MicroRNAs (miRNAs) are newly discovered endogenous small non-coding RNAs (21-25nt) that target their complementary gene transcripts for degradation or translational repression. In animals and plants, microRNAs play very important roles in cell growth, development and death. The biogenesis of a functional miRNA is largely dependent on the secondary structure of the miRNA precursor (pre-miRNA). An accurate prediction of the pre-miRNA secondary structure is important in miRNA informatics. For many years, thermodynamics-based methods have been the dominant strategy for single-stranded RNA secondary structure prediction. Recently, probabilistic-based methods have emerged to replace the free energy minimization methods for modeling RNA structures. However, the accuracies of the currently available best probabilistic-based models have yet to match those of the best thermodynamics-based methods. So this situation motivates us to develop a new prediction algorithm which will focus on microRNA structure prediction with high accuracy. A new model, nucleotide cyclic motifs (NCM), was recently proposed by Major {\em et al.} to predict RNA secondary structure. We propose and implement a novel model based on a Modified NCM (MNCM) model with a physics-based scoring strategy to tackle the problem of microRNA folding. Our MicroRNAfold is implemented by making use of a global optimal algorithm based on the bottom-up local optimal solutions. Our experimental results show that MicroRNAfold outperforms the current leading prediction tools in terms of True Negative rate, False Negative rate, Specificity, and Matthews coefficient ratio.

Mathematics Subject Classification:

This research was supported ***