ABSTRACT
Integrated Discovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327 17.4 Supplementary Data Approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327 17.5 Tentative Function Assignment Approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 328 17.6 Future Directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 328
17.6.1 Database Integration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329 17.6.2 “Plug-and-Play” Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329 17.6.3 Combining Analytical Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329 17.6.4 Evaluating Statistical Significance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 330
17.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 331 Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 332 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 332
You complete the molecular portion of your microarray experiment. You run the analysis most appropriate to your data set to see which genes are differentially expressed under the experimental conditions. Now you are left with the question: “What does the
DAAL: “dk2187_c017” — 2005/10/4 — 23:17 — page 322 — #2
array data tell me?” Sometimes this can be an easy question to answer, with only a few genes standing out from the many you have probed. Sometimes the question is not so easy with dozens or hundreds of genes, or even a representation of clustering results, to wade through. In the former case, an expert biological researcher can decide how to interpret the results, come up with new questions, and design new experiments. In the latter case, however, a computational analysis of the identified genes would be considerably helpful and sometimes even necessary.
