Non-coding RNAs (ncRNA) are RNA molecules that function in the cells without being translated into proteins. In recent years, much evidence has been found that ncRNAs play a crucial role in various biological processes. As a result, there has been an increasing interest in the prediction of ncRNA genes. Due to the conserved secondary structure in ncRNAs, there exist pairwise dependencies between distant bases. These dependencies cannot be effectively modeled using traditional HMMs, and we need a more complex model such as the context-sensitive HMM (csHMM). In this paper, we overview the role of csHMMs in the RNA secondary structure analysis and the prediction of ncRNA genes. It is demonstrated that the context-sensitive HMMs can serve as an efficient framework for these purposes. |