Loading...
Please wait, while we are loading the content...
Similar Documents
Genechip ® Expression Analysis Data Analysis Fundamentals Genechip ® Expression Analysis Chapter 4 First-order Data Analysis and Data Quality Assessment..................................... 27 Genechip ® Expression Analysis Affymetrix ® Netaffx Analysis of Promoter Sequences of Regulated Transcripts
| Content Provider | Semantic Scholar |
|---|---|
| Abstract | Foreword Affymetrix is dedicated to helping you design and analyze GeneChip ® expression profiling experiments that generate high-quality, statistically sound, and biologically interesting results. This guide provides information, resources, and tools to help you easily design and analyze experiments and maximize the value derived from your GeneChip data. There is a diverse range of experimental objectives and uses for GeneChip microarray data, which makes the areas of experimental design and data analysis quite broad in scope. As such, there are many ways to design expression profiling experiments, as well as many ways to analyze and mine data. This guide focuses on experimental design elements, statistical tests, and biological interpretation relevant to functional genomics expression profiling experiments, including transcriptional analysis of normal biological processes, discovery and validation of drug targets, and studies into the mechanism of action and toxicity of pharmaceutical compounds. The best designed microarray experiments begin with well-defined goals, anticipated technical pitfalls, and minimized cost. This design phase is critical, as overlooking these key elements can result in highly variable or un-interpretable data. The initial task is to define the objectives of the experiment. Each experimental design should optimize the chances of answering a key hypothesis. There is a natural temptation to test all of the interesting questions in a single experiment, but this approach is dangerous, as overly complex experiments may be un-testable, meaning that the data from these experiments are not statistically powerful enough to answer all questions. In practice this is the direct result of too few replicates or too little experimental control. It is recommended that initial experiments focus on a thorough test of a single key hypothesis which will minimize the arrays required and simplify your data analysis. Testing of more complex hypotheses is best postponed for follow up studies. This serial approach minimizes cost, maximizes statistical power, and simplifies biological interpretation. For example, in a study of the toxic effect of a drug in mice, the critical variable is dose. It may seem desirable to maximize the number of doses, minimize the number of time points, and maintain a single controlled rodent diet. However, the temptation to test many time points or a new diet at the same time may undermine the ability to statistically test the dose response. Ideally, one would want replication to be maximized. True statistical replication means that all test variables are changed independently, one at a time. To … |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | http://www.affymetrix.com/support/downloads/manuals/data_analysis_fundamentals_manual.pdf |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
