Center of System Biomedical Sciences, University of Shanghai for Science and Technology, Shanghai, People's Republic of China.
PLoS One. 2010 Nov 8;5(11):e13857. doi: 10.1371/journal.pone.0013857.
Technology advances have promoted gene-based sequencing studies with the aim of identifying rare mutations responsible for complex diseases. A complication in these types of association studies is that the vast majority of non-synonymous mutations are believed to be neutral to phenotypes. It is thus critical to distinguish potential causative variants from neutral variation before performing association tests. In this study, we used existing predicting algorithms to predict functional amino acid substitutions, and incorporated that information into association tests. Using simulations, we comprehensively studied the effects of several influential factors, including the sensitivity and specificity of functional variant predictions, number of variants, and proportion of causative variants, on the performance of association tests. Our results showed that incorporating information regarding functional variants obtained from existing prediction algorithms improves statistical power under certain conditions, particularly when the proportion of causative variants is moderate. The application of the proposed tests to a real sequencing study confirms our conclusions. Our work may help investigators who are planning to pursue gene-based sequencing studies.
技术进步促进了基于基因的测序研究，旨在识别导致复杂疾病的罕见突变。在这些类型的关联研究中，一个复杂的问题是，绝大多数非同义突变被认为对表型是中性的。因此，在进行关联测试之前，区分潜在的致病变异与中性变异至关重要。在这项研究中，我们使用现有的预测算法来预测功能氨基酸替换，并将该信息纳入关联测试中。通过模拟，我们全面研究了几种有影响因素的影响，包括功能变体预测的灵敏度和特异性、变体数量和致病变体的比例，对关联测试的性能的影响。我们的结果表明，在某些情况下，特别是当致病变体的比例适中时，将来自现有预测算法的功能变体信息纳入其中可以提高统计功效。拟议测试在真实测序研究中的应用证实了我们的结论。我们的工作可能有助于计划进行基于基因的测序研究的研究人员。
