Network modeling of RNA secondary structures based on integrative randomized clustering

Ribonucleic acid (RNA) sequences have important cellular functions that attribute to sequence folding mechanisms in forming secondary structures. In recent computational approaches that predict or elucidate secondary structures, probabilistic inference techniques are used to generate a large pool of plausible secondary structures without solely relying on single criteria such as minimum free energy. Efficient computational approach is therefore needed for large data sets to enable further analytical models in computational domain. In this article, we introduce a randomized method to efficiently cluster secondary structures from RNA primary sequences. The clustering method is integrative in that it captures both structural similarity and base-pair probabilities, and further empowers network analysis that unravels important mathematical properties and visualizes global knowledge of folded sequences. The reported results show that our method is effective and efficient in comparison with other algorithms.