Biology Reference
In-Depth Information
Chapter 4
Integrated Analysis of Gene Expression
Profiling Studies — Examples in
Breast Cancer
Pratyaksha Wirapati, Darlene R. Goldstein
and Mauro Delorenzi
1. Introduction
The high-throughput nature of biological assays such as microarrays has con-
tributed to their rise in importance for studying the molecular basis of
fundamental biological processes and complex disease traits. The widespread
use of microarrays has resulted in a large-scale, rapid expansion of data.
Data from many microarray studies are deposited in databases such as
Gene Expression Omnibus (GEO)
1,2
and ArrayExpress,
3
often to satisfy
journal requirements to make the primary data publicly available. It is
hoped that ready access to the data will facilitate the integration of infor-
mation across different studies.
Information from different studies may be combined at various levels.
Combining test statistics from different studies allows ready integration
of diverse data types without great loss of information. We present a
framework for the combination of results across possibly very heteroge-
neous studies. This framework is based on meta-analytic combination of
z
-scores derived from generalized linear models fitted to primary data for
each study. Multiple hypothesis testing correction is based on the final
combined
Z
.
85
