ANALYSIS OF OLIGONUCLEOTIDE GENE EXPRESSION ARRAY DATA

高密度寡核苷酸阵列提供了一种同时监测数万个基因表达水平的手段。这项技术产生的大量数据将成为信息的关键来源，有助于阐明人类基因组计划中确定的大量基因的作用、功能和调控。该项目的长期目标是开发统计方法和计算机程序，以有效分析高密度寡核苷酸阵列的基因表达数据。具体来说，我们的施舍如下。1)改进原始图像数据的分析方法，包括动态网格化和去模糊化、细胞强度和质量分数的计算、背景和梯度校正。2)设计数组对数组比较的归一化程序。3)通过分析一个大型表达式数组数据库，学习和积累关于探针特定效果的统计知识。这些知识将被用于设计分析方法，这将大大减少由于估计和比较基因表达水平的探针效应而产生的可变性。这些知识也将用于设计基于模型的方法来评估基因存在和差异表达。此外，探针序列信息将在可用时纳入分析。4)根据上述方法学研究开发计算机程序。这样的程序将提供给研究界，以促进有效的分析寡核苷酸表达阵列数据。用目前的分析方法，寡核苷酸阵列技术在基因表达研究中还没有充分实现其极高的固有灵敏度。在这个项目中开发的方法将允许改进分析，并将大大提高表达式数组数据的效用。

High-density oligonucleotide arrays provide a means to monitor the expression levels of tens of thousands of genes simultaneously. The massive amount of data generated by this technology will be a key source of information that will help to clarify the roles, functions and regulation of the large number of genes identified in the Human Genome Project. The long-term objective of this project is to develop statistical methods and computer programs for effective analysis of gene expression data from high-density oligonucleotide arrays. Specifically, our alms are as follow. 1) To improve on methods for the analysis of raw image data including dynamic gridding and de-blurring, computation of cell intensity and quality scores, background and gradient correction. 2) To design normalization procedures for array-to-array Comparison. 3) To learn and accumulate statistical knowledge about probe-specific effects through the analysis of a large database of expression arrays. This knowledge will be exploited to design methods of analysis that will greatly reduce variability due to probe effects in the estimation and comparison of gene expression levels. This knowledge will also be used to design model- based methods to assess gene presence and differential expression. In addition, probe sequence information will be incorporated into the analysis when it becomes available 4) To develop computer programs based on the above methodological research. Such programs will be made available to the research community to facilitate effective analysis of oligonucleotide expression array data. With current methods of analysis the very high intrinsic sensitivity of oligonucleotide array technology has not been fully realized for gene expression studies. The methods developed in this project will allow improved analysis and will greatly enhance the utility of expression array data.