Page 218 - Biobehavioral Resilence to Stress

Posttraumatic Stress Disorder 195

Although there are many difficulties associated with this broad approach,

some difficulties are gradually being overcome such that whole-genome
association studies may soon hold promise for the genetic study of PTSD.
If the purpose of a genome-wide study is to query all the essential com-
ponents of the human genome (∼3 billion nucleotides in length), then the
researcher must have access to a large number of polymorphic sites in order
to avoid inadvertent neglect of potentially important genomic locations. Th e
IHP was initiated to address this need among others (Altshuler et al., 2005;
www.hapmap.org). The project’s recent data release and resources provide

comprehensive information about human DNA polymorphism. In addition
to providing information about the locations and frequencies of specifi c DNA

sequence variations in four different populations (West African, Northern
European, Chinese, and Japanese), the IHP also provides detailed informa-
tion about polymorphic site “coverage” (correlation of variations, or “linkage

disequilibrium” between variations). This information helps to ensure that
all (or most) genomic regions can be interrogated as part of a whole-genome
association study.
Whole-genome association studies are currently very expensive. For
example, a researcher may need to test ∼400,000 polymorphic sites in
order to adequately “cover” all the polymorphic variation that exists in the
genome (Altshuler et al., 2005). Using traditional genotyping assays (∼$1.00
per assay), the total cost of such a study could reach as high as $80,000,000.
Fortunately, new technologies and strategies can be used to reduce these
costs. DNA microarrays can now be fabricated on chips to provide multiplex
genotyping assays that currently cost $800–1200 per chip (see, e.g., http://

www. aff ymetrix.com/index.affx and http://www.illumina.com/). Despite
the reduced costs of large-scale genotyping via chip technology, the cost of
genome-wide research is still very high (e.g., $1000 per chip × 2000 experi-
mental subjects = $2,000,000). It is hoped that cost will be further reduced
as chip technology continues to improve.
An alternative strategy for genome-wide association research is more

affordable, but also problematic. DNA can be “pooled” across individuals
within groups of interest (e.g., trait-positive versus trait-negative subjects).
Using genotyping technology such as a chip, allele frequencies can then be
estimated at each locus (Craig et al., 2005) and the resulting estimates can
be compared to identify variations related to the phenotype of interest. Th is
strategy dramatically reduces the cost of genotyping. However, pooled DNA
allele frequency estimates may be unreliable, and disequilibrium patterns
cannot be leveraged as they can for individual genotype studies (Salem,
Wessel & Schork, 2005).
Statistical methodology is another challenge in genome-wide research.
It is no small feat to yield compelling and meaningful results by the analysis
of hundreds of thousands of individual DNA sequence variations. In order

1/22/2008 9:36:19 PM
CRC_71777_Ch008.indd 195 1/22/2008 9:36:19 PM
CRC_71777_Ch008.indd 195