Sib-pair linkage analyses of alcoholism: Dichotomous and quantitative measures

We hypothesized that a quantitative alcoholism trait would have greater pow er than the Collaborative Study on the Genetics of Alcoholism (COGA) dichot omous alcoholism traits, ALDX1 and ALDX2, to detect putative alcoholism loc i. To test this, we performed nonparametric sib-pair linkage analysis to sc reen 285 polymorphic autosomal markers for evidence of linkage to ALDX1, AL DX2, and a quantitative trait, QUANT, defined from the 11 COGA latent class variables. We also examined the effects on the analyses of including covar iates (sex, age, and pack-years of smoking) and of transforming QUANT (log and square root). ALDX1 and ALDX2 showed the greatest evidence for linkage to markers on chromosome 1, by both the affected sib-pair and the Haseman-E lston tests. Regions of interest were also identified on chromosomes 4, 8, 16, and 17. QUANT showed little evidence for linkage to any chromosomal reg ion, having no more significant results than were expected by chance. Inclu ding covariates or transforming QUANT had little effect on the analyses. A quantitative trait based on all 37 latent class variables, with each variab le appropriately weighted, may have had more power than QUANT to detect gen omic regions of relevance to alcoholism. (C) 1999 Wiley-Liss, Inc.