Genetic epidemiology of cancers is on the verge of burgeoning. Recent advances in genetic technology allow epidemiologist to readily to obtain information on environmental risk factors as well as genetic markers in population-based studies involving families. In order to quantify relevant information from complex human pedigree data, development of statistical methods becomes more important that ever before. The broad objective of this proposal is to develop several statistical methods for genetic epidemiology of cancer, involving population-based cohort with pedigree information. These statistical methods will be applied to a population- based pedigree cohort of multi-ethnic groups in Hawaii. The results from the analysis will reveal patterns of familial aggregation of all major cancers. AS a long term objective, this project will provide leads to further research investigating roles of genes and environmental factors in cancer etiology. The proposal has seven aims. The first four aims are to establish four relevant statistical methods for the human pedigree analysis. The fifth aim is to evaluate case-control and diseased-proband ascertainment methods. The sixth aim is to investigate a two-stage ascertainment method that takes advantage of available cohort information. The seventh aim is to apply all these methods to the cohort data described above. The results from the analysis will suggest what type of cancers aggregates within families and how cancer is aggregated.