Data from the Framingham Heart Study were provided for Genetic Analysis Workshop 13. This paper summarizes six contributed papers that focused on the analysis of the longitudinal alcohol and/or cigarette use phenotypes available in these data, with the goal of detecting genes influencing these traits. For several of these contributions, the primary phenotype was maximum daily substance use reported over the longitudinal data. Others focused on a cross section by taking the daily use reported at a fixed time interval for all individuals, and others considered dichotomous phenotypes defined by thresholds of use. A variety of covariates were considered, including age (or year of birth) and sex. Most studies included unexposed individuals (those reporting no alcohol/cigarette use at all available assessments) in the linkage analyses, though one study compared results when defining the phenotype for unexposed individuals to be zero vs. defining the phenotype for unexposed individuals to be unknown. Linkage findings varied across studies. However, there was some concordance of evidence on chromosome 9 for alcohol traits, and of weaker evidence on chromosome 20 for cigarette traits. Analytical issues arose which may be crucial for genetic studies of substance use and dependence, including the choice of how to handle unexposed or substance-naive individuals, use of dichotomous vs. quantitative traits, and consideration of transformations and covariates.
Copyright 2003 Wiley-Liss, Inc.