Taken together, epidemiological, cytogenetic, linkage, and association studies in psychiatric genetics to date paint a picture of highly complex genetic influences on psychiatric disorders. As Kendler (2005) pointed out, the phrase "a gene for. . . " will very likely not apply to psychiatric genetics. And, as Kendler went on to note, "The impact of individual genes on risk for psychiatric illness is small, often nonspecific, and embedded in a complex causal pathway" (Kendler 2005, p. 1243). The field may need to adopt strategies that are better adapted to the most likely disease models. In the following we will mention some of the proposed strategies to address this issue.

First, we may need to reconsider the way we define cases or the phenotype of interest. Our current classification schemes are not likely directly reflective of the underlying biology—and thus the genetic determinants—of psychiatric disease. The currently used diagnostic algorithms (DSM-IV-TR [American Psychiatric

Association 2000] and ICD-10 [World Health Organization 1992]) group diagnoses by symptoms and clinical course, characteristics that may reflect not a common biology but rather a final common pathway of several different pathophysiological disturbances. That recognition has led some to propose the use of intermediate phenotypes—including neurophysiological, biochemical, cognitive, and endocrine measures (Gottesman and Gould 2003; Hasler et al. 2004)—in psychiatric genetic studies in order to create biologically more homogeneous subgroups of patients and thus to increase the power to detect case-control associations. Another important consideration is that a number of symptoms are common to several different DSM-IV-TR diagnoses, and the genetic susceptibility to develop these symptoms may be common across disorders. In fact, there is evidence that the major psychiatric disorders may share susceptibility genes. A series of linkage peaks and candidate gene associations overlaps between bipolar disorder and schizophrenia, for example (Craddock et al. 2006), and the cytogenetic disruption of DISC1 leads to a variety of severe psychiatric disorders, ranging from recurrent unipolar disorder to schizophrenia (St. Clair et al. 1990).

Second, environmental measures should be included more consistently in genetic studies, including whole-genome association studies. Epidemiological (Kendler 1995) as well as molecular genetic studies have now repeatedly demonstrated the importance of gene-environment interactions in psychiatric disease (Caspi and Moffitt 2006). Genetic effects may be obscured by unmeasured environmental effects, so that different environmental exposures in replication samples may be one source of nonreplication of genetic association.

Finally, one should not forget that SNPs are just the most common and convenient type of genetic variant. Other types of variation, such as CNVs, may be equally important (Redon et al. 2006; Sebat et al. 2004). Newer versions of whole-genome arrays now try to cover most copy number variations known to date, and association with these may lead to surprising findings.

0 0

Post a comment