Xu S, Gessler D D
Department of Botany and Plant Sciences, University of California, Riverside 92521, USA.
Genet Res. 1998 Feb;71(1):73-83. doi: 10.1017/s0016672398003115.
We present a multipoint algorithm to map quantitative trait loci (QTLs) using families from outbred populations with a variable number of sibs. The algorithm uses information from all markers on a chromosome simultaneously to extract information of QTL segregation. A previous multipoint method (Kruglyak & Lander (1995) American Journal of Human Genetics 57, 439-454) extracts information using a hidden Markov model. However, this method is restricted to small families (< 10 sibs). We present an approximate hidden Markov model approach that can handle large sibships while retaining similar efficiency to the previous method. Computer simulations support the notion that data sampled from a small number of large families provide more power than data obtained from a large number of small families, under the constraint that the total number of individuals for the two schemes is the same. This is further reflected in simulations with variable family sizes, where variance in family size improves the statistical power of QTL detection relative to a constant size control.
我们提出了一种多点算法，用于利用来自远交群体且同胞数量可变的家系来定位数量性状基因座（QTL）。该算法同时使用染色体上所有标记的信息来提取QTL分离信息。之前的一种多点方法（Kruglyak和Lander（1995年），《美国人类遗传学杂志》57卷，439 - 454页）使用隐马尔可夫模型来提取信息。然而，这种方法仅限于小家系（<10个同胞）。我们提出了一种近似隐马尔可夫模型方法，该方法可以处理同胞数量多的家系，同时保持与之前方法相似的效率。计算机模拟支持这样一种观点，即在两种方案个体总数相同的约束下，从少量大家系中采样的数据比从大量小家系中获得的数据具有更大的功效。这在可变家系大小的模拟中进一步得到体现，其中家系大小的方差相对于恒定大小的对照提高了QTL检测的统计功效。
