Biomedical Engineering Reference
In-Depth Information
in a logistic regression model [
12
,
17
-
19
]. Among these methods based on PCA,
EIGENSTRAT is the most widely used in GWAS [
20
-
22
]. Different studies pro-
posed alternative methods or integration to PCA [
23
-
25
]. These different methods
can use genotype information from a whole genome set of SNPs [
16
,
19
-
22
] or from
a set of selected informative markers (AIMs) [
14
,
15
,
26
,
27
]. AIMs are defined as
markers that show larger allele frequency differences between ancestral populations.
There are currently several existing AIMs panels [
5
] that can be genotyped to esti-
mate genetic ancestry, but all panels are not equivalent and may also not be robust
enough in genetic studies with populations of unknown origins. In summary, there
is currently no definitive consensus about a common strategy to adjust for PS. The
goal of scientists studying genetic associations with a complex disease in samples
that may include different (ethnic) groups is to examine the population ethnic back-
ground and to correct for stratification to find the genetic variants really relevant for
the disease by avoiding false-positive associations. This chapter shows some exam-
ples looking into PS applying different methods to find the most efficient strategy
to correct the observed findings based on the detected genetic variance in two as-
sociation studies on schizophrenia. We assessed PS in the UCI sample composed
by about 200 subjects and in the available CATIE-NIMH sample [
28
], made up of
1,492 individuals. Both samples are composed of different ethnic groups.
9.2
Materials and Methods
9.2.1
Subject Collection and Genotyping
We performed case-control association study in two different samples with different
sizes: the UCI sample and the CATIE-NIMH sample.
9.2.1.1
UCI Sample
We studied 107 patients with chronic schizophrenia who have been recruited at the
University of California at Irvine (UCI). Schizophrenia was diagnosed according
to the criteria of the diagnostic and statistical manual of mental disorders (DSM)
IV using structured clinical interviews. In addition, 91 healthy controls matched
for age and gender have also been recruited at the UCI without any mental disease
according to DSM IV. Written informed consent was obtained and blood samples
collected from each patient and control. Eighty-four per cent of the individuals were
Caucasians, 4.5% were Asians and the remaining 11.5% were African-Americans.
Ethnic status was assessed by the clinician according to information about the place
of birth of each individual together with that of their parents and grandparents,
as well as their mother language. Genomic DNA was extracted from blood by a
standard procedure. Genotyping was performed at the Department of Science and
Search WWH ::
Custom Search