REU: Correlated Evolution of Continuous Traits: A Simulation Study

比较研究是现代生物学的重要组成部分。通常，几个物种在两个或两个以上的特征上进行比较，例如大脑大小和身体大小，四肢长度和跑步速度，或者牙齿和饮食。性状之间的显著关联通常是通过应用标准的统计检验来推断的。不幸的是，标准测试不适合用于比较研究。这是因为物种并不代表统计上独立的数据点。最近，已经提出了几种方法，它们考虑了物种价值的非独立性，从而允许对连续性状之间的进化关系进行适当的统计检验。这些方法明确利用了物种之间已知的进化(系统发育)关系。这项拟议的研究通过确定它们在分析具有已知属性的计算机模拟数据时的表现，比较了几种此类方法的统计效率。

Comparative studies form a major part of modern biology. Typically, several species are compared with regard to two or more traits, such as brain size and body size, limb length and running speed, or dentition and diet. Significant associations between traits generally are inferred by applying standard statistical tests. Unfortunately, standard tests are not appropriate for comparative studies. This is so because species do not represent statistically independent data points. Recently, several methodologies have been proposed which take the non-independence of species values into account and hence allow appropriate statistical tests for evolutionary relationships between continuous traits. These methods make explicit use of the known evolutionary (phylogenetic) relationships among species. The proposed research compares the statistical efficacy of several such methods by determining how well they perform in the analysis of computer-simulated data with known properties.