|
Activity Number:
|
528
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Thursday, August 2, 2007 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Statistics in Epidemiology
|
| Abstract - #308804 |
|
Title:
|
Assessment of Impact of Protocol Changes on Classification Accuracy
|
|
Author(s):
|
Wei-min Liu*+ and Wei Wen and Barbara Brady and Sung Lee
|
|
Companies:
|
Roche Molecular Systems, Inc. and Roche Molecular Systems, Inc. and Roche Molecular Systems, Inc. and Roche Molecular Systems, Inc.
|
|
Address:
|
4300 Hacienda Drive, Pleasanton, CA, 94588,
|
|
Keywords:
|
cancer ; classification ; clinical ; microarray ; protocol ; visualization
|
|
Abstract:
|
It is possible to use microarray experiments to classify cancer and other diseases. The protocol including assay of microarray experiments can change during research and development. These changes may influence the signals of microarray experiments and hence influence the classification results. We propose a method called IASCCA (Impact of Additive Signal Changes on Classification Accuracy) to use limited number of cell line experiments and existing clinical training data set to generate many thousands or millions virtual data to simulate the signal changes of the clinical data due to protocol changes and assess the impact of protocol changes on the classification accuracy. We also developed a visualization tool, projection plot, which can be used to show the results of IASCCA.
|
