Genomewide linkage searches, which were taken up so enthusiastically in psychiatric genetics only a few years ago are now greeted with increasin skepticism. The high hopes have not been realized, and claimed linkages have been discredited. However, many searches have been initiated, and a great deal of apparently negative data now exists. Investigators searching for linkage to various diseases have not considered what they will do with their data if no single, overwhelming linkage appears. Are these data useless? Are the important psychiatric disorders, albeit highly heritable, only polygenic and without megaphenic genes? Despite recent disappointments, this has not been proved. Although the discover of a single gene that explains all affection in high density pedigrees appears less and less likely, this is not the only possibility. If ther are a half dozen major loci that substantially influence risk of schizophrenia in small proportions of high density pedigrees, they probably cannot be detected by the analytic methods heretofore employed, but theoretically, it may be possible to do so. We have recently developed new methods specially designed to disclose multiple megaphenic genes influencing complex traits. These methods wer developed employing simulated data, and we now propose to apply them to a large set of linkage data on schizophrenia collected by our group. We have investigated a few promising chromosomal regions for linkage to schizophrenia, and results are reviewed in this proposal. In one region in particular, suggestive but inconclusive single marker results were integrated into a coherent regional map that increased the evidence of linkage considerably. Since only two-thirds of the presently available data were employed in the original investigation of this region, we have a fortuitous opportunity for cross validation of our finding. Much work remains. Whether megaphenic genes exist for schizophrenia is unknown, but if they do, these new methods may give us a chance of finding them. Moreover, regardless of our success or failure on a specific sample, practical demonstration of our methods may be useful to investigators of a wide range of other complex traits.
Showing the most recent 10 out of 19 publications
