The applicant's overall project is to develop computationally tractable and efficient methods for accurate statistical analysis of research data from genetic epidemiology and data from more general medical studies. Basic issues underlying much of the work is biased sampling and censored and missing data. He poses six major aims: 1. allele sharing genetic linkage analysis of censored survival traits and traits with variable age at onset; 2. to develop computationally tractable methods for analyzing interval censored data. Emphasis will be on proportional hazards, proportional odds, and accelerated failure time regression models; 3. methods for analyzing the influence of risk factors on the time between the initiating and subsequent event from doubly censored current status data; 4. methods for correcting ascertainment bias when pedigrees are enrolled on the basis of phenotypic characteristics of family members; 5. to develop a general approach for approximating boundary crossing probabilities in order to find more accurate p-values and confidence limits for nonparametric estimates of risk; 6. methods to account for the variety of censoring patterns and missing data that occur in family history data, with special attention paid to studies of childhood biomarkers for adult onset disease.
