Applying Computational Methods to DNA Pattern Classification in Cancer Discovery

The accumulation of large-scale data gathered from experiments and tests in the medical field prompts for new computational technologies that can exploit this huge amount of data to its fullest extent. In recent years, several efforts have been made to develop computational models and tools based on Artificial Intelligence (AI) techniques. These tools have been used by biologists for data clustering and classification. In addition to their use in different areas related to knowledge discovery, modeling, and optimization, these tools can be used to develop computational models that might lead to predicting the response of complex biological systems to any perturbation. The discovery and analysis of gene expression patterns of several model organisms represents a fascinating opportunity to explore important normal and abnormal biological phenomena. In this thesis, we address the problem of classifying a subset of genes from broad patterns of gene expression data recorded on DNA microarray. Statistical microarray data 1 2 analysis using these available data from cancer and normal patients was covered as a preprocessing phase which involves filtering (genes selection) and normalization. We conducted a comparative study of two computational tools applied to cancer classification. The study evaluates these tools and compares their performance, classification accuracy, and ability to reveal meaningful gene informatio