PHENOTYPE DELINEATION AND GENETIC MODELING OF AUTISM

点击翻译按钮获取中文摘要

Investigating the genetics of autism presents particular challenges (due to features such as limited vertical transmission, and small family size), over and above the usual difficulties of modeling complex and heterogeneous disorders. Therefore, clarification of the phenotypic expression of the genetic liability underlying autism is a critical first step. A method for exploring the phenotypic boundaries of autism and identifying relatives of autistic probands affected with a broad or multidimensional phenotype will be developed using data from the Baltimore Family Study of Autism. The method will then be replicated, in a quasi- validation step, with data from two additional sets of autism families (one simplex and one multiplex) collected in Iowa with the use of instruments and measures similar or identical to those used in the Baltimore study. The method for identifying the affected will be tested for its utility in improving the power of genetic modeling, including segregation and linkage analyses. The segregation analyses will be carried out on the two data sets described above. The linkage studies will be carried out with simulated data; both trait and marker genotypes will be simulated for pedigrees with the phenotypic characteristics of the families we have actually observed. Then both likelihood-based and non-parametric linkage analysis will be carried out to determine statistical power under various models and misclassification schemes. More generally, I hope this work will contribute to the development of methods for correctly classifying affected relatives in families expressing other complex disorders with ambiguous phenotypic expression, and to the search for genes for other neuropsychiatric disorders.