Share Email Print
cover

Proceedings Paper

Toward 'smart' DNA microarrays: algorithms for improving data quality and statistical inference
Author(s): David J. G. Bakewell; Ernst Wit
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

DNA microarrays are a laboratory tool for understanding biological processes at the molecular scale and future applications of this technology include healthcare, agriculture, and environment. Despite their usefulness, however, the information microarrays make available to the end-user is not used optimally, and the data is often noisy and of variable quality. This paper describes the use of hierarchical Maximum Likelihood Estimation (MLE) for generating algorithms that improve the quality of microarray data and enhance statistical inference about gene behavior. The paper describes examples of recent work that improves microarray performance, demonstrated using data from both Monte Carlo simulations and published experiments. One example looks at the variable quality of cDNA spots on a typical microarray surface. It is shown how algorithms, derived using MLE, are used to "weight" these spots according to their morphological quality, and subsequently lead to improved detection of gene activity. Another example, briefly discussed, addresses the "noisy data about too many genes" issue confronting many analysts who are also interested in the collective action of a group of genes, often organized as a pathway or complex. Preliminary work is described where MLE is used to "share" variance information across a pre-assigned group of genes of interest, leading to improved detection of gene activity.

Paper Details

Date Published: 14 December 2006
PDF: 12 pages
Proc. SPIE 6416, Biomedical Applications of Micro- and Nanoengineering III, 64160R (14 December 2006); doi: 10.1117/12.705422
Show Author Affiliations
David J. G. Bakewell, Univ. of Liverpool (United Kingdom)
Ernst Wit, Lancaster Univ. (United Kingdom)


Published in SPIE Proceedings Vol. 6416:
Biomedical Applications of Micro- and Nanoengineering III
Dan V. Nicolau, Editor(s)

© SPIE. Terms of Use
Back to Top