Tan Qihua, Zhao J H, Iachine I, Hjelmborg J, Vach W, Vaupel J W, Christensen K, Kruse T A
Department of Clinical Biochemistry and Genetics, KKA, Odense University Hospital, Odense, Denmark.
Genet Epidemiol. 2004 Apr;26(3):245-53. doi: 10.1002/gepi.10304.
This report investigates the power issue in applying the non-parametric linkage analysis of affected sib-pairs (ASP) [Kruglyak and Lander, 1995: Am J Hum Genet 57:439-454] to localize genes that contribute to human longevity using long-lived sib-pairs. Data were simulated by introducing a recently developed statistical model for measuring marker-longevity associations [Yashin et al., 1999: Am J Hum Genet 65:1178-1193], enabling direct power comparison between linkage and association approaches. The non-parametric linkage (NPL) scores estimated in the region harboring the causal allele are evaluated to assess the statistical power for different genetic (allele frequency and risk) and heterogeneity parameters under various sampling schemes (age-cut and sample size). Based on the genotype-specific survival function, we derived a heritability calculation as an overall measurement for the effect of causal genes with different parameter settings so that the power can be compared for different modes (dominant, recessive) of inheritance. Our results show that the ASP approach is a powerful tool in mapping very strong effect genes, both dominant and recessive. To map a rare dominant genetic variation that reduces hazard of death by half, a large sample (above 600 pairs) with at least one extremely long-lived (over age 99) sib in each pair is needed. Again, with large sample size and high age cut-off, the method is able to localize recessive genes with relatively small effects, but the power is very limited in case of a dominant effect. Although the power issue may depend heavily on the true genetic nature in maintaining survival, our study suggests that results from small-scale sib-pair investigations should be referred with caution, given the complexity of human longevity.
本报告探讨了应用受累同胞对（ASP）的非参数连锁分析[克鲁格利亚克和兰德，1995年：《美国人类遗传学杂志》57:439 - 454]，通过长寿同胞对定位影响人类长寿的基因时的效能问题。通过引入一种最近开发的用于测量标记 - 长寿关联的统计模型[亚申等人，1999年：《美国人类遗传学杂志》65:1178 - 1193]来模拟数据，从而能够直接比较连锁和关联方法的效能。对在含有因果等位基因的区域中估计的非参数连锁（NPL）分数进行评估，以评估在各种抽样方案（年龄截断和样本量）下不同遗传（等位基因频率和风险）及异质性参数的统计效能。基于基因型特异性生存函数，我们推导了遗传力计算方法，作为对具有不同参数设置的因果基因效应的总体度量，以便能够比较不同遗传模式（显性、隐性）的效能。我们的结果表明，ASP方法是定位显性和隐性强效应基因的有力工具。要定位一个使死亡风险降低一半的罕见显性遗传变异，需要一个大样本（超过600对），且每对中至少有一个极其长寿（超过99岁）的同胞。同样，在大样本量和高年龄截断的情况下，该方法能够定位效应相对较小的隐性基因，但对于显性效应，效能非常有限。尽管效能问题可能在很大程度上取决于维持生存的真实遗传性质，但鉴于人类长寿的复杂性，我们的研究表明，小规模同胞对研究的结果应谨慎参考。
