Genetic algorithms and artificial neural networks in microarray data analysis: a distributed approach

Microarrays allow biologists to better understand the interactions between diverse pathologic states at the gene level. However, the amount of data generated by these tools becomes problematic. New techniques are then needed in order to extract valuable information about gene activity in sensitive processes like tumor cells proliferation and metastasis activity. Recent tools that analyze microarray expression data have exploited correlation-based approach such as clustering analysis. Here we describe a novel GA/ANN distributed approach for assessing the importance of genes for sample classification based on expression data. Several different approaches have been exploited and a comparison has been given. The developed system was employed in the classification of ER+/- metastasis recurrence of breast cancer tumors and results were validated using a real life database. Further validation has been carried out using Gene Ontology based tools. Results proved the valuable potentialities and robustness of similar systems.