Jørstad Tommy S, Langaas Mette, Bones Atle M
Department of Biology, Norwegian University of Science and Technology, NO-7491 Trondheim, Norway.
Trends Plant Sci. 2007 Feb;12(2):46-50. doi: 10.1016/j.tplants.2007.01.001. Epub 2007 Jan 16.
DNA microarray experiments have become a widely used tool for studying gene expression. An important, but difficult, part of these experiments is deciding on the appropriate number of biological replicates to use. Often, researchers will want a number of replicates that give sufficient power to recognize regulated genes while controlling the false discovery rate (FDR) at an acceptable level. Recent advances in statistical methodology can now help to resolve this issue. Before using such methods it is helpful to understand the reasoning behind them. In this Research Focus article we explain, in an intuitive way, the effect sample size has on the FDR and power, and then briefly survey some recently proposed methods in this field of research and provide an example of use.
DNA微阵列实验已成为研究基因表达的一种广泛使用的工具。这些实验中一个重要但困难的部分是确定要使用的合适的生物学重复次数。通常，研究人员会希望有一定数量的重复，以便在将错误发现率（FDR）控制在可接受水平的同时，有足够的能力识别受调控的基因。统计方法学的最新进展现在有助于解决这个问题。在使用这些方法之前，了解其背后的推理是有帮助的。在这篇研究聚焦文章中，我们以直观的方式解释样本量对FDR和检验效能的影响，然后简要概述该研究领域最近提出的一些方法，并提供一个使用示例。
