Comparative Analysis of Biological Networks Using Markov Chains and Hidden Markov Models

The diverse cellular mechanisms that sustain the life of living organisms are carried out by numerous biomolecules, such as DNAs, RNAs, and proteins. During the past decades, significant research efforts have been made to sequence the genomes of various species and to search these genomes to track down genes that give rise to proteins and noncoding RNAs (ncRNAs) [1], [2]. As a result, the catalog of known functional molecules in cells has experienced a rapid expansion. Without question, identifying the basic entities that constitute cells and participate in various biological mechanisms within them is of great importance. However, cells are not mere collections of isolated parts. Biological functions are carried out by collaborative efforts of a large number of cellular constituents, and the diverse characteristics of biological systems emerge as a result of complicated interactions among many molecules [3], [4]. As a consequence, the traditional reductionistic approach, which focuses on studying the characteristics of individual molecules and their limited interactions with other molecules, fails to provide a comprehensive picture of living cells. In order to better understand biological systems and their intrinsic complexities, it is essential to study the structure and dynamics of the networks that arise from the complicated interactions among molecules within the cell. In recent years, several high-throughput techniques for measuring protein-protein interactions, such as the two-hybrid screening [5] and co-immunoprecipitation followed by massspectrometry [6], have enabled the systematic study of protein interactions on a global scale. Since protein-protein interactions are fundamental to all biological processes, a comprehensive protein-protein interaction (PPI) network obtained by mapping the protein interactome (complete set of protein interactions) provides an invaluable framework for understanding the cell as an integrated system [7]. Furthermore, literature mining techniques have become increasingly popular to search through the vast amount of scientific literature to collect known biological interactions [8]. Nowadays, there exist a number of public databases, such as BioGRID [9] and DIP [10], that provide access to large collections of molecular interactions. In addition to these public databases, there also exist many commercial databases, such as the Yeast Proteome