Tran Doanh C, Brazeau Daniel A, Fung Ho-Leung
Department of Pharmaceutical Sciences, School of Pharmacy and Pharmaceutical Sciences, University at Buffalo, State University of New York, Buffalo, NY 14260-1200, USA.
Methods Enzymol. 2005;396:387-95. doi: 10.1016/S0076-6879(05)96033-3.
Gene array technology has been used to examine gene expression changes following drug treatments, including administration of nitric oxide (NO) donors. High-density arrays represent a powerful and popular method to analyze a large number of genes simultaneously. On the other hand, low-density arrays, available commercially at a lower cost, allow for the use of gene-specific primers, which reduces the risk of cross-hybridization among genes with similar sequence. For certain experiments in which the hypothesis is focused on a selected set of genes, use of low-density arrays might be more productive and cost-effective. Here, we describe our experience using low-density arrays to examine the effect of exposure to the NO-donor isobutyl nitrite on the expression of 23 cancer- and angiogenesis-related genes in mouse tissues. Detailed descriptions of data capture procedures, statistical tests, and confirmation studies using real-time quantitative (RTQ) reverse transcription polymerase chain reaction (RT-PCR) are presented. Three simple statistical methods, namely Student's t test, significant analysis of microarrays (SAM), and permutation adjusted t statistics (PATS), were applied on our gene array data, and their utilities were compared. All three methods yielded concordant results for the most significant genes, namely vascular endothelial growth factor (VEGF), VEGF receptor 3, Smad5, and Smad7. RT-PCR confirmed VEGF upregulation as observed via gene arrays. PATS appeared to be more robust than SAM in handling our small gene array data set. This statistical method, therefore, appears more suited for analyzing low-density gene array data. We conclude that low-density gene array is a useful screening method that can be performed with lower cost and less cumbersome data treatment.
基因芯片技术已被用于检测药物治疗后基因表达的变化，包括一氧化氮（NO）供体的给药。高密度芯片是一种强大且常用的同时分析大量基因的方法。另一方面，以较低成本商业可得的低密度芯片允许使用基因特异性引物，这降低了具有相似序列的基因之间交叉杂交的风险。对于某些假设聚焦于一组选定基因的实验，使用低密度芯片可能更具成效且更具成本效益。在此，我们描述了我们使用低密度芯片检测暴露于NO供体亚硝酸异丁酯对小鼠组织中23个癌症和血管生成相关基因表达的影响的经验。文中给出了数据捕获程序、统计检验以及使用实时定量（RTQ）逆转录聚合酶链反应（RT-PCR）进行的验证研究的详细描述。我们对基因芯片数据应用了三种简单的统计方法，即学生t检验、微阵列显著性分析（SAM）和置换调整t统计量（PATS），并比较了它们的效用。对于最显著的基因，即血管内皮生长因子（VEGF）、VEGF受体3、Smad5和Smad7，这三种方法都得出了一致的结果。RT-PCR证实了基因芯片观察到的VEGF上调。在处理我们的小基因芯片数据集时，PATS似乎比SAM更稳健。因此，这种统计方法似乎更适合分析低密度基因芯片数据。我们得出结论，低密度基因芯片是一种有用的筛选方法，它可以以较低的成本和较少繁琐的数据处理来进行。
