error
• relevant data analysis
• valid interpretation

4
Differences between linkage and allelic
association

• Linkage Genetic linkage refers to the locus on
  the genome it does not refer to specific alleles
  at the locus
• purpose of genetic linkage studies to test if
  the disease gene is at a particular locus on a
  chromosome or a genome
• Association A disease is associated with a
  specific allele of a gene
• purpose of association studies to test if a
  particular allele of a gene is associated with a
  disease

5
Advantages/disadvantages of linkage vs.
association

• Linakge - systematic a few hundred markers can
  be used to scan the whole genome
• systematic but not powerful
• Allelic association - powerful tens of thousands
  of DNA markers would be needed do a genome scan)
• powerful but not systematic
• Recently more researchers favoring allelic
  association studies using a genome scan approach

6
Allelic association

• Because tens of thousands of markers would be
  needed to do a genome scan with an allelic
  association
• association has been used primarily to
  investigate associations with candidate genes
• Problem with candidate gene approach
• often dont have strong hypotheses as to which
  genes are candidate genes

7
Types of allelic association studies

• 1. Case-control studies
• Simple test for association
• allelic frequency differences between affected
  and control individuals
• Advantage of case-control study
• very simple
• Disadvantage of case-control study
• could be biased if affected and control
  individuals are not adequately matched on
  ethnicity (i.e. population stratification)

8
Population Stratification

• Ethnic groups often yield different allelic
  frequencies for DNA marker
• Example in samples that included Chinese and
  Caucasian subjects associations would be found
  for any DNA marker that showed allelic frequency
  differences between Chinese and Caucasian
  subjects with use of chopsticks

9
Types of allelic association studies

• 2. Within family studies
• avoids the problem of ethnic stratification bias
  by examining associations within families

10
I. Discordant siblings

• Straightforward way to investigate associations
  within families
• Compare discordant siblings in which one sibling
  is affected and the other is unaffected
• The hypothesis
• allelic frequencies should differ in the two
  types of siblings (affected and unaffected)

11
I. Discordant siblings

• A potential problem
• if a trait is heritable, may not find many sib
  pairs who are discordant phenotypically
• Another way examine discordant siblings
• start with genotypically discordant siblings and
  see if they differ phenotypically

12
II. Transmission disequilibrium test (TDT)

• uses trios consisting of affected individuals and
  their biological parents
• affected individuals
• must have received the susceptibility alleles
  from their parents
• alleles transmitted from parents to affected
  individuals viewed as the case alleles
• alleles not transmitted from parents to affected
  individuals viewed as the control alleles

13
II. Transmission disequilibrium test (TDT)

• TDT
• testing departures from the expected equal
  frequency of transmitted and nontransmitted
  alleles (i.e., 50 transmitted and 50
  nontransmitted)

14
Questions to Ask in Association Studies

• How good a candidate is the gene?
• How strong is the case for the genetic variants
  that have been typed?
• How appropriate are the samples typed?
• Is the study size large enough?
• How good is the genotyping?
• How appropriate is the analysis?
• How appropriate is the interpretation?

15
Q1 How good a candidate is the gene?

• Evidence from a range of sources can be used to
  identify genes (so-called candidates) with higher
  prior odds for phenotypic involvement
• Detection of a true association demands
• the gene product be involved in pathways relevant
  to the development of the trait of interest
• the gene contains variants capable of influencing
  its regulation or function.

16
Q2. How strong is the case for the genetic
variants that have been typed?

• Necessary to examine, in large samples, every
  base at which variation might conceivably alter
  gene function or expression
• Too expensive to examine many hundreds of
  variants in several thousand people for every
  gene of interest.
• The subset of variants to be typed has a
  substantial effect on the power and quality of
  the study.

17
Q2. How strong is the case for the genetic
variants that have been typed?

• There are two complementary approaches to the
  selection of variants.
• Tagging exploits the extensive linkage
  disequilibrium in many parts of the genome
• Assessments of the likely functional effect of
  variation within a gene or region of interest
• Further advances in genotyping, difficult
  decisions about SNP selection will no longer have
  to be made.

18
Q3. How appropriate are the samples typed?

• As in conventional epidemiology, the prospective
  cohort study is often regarded as the gold
  standard
• Unless the disease is very common, the study
  samples generated will have far fewer individuals
  with disease than without, and the nested
  case-control samples that emerge will often be
  small

19
Q3. How appropriate are the samples typed?

• The case-control study remains the mainstay of
  genetic association studies, and the most
  important issues relate to choice of the two
  study groups.
• Selection of cases is the easier task because
  clinical presentation facilitates recruitment
• Controls must be selected from the same
  population as cases, or genomic control methods
  must be used to correct for population
  stratification.

20
Q4. Is the study size large enough?

• The key determinant of quality in an association
  study is sample size

21
Q4. Is the study size large enough?

• These calculations assume that the susceptibility
  variant itself (or a marker in complete linkage
  disequilibrium) has been typed, which is a
  best-case scenario
• However, large sample size alone is no guarantee
  of validity.
• Increasing size reduces sampling error but
  carries an increased danger of false positives as
  hypothesis testing becomes highly sensitive to
  the consequences of small bias effects due, for
  example, to population stratification.

22
Q5. How good is the genotyping?

• Most association studies assume implicitly that
  the genotypes are accurate. However, even with
  the best methods, some assays will be unreliable.
• Reducing the effect of genotyping error requires
  more accurate genotyping platforms. Such methods
  are now becoming available, but not all
  researchers yet have access

23
Q5. How good is the genotyping?
24
Q5. How good is the genotyping?
25
Q6. How appropriate is the analysis?

• There are many ways of inferring haplotypes from
  unphased genotype data (as obtained in a sample
  of unrelated individuals)
• The conceptual and computational differences
  between these methods translate into substantial
  differences in performance and, sometimes, in
  outcome

26
Q6. How appropriate is the analysis?

• The best guarantor of probity is likely to remain
  a significant input from experts in statistical
  genetics during study design, data analysis, and
  peer-review.
• Additional reassurance comes when a finding
  remains consistent when several complementary
  methods are used.

27
Q7. How appropriate is the interpretation?

• The perceived unreliability of association
  studies has generated much discussion about the
  level of evidence needed before an association
  can be regarded as proven.
• If there are around 106 variants within the
  genome that can, in principle, affect any given
  human trait, an appropriate threshold might be a
  p value of 5108 (that is the standard 005
  corrected for 106 tests).

28
Q7. How appropriate is the interpretation?

• Although the increases in sample size necessary
  to satisfy more stringent criteria are
  surprisingly modest (table 4), they can still
  exceed the scope of many current collections.

29
Strengthening the Reporting of Observational
Studies in Epidemiology (STROBE)

• STROBE StatementChecklist of items that should
  be included in reports of case-control studies
• PDF

30
References

• What makes a good genetic association study?
• Andrew T Hattersley, Mark I McCarthy
• Lancet 2005 366 131523
• Strengthening the Reporting of Observational
  Studies in Epidemiology (STROBE) - Explanation
  and Elaboration
• Jan P. Vandenbroucke et al.
• Epidemiology 2007 18 805-835

31
HOMEWORK

• ? ?????? My Sister's Keeper
• ?? ?????? Jodi Picoul
• ??/? ????? Blueprint
• ?? Charlotte Kerner
• ?? ????? The 6th Day
• ????????
• ?? ????The Island
• ?? ?????
• Anything regarding clone or related issue

32
HOMEWORK

• Select one of the books/movies
• Read / watch it
• Share your thoughts in 5 minutes on 6/18