Analysis of Oligonucleotide Gene Expression Array Data

描述(由申请人提供)：该项目的长期目标是开发统计方法和计算机程序，以有效地分析微阵列数据。这一发展将建立在我们以前关于基于模型的方法进行此类分析的工作的基础上，但将解决几个问题，以便显着提高绩效。我们的具体目标如下。 目的1.建立完全匹配和错配寡核苷酸探针对mRNA浓度变化反应的完全模拟方法，并利用这种模型来构建改进的表达指数的估计。通过这样做，我们希望将估计的线性扩展到低信号范围，在估计过程中必须考虑交叉杂交效应。我们还将研究基于完全匹配设计的寡核苷酸表达阵列的可行性。 目的2.发展数组到数组的归一化方法，适用于期望对应的样本表现出非常不同的表达模式的情况。 目的3.建立用于酵母突变株混合研究的标签探针阵列的统计分析方法。 目的4.开发基于上述方法的计算机程序模块，并将其集成到软件dChip中，以扩展其用于基于模型的寡核苷酸阵列数据分析的功能。我们还将与实验小组密切合作，分析真实数据，以确保我们的研究结果与用户群体的需求相关。我们的目标不仅是通过目标1直接产生的方法来改进DChip，而且还将有选择地实施并将我们团队和我们的合作者开发的其他有用的分析方法纳入DChip框架。 目的5.与基于R的生物信息学联盟合作，促进将我们的方法转移到该平台，并通过R函数调用扩展dChip的功能。

DESCRIPTION (provided by applicant): The long-term objective of this project is to develop statistical methods and computer programs for effective analysis of microarray data. The development will build on our previous work on the model-based approach to such analyses but will address several issues in order to significantly improve the performance. Our specific aims are as follows. Aim 1. To develop methods for the full modeling of the responses of perfect match and mismatch oligonucleotide probes to changes in mRNA concentration, and to exploit such models to construct estimates of improved expression indexes. By doing so, we hope to extend the linearity of the estimates to the low signal range where cross-hybridization effects must be accounted for in the estimation process. We will also investigate the feasibility of oligonucleotide expression arrays based on perfect match-only design. Aim 2. To develop array-to-array normalization methods applicable to situations when the corresponding samples are expected to exhibit very different expression patterns. Aim 3. To develop statistical methods for the analysis of tag-probe arrays for the study of mixtures of yeast mutants. Aim 4. To develop computer program modules based on the above methods, and to incorporate them into the software dChip in order to extend its functionality for model-based analysis of oligonucleotide array data. We will also work closely with experimental groups in the analysis of real data, in order to ensure that the results from our research will be relevant to the need of the user community. We will aim to improve dChip not only by the methods directly resulting from Aim 1, but will also selectively implement and incorporate other useful analysis methods developed by our group and our collaborators into the dChip framework. Aim 5. To collaborate with the R-based bioinformatics consortium to facilitate the transfer of our methodology to that platform, and to extend the functionality of dChip through R function calls.