Design of a Hidden Markov Model for the Analysis of Genomic Changes in Cancer

This paper mainly deals with the recent advances in DNA Micro array technologies and the abundance of Genomic information which plays a vital role for analyzing the molecular mechanisms. Statistical and Machine Learning Algorithms are used to analyze the biological implications in order to discover complex gene expression patterns. A Hidden Markov model (HMM) is designed and the maximum penalized likelihood is used to estimate the parameters in this model. This method is applied to lung cancer micro array experiment. Several regions identified through the HMM are consistent with known recurrent regions of amplifications or deletions in cancer. The association of these abnormal expression regions with the measures of disease status, such as tumor stage, differentiation, and survival are being analyzed. Numerical calculations and graphical representations reveals that genes in these regions may play a major role in the process of carcinogenesis of the lungs.