Analysis of microarray data by genetic algorithm

Microarray gene expression data is used to understand the actions of thousands of genes. Just a few genes out of thousands have a significant impact in any cancer process. Finding these defective genes using experimental data is impractical. To locate the relevant genes, computational techniques are required. A method to identifying cancer candidate genes from microarray data is created. Clustering of similar genes is necessary to find co-expressed genes in different biological conditions. It is important to develop methods to find the few candidate genes for cancer. An optimization process is used for such purpose. A genetic algorithm employs the principles of evolution: selection, recombination, and mutation to solve an optimization problem. Mutual information is used to find the dependency between genes. The two genes are similar if their expression levels are comparable. The similarity as well as positive and negative correlations between genes is considered while clustering them. Interdependence measure tells how the genes are correlated. The genes responsible for a sick state have higher interdependence measures. These genes are defective genes having cancer diagnostic information. Here microarray gene expression datasets from gastric cancer and colon cancer from the public domain are considered.&nbsp