measuring smoking phenotype were selected from the Framingham Heart Study data. As indicated earlier, we only used the smoking information obtained from the exam conducted during 1970–1971, instead of using the maximum number of cigarettes smoked per day across multiple exams over many years. Epidemiological studies have indicated that there has been a steady and dramatic decline of 40% in the prevalence of cigarette smoking by people 18 years or older in the US from 1965 to 1990 [13]. This was also true in the Framingham Heart Study data (data not shown). Therefore, using smoking information obtained from multiple exams over a long period of time may affect estimation of the genetic and environmental parameters, and thus eventually the linkage analysis results. Nicotine dependence is a complex trait with strong genetic and environmental influences. Many years of genetic epidemiological studies have documented that smoking behavior is determined by multiple genetic and environmental factors, and interaction among these factors. Strong evidence for linkage of smoking behavior to chromosome 5q has been reported from an analysis of the COGA data [5]. The linkage to smoking behavior on chromosome 5 was also reported by another study with a different linkage analysis method but
