Department of Biological Statistics and Computational Biology, Cornell University, Ithaca, New York, USA.
Genet Epidemiol. 2011 Jul;35(5):371-80. doi: 10.1002/gepi.20585. Epub 2011 Apr 25.
Multi-symptom diseases without a consistent continuous measurement of severity may be best understood with a categorical interpretation. In this paper, we present LOCate v.2, a fast, exact algorithm for linkage analysis of all types of categorical traits, both ordinal and nominal. Our method is able to incorporate missing data and analyze complex genealogical structure, including inbreeding loops. LOCate v.2 computes exact likelihoods efficiently through an elimination algorithm, similar to that used by Superlink for binary traits. We compare LOCate v.2 to LOT and QTLlink, two existing methods of linkage analysis for ordinal traits. We find that LOCate v.2 outperforms both methods when used to analyze simulated nominal traits. In addition, LOCate v.2 performs as well as QTLlink on simulated ordinal traits, and better than LOT due to the necessity of cutting large pedigrees for analysis in LOT. To demonstrate the versatility of LOCate v.2, we conduct an ordinal and nominal linkage analysis of ventricular arrhythmias in a large, inbred pedigree of German Shepherd dogs. We find that a trichotomous ordinal or nominal interpretation strengthens the evidence in favor of linkage to a region on chromosome 6, and provides new evidence of linkage to a region on chromosome 11. LOCate v.2 is a unified, fast, and robust method for linkage analysis of ordinal and nominal traits which will be valuable to researchers interested in investigating any type of categorical trait.
多症状疾病没有一致的连续严重程度测量，可能最好用分类解释来理解。在本文中，我们提出了 LOCate v.2，这是一种用于分析所有类型分类特征（有序和名义）的快速、精确的连锁分析算法。我们的方法能够处理缺失数据，并分析复杂的系谱结构，包括近亲繁殖环。LOCate v.2 通过消除算法高效计算精确似然，类似于 Superlink 用于二进制特征的算法。我们将 LOCate v.2 与 LOT 和 QTLlink 进行比较，这是两种现有的有序特征连锁分析方法。我们发现，在分析模拟名义特征时，LOCate v.2 优于这两种方法。此外，LOCate v.2 在分析模拟有序特征时与 QTLlink 表现相当，并且优于 LOT，因为 LOT 必须对大型系谱进行切割以进行分析。为了展示 LOCate v.2 的多功能性，我们对德国牧羊犬的一个大型近亲系谱中的室性心律失常进行了有序和名义连锁分析。我们发现，三分类有序或名义解释加强了对 6 号染色体区域连锁的证据，并提供了对 11 号染色体区域连锁的新证据。LOCate v.2 是一种用于分析有序和名义特征的统一、快速和强大的连锁分析方法，对研究任何类型的分类特征的研究人员将具有重要价值。
