Biomedical Engineering Reference
In-Depth Information
CHAPTER 13
A Cooperative Genetic Algorithm
for Knowledge Discovery in
Microarray Experiments
MOHAMMED KHABZAOUI, CLARISSE DHAENENS, and
EL-GHAZALI TALBI
DNA microarrays allow to measure the expression levels of thousands of genes simul-
taneously. This is a great challenge for biologists who see in this new technology the
opportunity to discover interactions between genes. The main drawback is that data
generated with such experiments are so huge that very efficient knowledge discov-
ery methods have to be developed. This is the aim of this work. We propose to
study microarray data on the basis of association rules and to adopt a combinatorial
optimization approach. Therefore, a cooperative method based on an evolutionary
algorithm is proposed and several models are tested and compared.
13.1 INTRODUCTION
DNA microarray experiments have a great interest for biologists, because of their
ability to measure the expression and interactions of thousands of genes simultane-
ously [1]. Although microarrays have been applied in many biological studies [2-7],
the analysis of the large volumes of data generated (large matrices of expression levels
of genes under different experimental conditions) is not trivial and requires advanced
knowledge discovery methods. There exist several kinds of representations to express
knowledge that can be extracted from microarray data. Many data mining techniques
have been proposed to analyze microarray data.
Here, we propose to deal with such data through association rules. This is a general
model that allows to find associations between subsets of genes. Moreover, associa-
tions obtained are interesting to understand the relations between genes on the basis
of the notions of antecedent and prediction.
Search WWH ::
Custom Search