Huntington disease: Linkage analysis with age‐of‐onset corrections

Abstract

Eleven extended pedigrees segregating autosomal dominant Huntington disease (HD) were analyzed for linkage with 15 gene markers. A straight‐line age correction (STLN), fitted roughly to published age‐of‐onset data but excluding the oldest and youngest ages, was used. The results failed to confirm a positive lod score reported by Brackenridge et al [1978] between HD and haptoglobin (Hp); our lod scores, taken alone, would rule out this linkage. No lod score in our study exceeded 0.40. Three informative markers, adenosine deaminase (ADA), Hp, and MNSs, were used to compare results obtained by employing different age corrections during lod score calculation. In addition to STLN, a least‐squares line (LSQR) and a cumulative normal curve (CUMN) were used. CUMN was assumed to be the most accurate, since it fitted the data most closely. For comparison, two methods involving no age correction were also examined: one in which all unaffected persons were uniformly assigned a 50% penetrance (NONE) and one in which all younger at‐risk persons were removed from the analysis (REMV). NONE and REMV distorted lod scores, in some cases strikingly. More importantly, they also sometimes biased the estimate of the recombination fraction. In contrast, STLN and LSQR gave results generally very similar to those found using CUMN. Thus, it was concluded that the exact shape of the age correction used is not critical to linkage analyses, but that failure to use any correction may lead to unacceptable results.