|
Activity Number:
|
173
|
|
Type:
|
Other
|
|
Date/Time:
|
Monday, August 7, 2006 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
The ASA
|
| Abstract - #307660 |
|
Title:
|
Data Analysis for Multiplex Assays
|
|
Author(s):
|
Shlomo Ta'asan*+
|
|
Companies:
|
Carnegie Mellon University
|
|
Address:
|
Department of Mathematical Sciences, Pittsburgh, PA, 15213,
|
|
Keywords:
|
multiplex assays ; influenza a ; mathematical modeling ; logistic regression
|
|
Abstract:
|
Multiplex assays are becoming standard in immunology research due to their reduced cost and their ability to produce abundant information regarding multiple cytokines in a single run. Common statistical approach for interpreting the data use standard curves based on logistic fitting. In this talk we present an alternative approach which is based on a mathematical model for the binding process of antibodies to the beads. It uses just a few physical parameters, including binding affinity and fluorescence to concentration conversion parameter. We show the model's ability to reproduce experimental results. We then describe how to use the model in conjunction with an inverse problem formulation in order to define standard curves. The advantage of the method over purely statistical models is that it identifies physical parameters which are invariant over a wide range of conditions. Results regarding Influenza A infection will be presented.
|
