Date of Award


Degree Type


Degree Name

Doctor of Philosophy


Faculty of Communications, Health and Science

First Advisor

Associate Professor L. Kalaydjieva

Second Advisor

Associate Professor J. Hallmayer


Autism is a severe developmental disorder that was first described by Kanner in 1943. It is characterised by four major criteria: marked social deficits, delay in language development, a restricted range of stereotyped repetitive behaviours and onset of the disease within the first three years of life. The last decade of research has provided support for a strong genetic basis in the aetiology of autism. Firstly, a number of genetic conditions, such as fragile X syndrome, chromosome 15 anomalies and tuberous sclerosis, have been associated with autism. Secondly, family studies have demonstrated that the recurrence risk for autism among siblings of autistic children is 50-200 times higher compared to the risk of autism in the general population (4-5 in I 0 000 births). Thirdly, twin studies show a significantly higher concordance rate among monozygotic twins compared to dizygotic twins. To explore the theory of a biological basis of autism, molecular genetic strategies were employed in this study. Two collaborative groups (Stanford University in the U.S.A. and Edith Cowan University/Graylands Hospital in Perth, Australia) scanned the entire human genome for autism susceptibility genes in 90 American multiplex families, making this the largest genome screen in autism to date. Candidate regions were also run in an additional group of 41 Australian multiplex families. One hundred of the total of 519 markers were analysed as part of this PhD project. Since autism is most likely a complex disorder involving interacting genes with unknown modes of inheritance, non-parametric linkage analysis utilising the multipoint sib-pair approach is the best method presently available for identifying these susceptibility genes. Using this method and given the power of the study to reliably detect genes of moderate effect, findings from this PhD project and the overall genome scan provide substantial evidence for the absence of such genes in autism. Regions that may, however, contain these disease-predisposing loci were identified. These include regions on chromosomes Ip and 22q. Results from other smaller genome scanning studies also indicate these two candidate regions as possibly containing loci involved in the aetiology of autism. Our results could also not exclude chromosomes 7q, 9q, 10q, 11p and 17, regions that have previously been suggested by other studies to contain susceptibility loci. Overall, our data were most consistent with a model of ≥15 susceptibility loci, each with a small effect on autism. These results are important for future studies. In particular, they have shown the requirement for a large sample size to detect genes of small effect contributing to a disorder. Whilst it may be difficult for individual research groups to recruit an adequate sample size, collaboration with other groups may provide sufficient power to detect these susceptibility loci. By combining our results with those from other linkage studies on autism, we may also be able to eliminate some genetic models for this disease and provide an insight into the boundaries of the autistic phenotype. With advances in the Human Genome Project and the development of rigorous statistical approaches to the analysis of multifactorial diseases, the molecular characterisation of genes that have a small effect on the overall aetiology of complex polygenic disorders, such as autism, could soon be feasible.