Adaptation of New Statistical Ideas for Medicine

我们的MERIT奖工作将继续有两个主要组成部分：参与。 研究项目，如NHBLI的FEHGAS研究，并开发新的统计方法 用于分析大型复杂数据集。这些努力是相辅相成的， 提出哪些统计方法是mofit需要的，也作为新方法的测试案例。 例如，FEHGAS研究试图从SNP数据预测高血压发病年龄（以及 年龄和性别等背景变量）。有550，000个SNP可用于预测，其中大部分 这将被证明是无用的，使问题的规模更具挑战性，比在 表达微阵列的情况。埃夫隆计划将经验贝叶斯方法从他最近的 本文在此背景下，希望克服通常较弱的预测能力所造成的困难 单个SNPs。Olshen计划扩展CART（计算机辅助锡永树）和bootstrap 选择有前景的预测SNP组的方法。 大规模的显著性测试，例如在微阵列癌症研究中选择“显著”基因， 已成为统计学发展的热点。然而，在文献中，适当实施的关键问题仍然模糊不清：适当的零假设的选择; 比较组（所有550，000个SNP应该一起测试还是通过染色体单独测试？）;和效果 的相关性。我们在回答这些cf问题方面取得了一些进展，如《进展》中所述。 次报告.我们的持续努力是方法实施和理论发展的结合。 相关性对标准统计技术有特别的影响。是一组微阵列 彼此独立？“一项涉及20，000个基因的研究有其有效样本， 由于严重的基因相关性，大小减少到约17。我们目前正在开发诊断 在大量数据集中发现相关性困难的方法，以及评估它们对假设检验的影响的方法， 估计和预测。一个20,000个基因的微阵列研究产生了20000000个相关性，这听起来 对于实际的洞察力来说大得令人压抑。但我们在经验Ba 5 'es近似上取得了进展 它将相关性、效应总结在一个数字中，适合于简单分析。 20世纪的生物统计学应用在本质上是压倒性的频率论。纯粹：频率主义， 然而，对于分析现代生物医学产生的大型复杂数据集来说， 设备，其中必须考虑数千个参数和数百万个数据点的关系 一起我们正在继续开发经验贝叶斯方法，使贝叶斯的想法， 承担多重推理的问题，而不需要特定的先验分布。 一个长期的项目是了解经验贝叶斯信息在医学研究中的积累速度。 错误发现率是对基因（或SNP，br体素） 是“空”，给出了观察到的数据。我们需要观察多少个受试者和多少个基因才能按顺序进行 得到一个精确的经验贝叶斯估计的后验概率？ 在我们自己版本的摩尔定律中，生物医学数据集的大小每增加一个数量级， 自20世纪90年代以来的几年。新兴技术（平铺阵列、珠阵列、适体芯片、甲基化阵列、 外显子芯片和各种新的成像设备）承诺进一步增加， 设备和统计方法。我们的长期目标是提供算法和理论 适当的数据收集和生物医学要求。

Our MERIT award work will continue to have two main components: involvement in .specific biomedical reseai-ch projects sucli as NHBLI's FEHGAS study, and development of new statistical methods appropriate for the analysis of large, complex data sets. These efforts are complementary, with the speciflc projects ¿suggesting which statistical rnethods are mofit needed, and also serving as test cases for new methodology. The FEHGAS study, for exarhple,- seeks to predict age of onset of hypertiension from SNP data (and background variables such as age and gender). There are 550,000 SNPs available for prediction, most of which will turn out to be useless, making the problem an ijrder of magnitude more challenging, than in expression microarray situations. Efron plans to extend the empirical Bayes liiethodology from his recent paper to this context, hopefully overcoming the difficulties caused by the usually weak predictive power of individual SNPs. Olshen plans to extend CART (Computer Assisted Regre.s.sion Trees) and bootstrap methodology to the selection of groups of promising predictive SNPs. Large-scale significance testing, for instance selecting 'significant' genes in a microarray cancer study, has become an area of iiitense statistical development. Nevertheless, crucial questions of appropriate implementation remain vague in the literature: the choice of an appropriate null hypothesis; the selection of a comparison set (Should all 550,000 SNPs be tested together or sepai-ately by chromosome?); and the effects of correlation. We have made some headway in answering thescf questions, as described in the Progress Report. Our continuing efforts are a combination of methodological implementation and theoretical development. Correlatiion can have particularly dra.stic effects on staiidard statistical techniques. Iii "Are a .set of microarrays independent of each other?" it is shovyn that a study involving 20,000 genes has its effective sample size reduced to about 17 because of severe gene-wise correlation. We are currently developing diagnostic methods to spot correlation difficulties in massive data sets, and to assess their effects on hypothesis tests, estimates, and predictions. A 20,000 gene microarray study produces 200,000,000 correlations, which sounds oppressively large for practical insight. But we are making progress on an empirical Ba5'es approximation that summarizes correlation, effects in a single number, suitable for simple analysis. Twentieth Centiiry biostatistical applications were overwhelmingly frequentist in nature. Pure: frequentism, though, becomfSi impra<;tical for analyzing the large, complex data sets produced by modem biomedical devices, where the relationships of thousands of parameters and millions of data points have to be considered together. We are continuing to develop empirical Bayes methods that allow Bayesian ideas to be brought to bear on questions of multiple inference, without requiring specific prior distributions from the .scientist. A long-term project is to understand how quickly empirical Bayes information accrues in a medical study. A False Discovery Rate is an estimate of the Bayes posterior probabiUty that a gene (or a SNP, br a voxel) is 'null', given the observed data. How many subjects and how many genes do we need to observe in order to get an acciurate empirical Bayes estiinate of the posterior probability? hi our own version of Moore's law, biomedical data sets have increased an order of magnitude in size every few years since the 1990s. Emerging technologies (tiling arrays, bead arrays, aptamer chips, methylation arrays, exon chips, and a variety of new imaging devices) promise further increases, taxing both computational equipment and statistical inethodology. Our long-term MERIT goal is to provide algorithms and theory appropriate tp massive-data biomedical requirements.