Assessment of estimation procedures for risk and onset hazard with dependent data

Analysis of the role of candidate genes as risk factors for age-dependent hereditary conditions often ignores the importance of dependence among sibships or other family clusters for age of onset. We examined the performance of several methods of survival analysis with dependent data using Collaborative Study on the Genetics of Alcoholism families as submitted for GAW11. Additionally, an arbitrary truncation of cluster size was performed to explore the potential impact of heterogeneity of family size on the resulting inferences concerning the role of candidate genes. Our results showed substantial differences in attribution of risk to candidate genes according to whether the method utilized allowed for dependence in onset age and according to whether the sample was truncated or arbitrarily stratified. Further work needs to be done to clarify the importance of properly accounting for dependent data in age-dependent phenotypes and in integrating these methods into widely used genetic analysis computer programs.