A new Memetic Algorithm for ordering datasets: applications in microarray analysis

Microarray technologies introduced the ability to monitor gene expression for tens of thousands of genes in different conditions. This relatively new tool has transformed genomic research because of the quantity of data that can be collected. However, the amount of data available from microarray data sets is huge and the development of useful tools for its analysis has become one of the major challenges for the utilization of this technology. We present a powerful multi-populational Memetic Algorithm metaheuristic with an embedded Tabu Search, which improves our original algorithm introduced in 2003 for the Gene Ordering problem. We then compare the performance of our method with the well-known hierarchical clustering algorithm developed by the European Bioinformatics Initiative. The evaluation of different approaches using instances from microarray data is one of the challenges in clustering/ordering problems, since it is done mostly visually. To this end, we have evaluated the performance of the methods using images as a way of having a controlled setting. This approach allows a direct visual comparison of the results and provides a workbench for computational experiments that address the sensitivity of the algorithms to noisy measurements.