Univariate analysis of the "persistent tobacco use" phenotype is shown in the heavy dark line and univariate analysis for the "drinking severity" phenotype in the thin line. The bivariate analysis of the "persistent tobacco usage and drinking severity" is shown in the dark line. The analysis assumes a latent normally distributed variable with a threshold above which an individual is affected. The maximum LOD score for the bivariate analysis for the region is 3.4 compared to a maximum of 2.8 for drinking severity score and 1.7 for persistent tobacco use. Although there was insufficient power to exclude coincident linkage (p = 0.26), complete pleiotropy could be excluded (p = 0.02).