AbstractIn cell and molecular biology, the microarrays technology (DNA chips) is a fundamental tool to provide information about the behavior of thousands of genes. In this context, and particularly as regards to the analysis of their expression data, a major problem is to group genes with similar behavior according to their activity, namely "discover biclusters". To resolve this problem, we relied on a new approach based on the discovery of biclusters using genetic algorithms in combination with the ideas of Cheng & Church (mean squared residue, row variance, etc). Our work is based on an algorithm of this family : the SMOB , a multi-objective genetic algorithm, developed by Professors Federico Divina and Jesus S. Aguilar-Ruiz. The purpose of our approach is to improve this algorithm by incorporating local search procedures based on the formula of Cheng & Church and intelligent genetic operators. Then, in order to determine the quality of this new method, several tests were conducted using three datasets named Yeast, Human Lymphoma and the Colon Cancer dataset. Finally, from the obtained results, we can conclude that our approach works properly and fully meets expectations by improving the overall classic SMOB.
|Date of Award||2010|
|Supervisor||Wim Vanhoof (Supervisor)|
- gene expression data
- multi-objective evolutionary algorithm
Memetic algorithm for discovering biclusters in microarray data
Amant, S. (Author). 2010
Student thesis: Master types › Master in Computer science