生态学和进化的经典话题,面对快速的全球环境变化,数量性状的表型微进化重新受到关注。然而,对于植物来说,由于标准指标的有限使用,合成受到了阻碍,这使得经验信息难以系统化。在这里,我们展示了将荟萃分析工具纳入微观进化率审查的优势。我们对已知时间段内数量性状的微观进化的植物文献进行了系统的调查,基于scopus数据库。我们通过标准平均差异来量化变化量,并开发一组效应大小来分析此类数据。我们表明,将荟萃分析工具应用于系统的文献综述可以比直接计算微观进化率提取更大的信息量。我们还提出了派生的荟萃分析效应大小(h,LG和LR)适合研究进化模式,第一个类似于haldanes,第二个和第三个允许应用现有的分析框架来推断进化机制。这种新颖的方法学发展适用于任何分类单元的微进化研究。为了试验它,我们建立了一个开放获取数据库,该数据库包含来自128项研究的152种被子植物物种的1,711种微观进化率,这些研究记录了已知时间(<260年)的环境新颖性后数量性状的种群变化.建议的度量标准的性能(h,LG和LR)与先前存在的类似,同时,它们带来了较低的估计偏差和较高的荟萃分析典型的可用观察数的优势。
A classic topic in ecology and evolution, phenotypic microevolution of quantitative traits has received renewed attention in the face of rapid global environmental change. However, for plants, synthesis has been hampered by the limited use of standard metrics, which makes it difficult to systematize empirical information. Here we demonstrate the advantages of incorporating meta-analysis tools to the
review of microevolutionary rates. We perform a systematic survey of the plant literature on microevolution of quantitative traits over known periods of time, based on the scopus database. We quantify the amount of change by standard mean difference and develop a set of effect sizes to analyze such data. We show that applying meta-analysis tools to a systematic literature
review allows the extraction of a much larger volume of information than directly calculating microevolutionary rates. We also propose derived meta-analysis effect sizes (h, LG and LR) which are appropriate for the study of evolutionary patterns, the first being similar to haldanes, the second and third allowing the application of a preexisting analytical framework for the inference of evolutionary mechanisms. This novel methodological development is applicable to the study of microevolution in any taxa. To pilot test it, we built an open-access database of 1,711 microevolutionary rates of 152 angiosperm species from 128 studies documenting population changes in quantitative traits following an environmental novelty with a known elapsed time (<260 years). The performance of the metrics proposed (h, LG and LR) is similar to that of preexisting ones, and at the same time they bring the advantages of lower estimation bias and higher number of usable observations typical of meta-analysis.