|
 Genes
are responsible for many aspects of normal development. Mutations in genes
are thought to underlie the etiology and pathophysiology of a number of
psychiatric disorders. For decades, scientists have been trying to identify
these genes. As reviewed in the last column, linkage analyses have been
used to determine the approximate chromosomal location of genes involved
in many different disorders. The inheritance pattern of a disorder is
often studied in large pedigrees and is compared with the inheritance
pattern for other genes or markers whose locations in the human genome
are already known. These genetic markers are chosen either because of
previous information about the supposed location of the disease gene or
as part of a random genome scan. If a tight linkage is established between
the inheritance of the disease and the genetic marker, then the disease
gene must lie relatively nearby.
Linkage
analysis has been a powerful method for studying genetic disorders with
a mendelian mode of inheritance, such as cystic fibrosis or Huntington
disease. The analysis, however, depends on a set of assumptions. First,
the investigator must establish whether the mode of inheritance is autosomal
dominant, recessive, or another mode. In addition, it is essential to
define who has the disorder and who does not. The penetrance—the
chance that a subject who inherits the disease gene will express the disorder—must
also be estimated. The gene frequency in the population must be determined,
as well as whether there is genetic heterogeneity or not. The latter refers
to the probability that different genes can separately cause the disorder
under study. Finally, the rate of phenocopies must be determined, which
establishes whether a disorder is also caused by nongenetic factors. These
parameters are not known for most psychiatric disorders, and they must
be guessed at before traditional linkage analysis can be performed. The
chance of missing positive linkage is increased if any of these assumptions
are incorrect.
The
search for the gene for Tourette’s disorder provides an interesting
example of some of these pitfalls. Tourette’s disorder is a neurodevelopmental
disorder characterized by motor and vocal tics that begin in childhood.
The importance of genetic factors contributing to the expression of this
disorder was appreciated long ago. The expectations were high that a gene
would be localized for this disorder by linkage analysis. These hopes
relied on the relatively easy diagnosis of the phenotype and the apparent
autosomal dominant mode of inheritance. Nevertheless, after 10 years of
international collaboration in which dozens of pedigrees have been studied
with markers spanning more than 90% of the human genome, no linkage has
been established. In retrospect, parameters were assumed that were probably
incorrect. For example, investigators now question whether Tourette’s
disorder is an autosomal dominant disorder. Also, the assumption that
this disorder is caused by a single major gene is probably not true, and
several genes acting in conjunction are likely to be responsible for expression
of tics in the vulnerable child. Finally, separate and distinct genes
might contribute to the expression of the disorder in different pedigrees.
Thus the parameters that were initially specified were inaccurate and
linkage could not be established. In other psychiatric disorders such
as schizophrenia and bipolar disorder, the initially positive linkage
results that were obtained have not been replicated in later studies,
and researchers are similarly questioning the parameters that were assumed
for these disorders.
To
overcome some of these problems, several nonparametric approaches were
developed that do not rely on the a priori determination of these parameters.
There are 2 main types of approaches: association and allele-sharing.
In association studies, the distribution of different forms of a given
gene (called alleles) is studied in the general population. One compares
the frequency of the various alleles in a group of unrelated patients
with the frequency of these alleles in a group of normal controls. A higher
than normal frequency of a certain allele would suggest a role for this
allele in the expression of the disorder. This is how ApoE4, a specific
allele for a lipoprotein, was found to be related to Alzheimer disease.
More recently, an association was found between certain alleles of the
dopamine D4 receptor (D4DR) gene and the novelty-seeking trait, and between
the serotonin transporter gene and anxiety-related traits.
The
advantages of association studies are the simple statistics that are used
and the relative ease of obtaining a subject group as there is no need
to locate and interview relatives of the probands. On the other hand,
as linkage usually cannot be detected by association studies, genome scans
cannot be performed and alleles are chosen on the basis of some knowledge
about their function and presumed involvement in the disorder. Thus the
chance of finding the location of the gene for the disease is decreased.
Moreover, genes that are found to be associated with the disorder often
have a minor role in its etiology. For example, the serotonin transporter
gene explains as little as 4% of the variance of the anxiety-related traits
that were found to be associated with it.
Another
significant difficulty with association studies is that allelic differences
can stem from a number of factors. For example, differences in ethnic
origin between the affected and the control group are likely to cause
false-positive results. This is because allele frequencies are now known
to vary tremendously among different ethnic groups. This might be one
of the reasons for the difficulty in replicating the D4DR and serotonin
transporter findings that followed the initial results of a positive association.
To protect from this, a well-matched control group must be used. This
is often accomplished by analyzing the nonaffected relatives of the subjects
who are therefore of the same ethnic background.
The
second nonparametric technique is allele-sharing analysis. In this method,
related individuals with a disorder are studied. The sharing of the same
allele by all affected members of a pedigree is calculated and compared
with what is expected by chance. This can be done either in families with
2 affected siblings, called the sib-pair analysis, or in larger multigenerational
pedigrees, called affected pedigree analysis. Affected Pedigree Member
(APM) and Genehunter are examples of computer programs that use allele-sharing
methods that are currently used for the study of larger pedigrees. Once
again, the investigator looks for increased sharing of a particular allele.
Increased sharing of an allele argues for an association between the studied
marker and the disorder under study. This could imply linkage and help
localize the disease gene. One of the important advantages of allele-sharing
techniques is that no control group is required (Fig.
1).
In
contrast with association studies, allele-sharing methods can be used
for genome scans to detect genes with either a minor or major etiological
role. Allele-sharing methods are used in the study of many disorders with
complex inheritance. In diabetes, for example, sib-pair analysis was useful
in assessing the role played by genes within the HLA complex, as well
as finding other genes that are involved in its etiology. In the study
of schizophrenia, sib-pair analysis has proved useful in providing additional
support for the positive parametric linkage results of loci on chromosomes
6p and 22q that were reported. Researchers were also able to show increased
allele-sharing on the 6p chromosome in affected individuals from large
pedigrees with specific inherited forms of dyslexia. Recently, increased
allele-sharing of markers on chromosome 7q was found in sib pairs with
autism.
Although
very appealing, allele-sharing methods have one serious limitation. They
are much less powerful than parametric linkage analysis. Thus hundreds
of sib pairs and dozens of large pedigrees are needed to detect significant
findings. Researchers can overcome this problem by combining samples from
different populations. By doing so, however, they run the risk of genetic
heterogeneity interfering with the results, and thus they actually may
decrease the chance of finding increased allele-sharing. For this reason,
allele-sharing methods are often used as an adjunct to more traditional
parametric linkage analyses.
In
summary, newer methods of genetic analysis enhance the ability to investigate
the complex genetics of human behavior and development. Yet much effort
is still needed to find sufficient numbers of related subjects with a
disorder of interest and to search the entire genome for evidence of increased
allele-sharing. It remains a difficult task to localize genes for complex
disorders. Identifying the genes that are involved in these disorders
is the first step toward understanding the underlying molecular mechanism.
One must also appreciate that several genes are likely to be found to
contribute to a particular disorder. Understanding how these genes interact
with each other as well as with environmental factors poses the next challenge
to researchers.
top of page
|
