The impact of redefining affection status for alcoholism on affected-sib-pair analysis

The analysis of a complex disease such as alcohol dependence requires a more precise definition of affection status. Collaborative Study on the Genetics of Alcoholism (COGA) provided a variety of qualitative and quantitative measures as well as genotype information, in addition to two criteria of affection status. To identify two groups of phenotypically "more homogeneous" individuals among alcoholics (COGA criterion), we redefined affection status by using cluster analysis and classification and regression tree, incorporating some important covariates such as event related potentials, monoamine oxidase B activity, status of smoking, age of onset, three variables of personality assessed with the Tridimensional Personality Questionnaire and three latent class variables. With redefined affection status, we repeated nonparametric analysis by three sib pair analysis programs (SIBPAL, SIBPAIR, and BETA) using nine candidate DNA markers identified by Reich et al. [1998] and Long et al. [1998]. The goals of our analysis are 1) to confirm previous results for these nine markers with redefined affection status and 2) to compare the performance from these three programs.